ChatGPT

Shell will use agents from C3 AI to shift from basic anomaly detection towards fully-automated predictive maintenance. The global energy giant is building on their current use of the C3 AI Reliability Suite, which already keeps tabs on more than 30,000 crucial pieces of equipment across upstream and downstream operations. Shell now intends to lean heavily into autonomous AI agents, putting them in charge of the entire maintenance lifecycle. Going from that first warning sign all the way to a completed repair, this level of automation strips away the need for constant human oversight and makes sure the company’s resources are pointed exactly where they are needed most. “This expanded partnership with Shell proves what’s possible when enterprise AI is fully operationalised at global scale for predictive maintenance—reducing unplanned downtime and delivering hundreds of millions of dollars in economic value,” said Stephen Ehikian, President of C3 AI. “Shell has built mature AI predictive maintenance programs on our platform, and together we’re now pushing into agentic AI, advancing how this technology can further transform reliability, safety, efficiency, and operational performance.” C3’s AI agents help Shell move past basic anomaly detection In the beginning, Shell used machine learning simply to spot odd patterns in sensor data, giving engineers an early heads-up before things broke. To pull this off, the system ingests a massive amount of real-time operational technology (OT) data and mixes it with business context from ERP platforms such as SAP. The next step introduces AI agents built for actual reasoning and independent action. While older systems stopped at pinging an engineer when things looked unusual, this next-generation framework independently investigates why an alert fired in the first place. Once it pinpoints the root cause, the agent steps up to draft precise work orders, confirm part availability in the inventory, and generate procurement requests. C3 AI’s platform handles the heavy lifting, providing a model-driven space to easily integrate high-frequency sensor feeds with structured financial and maintenance logs. These AI capabilities are trained to learn the normal operating baselines for specific gear, like pumps, turbines, and compressors. The agentic layer sits on top of this foundation. Operators configure an individual agent for a given piece of equipment by defining its objectives and permitted responses. If the core machine learning models detect a deviation from normal operations, this agent activates, gathering extensive contextual data to build a complete picture of the situation. This context usually includes recent maintenance history, environmental conditions, and upstream process variables. Using all that information, it suggests a fix backed by solid evidence. Human operators can then easily approve or override the plan. As the system proves itself over time, Shell can fully automate its responses to certain types of alerts. Connecting straight into systems like SAP is critical here, allowing the agent to work inside the exact same workflows that human planners already use. The real impact of agentic AI for predictive maintenance Putting agentic AI to work at this scale tackles the classic “last mile” headache in predictive maintenance. Many industrial companies can predict failures just fine, but turning those insights into fast, efficient action remains a challenge. Usually, engineers still have to manually dig through alerts, investigate the causes, and write up the work orders themselves. Shell wants to shrink that timeline. By letting AI handle root cause analysis and work orders, the delay between a predicted failure and the actual fix drops. That directly improves equipment uptime and protects production. Moving to a model where repairs only happen when the equipment condition actually demands it naturally saves money, simply because nobody is wasting time tinkering with perfectly fine machinery. Leaving healthy hardware alone also means it lasts much longer. On top of the cost savings, stepping in before a catastrophe hits makes the whole operation much safer and cuts down on environmental risks, which is always top of mind in the energy sector. “What Shell and C3 AI have built on Azure over the past several years is exactly what enterprise AI should look like—real applications, running in production, delivering measurable value at global scale,” commented Sandy Gupta, VP GISV, Software Development Companies at Microsoft. This expanded rollout shows that we are finally talking about practical industrial AI production workflows instead of just algorithms. Rather than just the prediction itself, the real value comes from the system’s ability to act on it with barely any human oversight. See also: Meta Business Agent drives AI-powered conversational commerce Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post How C3 AI agents will automate predictive maintenance for Shell appeared first on AI News. View the full article

Meta has launched Business Agent to automate conversational commerce workflows directly inside its messaging applications. The software allows global retail brands to execute transactions and field support tickets without human intervention. Deploying this architecture places agentic AI directly at the core of social commerce. Meta integrated these workflows natively into Instagram, Messenger, and soon WhatsApp. High volumes of customer interactions overwhelm traditional contact centres. Meta’s platform creates a persistent digital sales representative capable of operating globally. The software operates far outside basic chatbot parameters and can execute concrete administrative tasks. How Meta Business Agent collapses the checkout funnel Consumers frequently discover merchandise on Instagram and initiate a Messenger chat regarding sizing variations. The agent intercepts the query and guides the buyer through the checkout process inside the host application. This architectural model eliminates the high cart-abandonment rates associated with external payment portals. Support operations gain massive efficiency by letting the automated system handle repetitive tier-one tickets. Human support staff gain the bandwidth to manage complex account issues. Contact centre directors can reallocate human capital to specialised retention units. Meta markets this capability as an “infinite team” for retail operators. The software assumes full responsibility for initial contact management. It functions as a first-tier response mechanism operating around the clock. Integrating direct business information allows the system to generate highly specific product recommendations. The underlying models learn and adapt from ongoing consumer interactions. Continuous learning improves performance over time without requiring constant manual reprogramming by internal developers. Retailers with seasonal catalogue changes and volatile consumer demands require such adaptability. Product database updates push directly to the conversational interface via automated syncing protocols. Platform-native architecture design Embedding an agent directly within the Meta ecosystem represents a distinct departure from deploying third-party customer service platforms. A native application integrates deeply with a user’s social graph and historical interactions. External API calls struggle to replicate this level of deep consumer profiling. Tight system integration enables secure, in-chat payment processing. Replicating this complex transaction workflow natively remains exceptionally difficult for external vendors. Lower technical barriers accelerate deployment timelines for small and medium-sized operators. However, large enterprises will need to evaluate how this managed service aligns with their existing CRM databases. Software fed with incomplete or poorly structured information generates subpar consumer interactions. Bad automated outputs actively damage consumer trust and corporate equity. Operations teams will need to ensure that support documentation and product details remain clean and machine-readable. Massive corporate data hygiene projects precede any successful product launch. Engineering teams must establish definitive escalation paths. Business leaders determine the exact scope of tasks the automated system is permitted to handle. Hard-coding operational limits prevents unauthorised internal actions. Creating precise handover protocols for human intervention helps to prevent major service outages. Customers trapped in automated conversational loops experience intense brand frustration. Quality assurance teams consume large portions of the pre-launch phase testing these specific escalation triggers. Engineers run thousands of simulated conversations to locate operational edge cases. Security design presents another major implementation consideration. Firms need highly secure authentication methods to verify a customer’s identity before processing returns or checking order statuses. Identity verification adds a heavy layer of process design to the core engineering timeline. Authentication workflows must integrate perfectly with existing internal Single Sign-On providers. Evaluating vendor dependency The core decision for marketing leaders pits adopting a powerful, integrated platform against maintaining an open, custom-built architecture. Selecting the Meta product secures immense distribution advantages. Platform adoption offers a lower initial development cost compared to building architecture from scratch. The target consumer base already exists natively on the application and Meta manages the heavy core processing infrastructure internally. Independent engineering stacks demand heavy internal maintenance and high operational expenditures. However, they offer greater flexibility and long-term application portability. Engineering departments can select distinct large language models for different departmental tasks. Legal teams can dictate exact data residency policies based on regional government regulations. Many organisations will likely deploy hybrid architectural designs to capture the best of both worlds. In this model, platform-native agents serve as a high-volume concierge, handling initial product discovery and routine catalogue routing. Meanwhile, high-value financial transactions and complex account resolutions are seamlessly handed off to proprietary, secure internal systems. By striking this architectural balance, enterprises can capitalise on Meta’s distribution while maintaining the technical autonomy required for long-term operational security. See also: Amazon brings AI shopping assistant to retailers with Kate Spade Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Meta Business Agent drives AI-powered conversational commerce appeared first on AI News. View the full article

Microsoft has announced the wider testing of its new Autopilot feature at the Microsoft Build event this week, backed by a post on the company’s’ website. Autopilots are described as a new category of agents that can work autonomously on a user’s behalf. Microsoft says each Autopilot has its own identity, and so multiple agents can co-exist within different rule sets, letting users run Autopilots at home, or at work, with separate governance and stipulations limiting or allowing specific activities, according to context. Microsoft’s first Autopilot is Scout, which some internal users at Microsoft have been able to test in beta. The project is now being rolled out to “a select group of customers…and Frontier organizations,” according to the company’s blog. Scout’s initial home will be in acting agentically in Microsoft 365 applications, working across Outlook, OneDrive, SharePoint, and Teams, and be able to coordinate data from each platform to schedule meetings, flag important messages, and generate calendar events to keep workers on track with their tasks. Over time, Microsoft Scout learns about each user’s preferences and work patterns, aligning its activities and priorities to become more efficient and tailored. Under the hood, Scout is built using OpenClaw, the vibe-coded project created over the course of a weekend by Peter Steinberger. Microsoft says Scout comes with enterprise-grade security and controls “so it can be trusted in your organization from day one.” Microsoft has stated that it intends to contribute upstream to the open-source OpenClaw project. Administrators whose organisations adopt Microsoft Scout will be able to validate that any Scout implementations operate securely within the bounds of IT and security policies, and be able to validate agent identities via dedicated Entra entries. The agentic platform will be “managed with the same rigor you expect from any first-party Microsoft service,” the company statement reads. The algorithm takes its data protection policy from Microsoft Purview, and the credentials behind a machine identity are redacted from logs and diagnostics to preserve anonymity. Humans are required to sign off on actions deemed sensitive by the algorithm. The early internal trials at Microsoft have allowed it to expose risks to testers using Scout on the desktop, and the company has tuned the agent to balance any security issues found with an ability to “keep work moving without constant prompting.” Letting Autopilots take the burden of low-level tasks can “keep work in motion so it continues even when your attention is elsewhere.” One feature will be to identify deadlines, block book a user’s calendar so preventing other activities from taking place in the run-up to a deadline, and provide the materials necessary to get around what it’s identified as a bottleneck to progressing an important, focused project. The announcement on the Microsoft website was penned by Omar Shahine, Corporate Vice President of Microsoft Scout, a Redmond lifer whose previous experience includes positions in the Windows Live, OneDrive (previously SkyDrive), and Mac Office divisions at the company. Early adopters keen to try Scout will need to be enrolled in Microsoft’s Frontier programme, have an Intune policy configuration, offer an “opt-in attestation”, and have an active GitHub Copilot licence. (Image source: Pixabay, under licence.) Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Scout from M’Soft is the agentic Autopilot that works across M365 appeared first on AI News. View the full article

Amazon is offering its AI shopping technology to other retailers through a new Agentic Shopping Assistant built on AWS, with Kate Spade among the first brands to use it. The service allows retailers to build AI shopping assistants for their own websites and apps. Amazon said each deployment can be customised to a retailer’s catalogue, customer base, shopping environment, and brand voice. The service is based on technology first developed for Amazon’s own online store. Now, it’s packaging architecture, starter code, and lessons from Alexa for Shopping for use by other retailers. More than 300 million customers used Amazon’s AI shopping assistant last year, according to the company, the assistant generating nearly US$12 billion in incremental sales in the same *******. Amazon said the service lets retailers deploy conversational agents “in weeks,” rather than the years building from scratch might take. The offering includes architecture guidance, starter code, and help from AWS experts and system integrator partners. Kate Spade uses AI for gift shopping Kate Spade, owned by Tapestry, is one of the first brands to use the technology, introducing an AI Gift Concierge that helps shoppers find gift options through a conversational interface. The assistant is focused on gift-buying, with Amazon citing its own data that reveals 53% of shoppers “report stress” during gift purchases. The assistant can recommend gifts based on occasion and customer inputs. Fabio Luzzi, Tapestry’s chief data and analytics officer, told Digital Commerce 360 that the tool came from listening to consumers and identifying what they needed when shopping for gifts. Tapestry tested the assistant for about two and a half months before making it available to consumers, according to Amazon’s announcement. AWS services behind the assistant The system uses Amazon Bedrock, AgentCore, and OpenSearch. Bedrock is the basis for generative AI applications, AgentCore operates AI agents, and OpenSearch is for search and retrieval. Amazon said conversational shopping sessions generate conversion rates 3.5 times higher than traditional keyword-based product searches. The AWS launch follows Amazon’s rollout of Alexa for Shopping in the US in May. That allows users to type shopping-related questions into Amazon’s search bar and receive conversational answers. Under the hood, Alexa for Shopping brings together elements of Rufus and Alexa+ (Rufus was the AI shopping assistant Amazon launched in 2024). (Photo by Shutter Speed) See also: AWS’s legacy will be in AI success Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Amazon brings AI shopping assistant to retailers with Kate Spade appeared first on AI News. View the full article

Standardising grid data through SAP S/4HANA allows E.ON to modernise infrastructure and execute AI deployments. The utility giant manages infrastructure across three distinct domains: energy grids, customer solutions, and energy infrastructure solutions. Maintaining operations across this scope requires continuous capital expenditure on IT hardware and software maintenance. Leadership initially questioned the business case supporting large-scale technology spending. The engineering team proved that persistent financial investment guarantees system stability, affordability, and resilience within a digitised energy network. E.ON prioritises growth, sustainability, and digitalisation as primary corporate objectives. Falling behind in technical capabilities carries long-term financial costs. Infrastructure standardisation drives uptime E.ON executes a cloud ERP migration alongside its SAP S/4HANA implementation. Legacy ERP systems in the utility sector often suffer from extreme customisation. The engineering department rejects fragmented custom builds to avoid this technical debt. Developers integrate established software packages directly into a cohesive architecture. This design methodology guarantees data scalability across the enterprise. The focus on foundational infrastructure delivers highly visible production outcomes. E.ON reports a 77 percent reduction in IT downtime over a five-year *******. Achieving these uptime metrics requires standardising data tables and removing redundant middleware from the technology stack. SAP S/4HANA uses an in-memory database architecture. This design choice accelerates query processing times compared to legacy relational databases. The utility provider leverages this speed to process telemetry data streaming from grid assets in real-time. Fast data processing serves as the prerequisite for deploying any machine learning models against operational data. Technology leaders face intense pressure to match the pace of external software development. E.ON CIO Sebastian Weber notes this pressure creates tension. Consumer software sets expectations for enterprise application deployments. Weber finds consumer AI applications like ChatGPT solve domestic problems effectively, creating internal demands for similar workplace automation. The energy company must close the gap between external software capabilities and internal readiness. Internalising data and cybersecurity operations E.ON treats internal readiness as a primary business objective. The company expanded its internal engineering teams aggressively and hired over 1,000 specialists to bring technical capabilities in-house. The recruitment drive secured more than 500 data experts and 300 cybersecurity professionals. Bringing data engineering in-house allows the utility provider to build proprietary data lakes and audit data governance internally. Retaining internal cybersecurity talent ensures the company maintains strict access controls over the operational technology systems managing the physical energy grid. Engineering now acts as the primary vehicle for achieving commercial targets in the European green energy sector. Of course, managing digital ecosystems at this volume requires strict oversight. The technical team establishes centralised governance structures across all business units. Administrators deploy standardised contracting frameworks and unified IT system management consoles. Having such an administrative architecture in place enforces security standards and cost discipline without restricting feature development. Standardising vendor contracts accelerates software procurement timelines while capping runaway licensing costs. Deprecating isolated innovation hubs Enterprises often isolate experimental technologies in separate business units. E.ON completely abandoned this methodology and deprecated experimental garages and isolated digital labs. Management integrates digital tools directly into active business processes. Keeping innovation teams separated from production environments often prevents applications from surviving the transition to live servers. By forcing developers to build within the core architecture, the engineering department guarantees production viability. “Bringing the system up to speed requires internal readiness,” explained Weber. “It means we must think deeply about investments, prioritisation, and most importantly, people and culture.” Weber expects the operational velocity to remain high, noting the company will not return to previous delivery speeds. New software deployments require precise alignment with business requirements. E.ON enforces a “BizDevOps” operating model. This framework forces developers to build features that generate exact commercial value. Engineers collaborate directly with business analysts during the initial architecture phase. This methodology is paired with targeted employee training. Line workers and managers receive specific instruction on operating newly-deployed tools. This capacity building ensures staff can extract verifiable value from the modernised infrastructure. E.ON is taking a pragmatic approach to AI E.ON manages its AI deployments with deliberate caution and refuses to build proprietary AI platforms from scratch. Instead, leadership prefers to leverage partnerships with established technology vendors. This procurement strategy maintains flexibility across the corporate software portfolio. Engineers explore specific, bounded use cases for machine learning applications. The technical roadmap targets customer service automation, predictive maintenance, and operational optimisation. Applying predictive maintenance algorithms to energy grids prevents catastrophic hardware failures. Sensors detect voltage anomalies and transmit the data back to the central S/4HANA instance. Machine learning models analyse this telemetry to identify wear patterns on physical infrastructure. Maintenance crews receive automated dispatch orders before the equipment actually fails. This active mitigation strategy reduces emergency repair costs and prevents localised power outages. Testing these applications via third-party providers prevents the company from overcommitting capital to unproven frameworks. E.ON embeds these automation features directly into core systems rather than treating them as optional add-ons. The technology serves a customer base of 47 million users. Processing user requests through automated customer service workflows reduces call centre loads and accelerates incident resolution. “In essence, our experience highlights a broader truth about digital transformation,” Weber noted. He explained that pushing new software to production cannot compromise system stability, cybersecurity, or governance frameworks. Without proper alignment with business requirements, advanced technologies fail to deliver value. The modernised architecture provides E.ON with the necessary foundation to scale green energy infrastructure reliably. See also: Walmart’s AI workflows meet the realities of the balance sheet Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post How E.ON uses SAP S/4HANA to modernise the grid with AI appeared first on AI News. View the full article

Walmart has reportedly begun limiting employees’ use of an internal AI assistant called Code Puppy after demands placed on the LLM backing the tool were higher than expected. Employees of Walmart were encouraged to use Code Puppy without any stricture or stipulations as to the limits of use, but Walmart is now assigning employees a fixed number of AI tokens, which limits how much it can be used. Code Puppy was publicised as being able to help with tasks like spreadsheet analysis, creating presentations, and other automatable workplace activities. The change in internal policy is a cost control measure, as LLMs are increasingly transitioning to pay-per-use, rather than the fixed-price, subscription model that gave near-limitless access to AI inference. Walmart has roughly 2.1 million employees, so even modest per-employee queries and task requests can create significant costs. Walmart’s guidance to employees is to use AI where it can create value, and comes with guidance on how workers should choose the right AI tool for any given task. Reporting also says employees have access to other AI platforms paid for by the company. Walmart has expanded AI tool use in the company and provided training for its employees in how to use an AI, encouraging workers to experiment and adopt successful uses. Now the costs of each interaction are being billed directly, it’s among other large enterprises struggling to balance reported improvements in productivity with the cost of achieving the same. At least part of the issue may stem from the methods used to measure productivity in workflows based on AI. Previously, tracking the number and complexity of uses of AI tools as measure of productivity has led to many employees ‘gamifying’ their KPIs – so-called ‘token maxxing’. As recently as April this year, a partner at Sequoia Capital told The Wall Street Journal, “We all should be tokenmaxxing”, an approach that resulted the emergence of AI leaderboards in companies to celebrate those making best use of AI software. Such performative practices at companies will increasingly incur costs relative to the number and complexity of AI tasks, and the model chosen to perform them. Larger models that perform recursive actions (‘thinking models’) use more tokens to process inputs introspectively, leading to higher bills for users. Walmart’s encouragement of workers to choose their model carefully is an attempt to limit spending on expensive, frontier models to achieve relatively trivial tasks, such as spreadsheet analysis and creating presentations. Multi-agentic AI work may also create unexpected costs for employers. When employees instigate iterative loops running on multiple agents in order to create a desired outcome, the real cost of sub-optimal results (and the necessary refining and re-submission of prompts) is now measurable in hard cash. While not all AI providers have changed the entirety of their billing models from fixed subscriptions to per-token, both Anthropic and OpenAI have already moved their higher tier enterprise plans to the new footing. Microsoft’s decision to charge for its GitHub Copilot software development tools as of June 1st is in line with what is rapidly becoming the new financial normality for model providers. Uber recently revealed that it had used up its 2026 budget for AI spend in the first four months of the year; a testament to changes in charging policy affecting end-users. By setting limits on token use on a per-employee basis, Walmart is striving to keep a lid on its ongoing costs, enforce more thoughtful use of AI tools, and enable it to establish the metrics of return on investment in AI. (Image source: Pixabay, under licence.) Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Walmart’s AI workflows meet the realities of the balance sheet appeared first on AI News. View the full article

Phishing protection refers to the category of cybersecurity products that are specifically designed to help companies detect, prevent, and respond to phishing attacks before they develop into full-blown data loss situations. Modern phishing attacks are no longer limited to deceptive emails. They now span the full attack chain, including impersonation, cloned websites, session **********, and real-time credential harvesting. AI-generated phishing attacks are making traditional detection methods less effective, increasing demand for adaptive phishing protection technologies. Given how rampant and complex phishing attacks have become, modern tools focus on the problem from different layers of the attack chain, whether that’s adding in protections directly within the inbox, the browser, or taking actions on the spoofed website itself. What Is Phishing Protection? Phishing protection technologies identify and block malicious attempts to steal sensitive information by misleading employees through deceptive communications. In most cases, the end goal of the attacker is to gain access to an account, steal credentials, and make it out with valuable data. One of the most common phishing tactics is when an adversary impersonates a trusted entity to try and trick the target into willingly handing over credentials, financial information, or even just sending out payments directly. Usually, phishing attempts arrive via email, but modern attacks are branching out and are also targeting employees through email, voice calls, fake websites, and cloned login portals. Why Phishing Protection Matters Phishing remains one of the most common causes of credential theft, business email compromise (BEC), and ransomware incidents. As phishing attacks become more personalized and AI-generated, organizations increasingly rely on dedicated phishing protection platforms to detect threats that traditional email filters miss. Even the most tech-savvy people within a company are falling victim to these more sophisticated, AI-powered attacks. This article will highlight some of the best phishing protection solutions in 2026, which include Proofpoint, Abnormal Security, Memcyco, Barracuda, and IRONSCALES. These platforms help organizations prevent phishing attacks through email filtering, behavioral AI, impersonation detection, browser-level protection, and real-time credential harvesting prevention. Best Phishing Protection Solutions to Evaluate in 2026 PlatformPrimary ApproachBest Suited ForProofpointEmail filtering, threat intelligence, VAP targetingLarge enterprises with high email volumeAbnormal SecurityBehavioral AI for BEC and vendor impersonationOrganizations facing targeted social engineeringMemcycoReal-time phishing site and credential harvesting detectionFinancial services, ecommerce, brand-targeted industriesBarracudaGateway filtering, AI inbox defense, bundled trainingMid-sized to enterprise organizations wanting a single vendorIRONSCALESAdaptive AI, SOC automation, phishing simulationSecurity teams integrating detection with employee training 1. Proofpoint Proofpoint is the most widely deployed anti-phishing email security solution among the Fortune 100 companies. Proofpoint scans over 3 trillion emails every year with its Nexus Threat Graph platform, using AI to examine language, visuals, and URL payloads in emails. Their main differentiator is its VAP (Very Attacked People) model, which identifies who is most at risk of being targeted by a phishing attack. This is based on things like their role, seniority, level of access, history, etc. The platform filters emails, isolates browsers, deploys anti-phishing training, and also comes with automated incident response. Best suited for: Large enterprises with high email volumes that need intelligence-driven phishing protection. 2. Abnormal Security Abnormal Security takes a behavioral approach to detecting phishing emails. The platform establishes a baseline of what normal email communications look like for each individual user, vendor, and partner within an organization’s ecosystem. Any email that deviates from these established baselines gets flagged and quarantined. This approach is especially effective for business email compromise (BEC) and vendor email compromise (VEC). According to the 2026 Attack Landscape Report from Abnormal Security, vendor email compromise accounts for 61% of all BEC attacks. Best suited for: Organizations facing sophisticated phishing attacks, particularly BEC attacks and vendor email compromise attacks. 3. Memcyco Memcyco represents a newer category of phishing protection focused on browser-level interception and credential harvesting prevention. It focuses on what happens outside of the inbox, monitoring activity on the spoofed websites and cloned login portals that hackers use to harvest credentials after someone clicks on a phishing link. The platform detects phishing websites and brand impersonation in real time, often before they show up in the threat databases. Memcyco works directly within the browser at the session level, intervening just at the moment when hackers are about to capture login information. The added kicker is that it can swap out sensitive information that employees enter with decoy credentials. This not only renders them useless, but completely turns the tables and gives companies visibility into the hacker as soon as they attempt to log in with the decoy. Best suited for: Companies that are targeted by brand impersonation, especially in the financial and eCommerce spaces. 4. Barracuda Barracuda Email Protection package includes gateway filtering, AI inbox defense, automated response to security threats, and security awareness training. The platform defends against 13 categories of email threats and can be deployed via API to Microsoft 365. The AI engine learns the organization’s communication patterns and can recognize anomalies that may indicate phishing, spear-phishing, or account take over attacks. Any malicious emails are automatically pulled from all affected inboxes. Over 200,000 organizations use Barracuda for their cybersecurity needs. Best suited for: Mid-sized to enterprise-level companies looking for a single vendor to provide them with cybersecurity training and email protection. 5. IRONSCALES IRONSCALES is a cloud-native phishing protection platform that uses adaptive A, automated incident response, and simulation training to protect organizations from phishing attacks. The platform offers protection to over 17,000 organizations through its API integration with Microsoft 365. The latest Winter 2026 release of IRONSCALES introduces three AI agents. The Red Teaming Agent performs reconnaissance on publicly available data to simulate phishing attacks into the detection model. The Phishing SOC Agent investigates reported phishing attacks automatically. The Phishing Simulation Agent generates personalized training calibrated to what an adversary would actually send. Best suited for: Security teams looking for protection from phishing attacks employee training integrated into a single adaptive platform. Phishing Protection and the Broader Security Strategy First off, no phishing protection tool will ever replace the importance of having good fundamentals in cybersecurity. Solid endpoint security solutions, access controls, and incident response teams will always do the heavy lifting for an organization. Phishing protection tools help fill the gaps that these other tools leave behind. With real-time cloning and multi-channel impersonation threats bypassing the many legacy security systems, this is exactly where dedicated anti-phishing protection tools show their worth. These tools allow organizations to detect and respond to threats that would otherwise slip through their existing security system, whether those threats come in the form of a convincing lure that passed through the email gateway, a cloned login page harvesting credentials in real time, or a deepfake voice call targeting a finance team. How to Choose the Right Phishing Protection Solution This really comes down to your current level of exposure. If the biggest risk is email volume and socially engineered messages getting past filters, platforms like Proofpoint, Barracuda, and Abnormal Security are built for that problem. If the risk sits further down the chain, where users are already clicking through to credential harvesting sites, Memcyco covers that stage. If the priority is tying detection directly into training and SOC workflows, IRONSCALES brings those together. Start with the gap your current stack leaves open and work outward from there. Key Takeaways Modern phishing attacks extend beyond email into browsers and cloned login portals Behavioral AI and session-level protection are becoming increasingly important Different phishing protection tools focus on different stages of the attack chain Organizations should choose phishing protection solutions based on where threats bypass their existing defenses FAQs About Phishing Protection What Does Phishing Protection Do? Phishing protection tools detect and block malicious attempts to gain access to sensitive information through deceptive emails, fake websites, and impersonation campaigns. These tools work at different stages of the attack chain, from filtering emails before they enter the inbox to detecting fake login pages and alerting users in real time. How Do Phishing Attacks Work in 2026? In 2026, attacks use generative AI to create phishing emails that are highly personalized to each victim, deploy highly convincing cloned login portals, and impersonate trusted contacts across multiple channels at the same time. The use of AI removes many of the typical spelling and formatting mistakes that employees may formerly have been trained to spot. Can Phishing Protection Stop AI-Generated Attacks? Behavioral AI and session-level detection are effective because they analyze patterns and intent rather than matching known signatures, which fail against unique AI-generated messages. The best tools adapt continuously, learning what normal communication looks like within an organization and flagging anything that deviates from that baseline. What Is the Difference Between Email Security and Phishing Protection? Email security covers spam, malware, data loss prevention, and compliance. Phishing protection specifically targets credential theft and social engineering, often extending beyond email to web and SMS. Many organizations use both, with email security handling the broad filtering layer and phishing protection focused on the targeted, high-risk attacks that slip through. The post 5 Phishing Protection Solutions Security Teams Should Evaluate in 2026 appeared first on AI News. View the full article

Microsoft’s Majorana 2 quantum chip arrived this week with numbers that are genuinely difficult to contextualise: qubits 1,000 times more reliable than the first generation, a mean qubit lifetime of 20 seconds against an industry norm measured in microseconds, and a revised roadmap targeting a commercially scalable quantum computer by 2029. Behind those numbers is Microsoft Discovery agentic AI, and that platform is arguably the more consequential part of this announcement. To put that in plain terms: most quantum chips today can hold their fragile computational state for a fraction of a second before losing it. Majorana 2 holds it for up to a minute. Microsoft’s own analogy is a phone battery that, instead of dying in a day, lasts nearly three years on a single charge. Majorana 2 was developed with the help of Microsoft Discovery, the company’s agentic AI platform for scientific R&D, which also reached general availability this week. The timing is deliberate. The quantum chip is Microsoft’s proof that the platform works. What Microsoft Discovery agentic AI actually did here The common read on this story is that AI designed the chip. The reality is more specific, and arguably more interesting. The decision to switch the superconducting material from aluminium to lead, which Microsoft says is the single change most responsible for the reliability improvement, came out of years of conventional materials research, not an AI recommendation. What Microsoft Discovery’s agents did was everything around that: managing fabrication workflows, automating measurements that previously took weeks each, breaking down nearly two decades of siloed research data, and surfacing correlations that no single researcher could hold in their head across that volume and variety of information. “As you run AI agents on this data, they’re able to essentially resynthesize and make correlations that we as humans cannot see because no single individual has that much vision across that much data,” said Zulfi Alam, corporate vice president for quantum at Microsoft. That framing matters because it shifts the story from “AI built the chip” to something more accurate: agentic AI compressed the experimental cycle. What would have required extensive trial-and-error to find the right atomic-level recipe for the chip’s crystalline structure could, through AI-driven simulation, be narrowed to a single targeted experiment. “In the new world order, through simulations, you can see where the highly probable target is. And then with that knowledge, you ideally only have to experiment once,” Alam said. The measurement problem, solved One of the more concrete wins the team describes involves qubit measurement; the process of detecting quantum states by determining whether there’s an even or odd number of billions of electrons on a semiconductor wire. When done manually, this takes weeks. Microsoft tried to automate it a few years ago using earlier machine learning and couldn’t. With agentic AI built on Microsoft Discovery, they created a specialised agent that now runs the process automatically and continuously, building three-dimensional maps of qubit conditions at a pace no individual researcher could replicate. “Using agentic AI to automate the measurements was a game changer,” Alam said. The agent handles parallel voltage adjustments across hundreds of parameters simultaneously, something human researchers, thinking linearly and structurally, cannot do. Chetan Nayak, Microsoft technical fellow leading the quantum programme, said the shift has been thoroughgoing: “Agentic AI has permeated almost everything we do, it’s just become kind of a very natural part of our workflow.” Microsoft Discovery goes general The platform that underpinned all of this is now available to enterprise customers. Microsoft Discovery combines specialised AI agents for scientific research, a Discovery Engine for research and reasoning workflows, and enterprise-level security and governance. A free Microsoft Discovery app, usable locally with a GitHub Copilot account, is also in early preview, lowering the barrier for individual researchers who want to run the same kind of agentic workflows. The commercial pitch is clear: the same capability stack that the quantum team used to compress its development timeline is now available to any organisation running intensive R&D. Microsoft has already seen uptake in life sciences, chemicals and materials, energy and manufacturing. Syensqo, for instance, is using it to develop next-generation fluids for semiconductor manufacturing. The 2029 claim, in context Microsoft’s revised quantum timeline deserves a note of editorial distance. The company has moved its target from 2033 to 2029 based on Majorana 2’s progress, which is a significant acceleration, but quantum roadmaps have a history of optimistic compression. The 1,000x reliability figure refers specifically to improvements over Majorana 1’s qubits, not a direct benchmark against competing approaches from IBM or Google, which use fundamentally different architectures. Nayak’s own framing is honest about the incremental nature of this: “Where are we relative to last year? We’re 1,000 times better.” That’s a meaningful year-on-year milestone. Whether it holds at the pace required to reach utility-scale quantum computing by 2029 is the question no one, including Microsoft, can yet answer. See also: *** and Germany plan to commercialise quantum supercomputing Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Microsoft’s Majorana 2 quantum chip is also a case study for agentic AI in R&D appeared first on AI News. View the full article

Anthropic’s IPO filing marks the maturation of generative AI from a research-heavy venture phase into a stabilised enterprise utility. Model developers operating in private markets have prioritised rapid iteration and maximum compute performance over predictable billing cycles. Taking a foundational provider public aligns those engineering goals with standard corporate procurement, introducing structured release schedules and established pricing frameworks that decision-makers require for multi-year planning. William Samengo-Turner, Technology Sector Lead at A&O Shearman, said: “If Anthropic pursues an IPO, the most important question isn’t whether public markets are ready for AI—it’s whether AI is ready for public markets.” The enterprise consumer sits directly at the centre of this maturation. Companies integrating Claude into their proprietary workflows can now plan around how public market structures will formalise Anthropic’s pricing tiers, API rate limits, and enterprise service agreements over the coming years. Establishing a public valuation framework Institutions looking to capitalise on generative machine learning have largely invested in hardware providers and infrastructure layers. This indirect approach allowed companies to build out the necessary compute clusters without taking on the concerns around model hallucination or algorithmic copyright disputes. Samengo-Turner notes that public investors have focused on the surrounding ecosystem: “Investors have been able to buy the ‘picks and shovels’ of the AI *****—with infrastructure, semiconductor, and software businesses benefiting from it. Anthropic would offer one of the first opportunities to invest directly in a company building frontier models at scale.” Pricing that asset class presents immense difficulty. Anthropic and its competitors require continuous, massive capital expenditures to train successive model generations. Converting these capital requirements into a public structure introduces high operational drag for both the provider and the client. A public Anthropic will need to balance the need to buy tens of thousands of GPUs against the need to post favourable quarterly earnings, which requires passing those compute costs onto the end user in a predictable manner. Karthik Hariharan, Senior Engineering Manager at DoorDash, commented: “Both OpenAI and Anthropic are racing to IPO ahead of each other and catch up to SpaceX/xAI. The problem is whoever lands first probably sets the floor and ceiling for public market pricing that others will follow for at least 12–18 months.” If Wall Street demands aggressive margin expansion following the IPO, enterprises should anticipate tighter licensing terms and the potential deprecation of older and less profitable model versions. This creates forced migration cycles for corporate development teams, requiring them to constantly update their API integrations to maintain access to the most cost-effective models. The B2B dependency The commercial structure of these public listings relies heavily on enterprise adoption because the consumer market lacks the scale to offset computing costs. Suvrankar Datta, Principal Investigator at ****** Lab, explained: “There are eight billion human beings on the planet… of the eight billion, only 100 million can afford to pay for Claude at the current rate. Even if they pay $20 per month for Claude, it still won’t be able to survive without an IPO.” The $20 monthly consumer tier cannot fund billion-dollar server clusters. Therefore, model providers must extract their required revenue from corporate budgets, integrating their tools into daily enterprise operations such as human resources, legal document review, and customer support triage. Nate Elliott, AI Analyst at Emarketer, said: “We’re about to find out whether the market thinks AI is a consumer story or an enterprise story. Because while Claude has built a solid enterprise user base, it’s just not competitive as a consumer AI platform.” Emarketer forecasts that only 5.4 percent of US internet users will use Claude in 2026, far behind the 36.6 percent who will use ChatGPT and the 27.4 percent who will use Gemini. “The good news for Anthropic: more than 60 percent of US AI users say they use these tools for work, and we believe that percentage will only grow,” adds Elliott. Anthropic will need reliable, high-volume enterprise contracts to demonstrate steady revenue growth to prospective shareholders. Boardrooms can use this dependency to negotiate longer-term price locks and favourable data governance agreements before the public market forces Anthropic to prioritise short-term yield over market penetration. Margin pressures and market consolidation The impending public offering acts as a forcing function for commercial discipline across the entire generative computing sector. Rather than viewing this negatively, enterprises can see it as the end of unpredictable startup behaviour and the beginning of reliable vendor management. Smitarani Tripathy, Social Media Analyst at GlobalData, said: “Discussions reveal increasing concerns around the economics of the AI ecosystem, with several influencers questioning whether massive investments in model development and compute infrastructure can ultimately translate into sustainable profits.” Tripathy further explains that this filing initiates an “AI capital markets race,” where model providers must demonstrate revenue growth, operational efficiency, and defensible business models alongside innovation. If a vendor goes public and fails to achieve sustainable profits, they may aggressively alter their service-level agreements or sunset key API endpoints to reduce overhead. “Future valuations will hinge on enterprise unit economics, gross margins, and customer retention, forcing severe consolidation among smaller players unable to scale commercial revenue engines or achieve software-like operating leverage,” explains Tripathy. Companies building proprietary tools around smaller language models must prepare for those providers to be absorbed by larger entities or forced out of the market entirely. Designing middleware layers that allow smooth swapping of foundational models is a vital defensive measure against vendor bankruptcy or acquisition. In addition, enterprises should expect more aggressive rate limiting. In a private model, absorbing the compute cost of heavy user requests serves as a loss leader to build market dominance. In a public model, unmetered access destroys gross margins. Businesses will likely see the introduction of complex, tiered pricing structures that penalise erratic workloads and reward predictable, batch-processed data requests. The test for high-capital innovation Anthropic’s journey to the public exchange serves as a barometer for how institutional capital values resource-intensive technology. Samengo-Turner expands on the wider implications for venture-backed companies: “The significance extends well beyond the AI sector. A successful listing could become a reference point for how public markets assess a new generation of technology companies that combine immense capital needs, world-class research talent, and long-term strategic ambitions.” He notes that this event could “encourage more venture-backed technology companies to revisit public markets after a decade in which many of the sector’s biggest growth stories remained private.” If Anthropic successfully sets a public valuation framework, a wave of machine learning companies will likely follow, moving the entire vendor ecosystem toward strict financial compliance and margin protection. “Ultimately, investors will be evaluating more than Anthropic’s prospects,” Samengo-Turner concludes. “They will be testing whether public markets are prepared to support the next generation of technology champions.” See also: Anthropic releases Claude Opus 4.8 Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Anthropic IPO filing marks AI maturing into enterprise utility appeared first on AI News. View the full article

Since its announcement in April this year, the proposed changes to billing methods on GitHub Copilot were a source of much speculation: how much more or less would a pay-a-you-use AI cost an organisation or individual compared to a flat-rate, monthly subscription? Just a day into the changeover to token-based billing for the LLM-based service, software developers and IT departments have been reporting their findings online – and the shortened version is that, as of 1st June 2026, using GitHub Copilot in software development and deployment just got a whole lot more expensive. What are the changes to GitHub Copilot’s charging scheme? Although subscription prices have not changed (Copilot Pro $10 per month, Pro+ $39, Business tier $19 per user, and Enterprise $39 per user per month), the prices now refer to a monthly number of credits that can be spent on the various AI models made available on the GitHub platform. For a typical user, one credit costs a single cent, and depending on the model variant selected at the point of inference, credits are then deducted according to how much silicon effort is expended by the AI. Thus, a Copilot Enterprise user receives 3,900 credits per month ($39), a Copilot Business user receives 1,900 credits ($19). Users will burn up their credits in the form of tokens which are priced differently, according to the power and type of model used. For example, using ChatGPT-5.2, it costs $1.75 per million input tokens (a token can be thought of as nearly-a-word), output tokens cost $14 per million, and cached input (the information held by the LLM to provide ongoing context to a series of queries, for example) are priced at $0.175 per million tokens. When users reach the end of their allotted number of credits, they have the option to buy more. Code completions inside a developer’s IDE (integrated development environment) and ‘next edit’ suggestions will be free, but Code Review processes will cost at the same rates as other GitHub Copilot activities. Will users pay more to use Copilot? Whether or not an average user will end up paying more depends very much on the individual user, and to a certain extent on who you ask. The Comments section of the GitHub Community Discussions page that announced the changes back in April 2026 has many reports of users finding that their credits are being exhausted much more quickly than expected. User ‘rvs99’ said, “My 12% of total AI credits burned like anything for very minor task. I used Claude Sonnet 4.6 as usual and in response it barely updated 2-3 lines in total 6 files which costed like ~$0.35 per line updates.” ‘prhost’ posted a screen-grab of their account dashboard that showed 3,705 credits remaining of an allowance of 7,000 after one day’s use, and stated “It would be easier to shut down the project. [Microsoft] shot themselves in the foot.” User ‘zoomp05’ summarised the tone of most commentators: “The strategy is clear, but it would have been good to say from the beginning, ‘This is a subsidized trial’ or something similar, to promote our tool.” The initial subscription offerings from GitHub, now deprecated, were likely seen by the platform’s owners, Microsoft, as loss leaders. It was immediately apparent that allowing users to burn far more tokens than their subscription value represented was never going to be sustainable. Cursory reading around the internet away from the big model providers’ announcements and posts revealed that, as a business model, subscription-based billing could only be temporary. What is surprising, perhaps, is the surprise of many users that their coding platform is now being billed for at levels in keeping with suppliers’ costs. Running an LLM is not a cheap undertaking, especially considering the additional sums involved in developing new models, post-training, maintenance, data centre construction, future loan repayments, and so on. What businesses might do now Those invested in supporting their development teams with LLM-based coding tools have several options they might consider: Reassess the ROI that AI coding platforms bring, and adjust budget allocations accordingly. Consider which processes of the software development workflow may be cheaper to hand off to AI (junior developer-level code creation, for example), and which are cost sinks (code review, multiple-agent workflows, fast-cadence Actions, etc.). Look for alternative, lower-cost platforms. These fall into three main camps: Open models hosted on-premise. These are not frontier LLMs, and lack many of the features of the coding ‘harnesses’ that professional coding platforms offer. Hosted near-frontier models from LLM providers such as Huawei and Alibaba. ‘Secondary’ coding platforms such as Cursor may offer temporary respite, although be aware that many of the alternatives use frontier, better-known models from OpenAI and Anthropic, and are likely to therefore adopt the same per-use billing as GitHub Copilot. (Image source: Pixabay, under licence.) Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post GitHub Copilot users see token-based price hikes appeared first on AI News. View the full article

Automation is becoming a ******* part of how financial markets are approached, and forex trading is one area where this is becoming easier to notice. As the tech world improves, more traders are looking for ways to stay involved in the market without the need to sit in front of charts for hours at a time. A large part of this thought process comes down to forex robots, which are designed to carry out trades based on a set of pre-made rules. These tools are not new, but they are becoming more refined and easier to use as time goes by. If you are to look at the future of automated trading through the best forex robot reviews, you’ll have a clearer idea of how these systems are being used today and how they may continue to develop over time. How automation use is growing in forex trading Automated trading has been around for a while, but the possibilities available today are more advanced than what traders had access to in the past. Forex robots are able to scan the market, look for specific trade setups, and place trades without a trader ever needing to lift a finger. These systems follow a set of rules that are usually based on technical indicators or past price behaviour. Basically, they are designed to look for patterns and react when certain conditions are met. Some systems are quite basic, and some are built to handle ******* amounts of data and more detailed strategies, so that you don’t have to constantly monitor the market. The growing role of data and AI Artificial intelligence is growing at a fast pace, and playing a more noticeable role in trading these days. Some systems are now able to identify patterns that might not be easy to spot when looking at charts manually. This doesn’t mean that every forex robot is fully driven by AI, but many are now starting to use data in smarter ways. In some cases, systems can adjust how they react based on current market conditions, not following the exact same response every time. This is where FXSentry has become especially useful, as they help break down how different systems work, making it easier for traders to understand what is happening behind the scenes before deciding which one to use. Efficiency and ease of use for traders One of the main reasons that automated trading continues to grow is how it makes trading easier to manage. Certainly not everyone has the time to sit and monitor charts throughout the day, especially in a market that operates almost around the clock. Forex robots run in the background and only jump in when certain conditions are met. This means trades can still be placed even when you’re not actively watching the market, making trading feel more manageable and less overwhelming. Why reviews are important Since there are so many forex robots available, choosing the right one can feel a bit stressful. Each system works differently depending on how it is built and what type of strategy it follows. Reviews can give a clearer picture of how a system operates, how it manages trades, and what kind of results it has produced over time, which makes it easier to compare options and avoid choosing a system without fully understanding how it works first. Important things to keep in mind Even though automated trading can be helpful, it is not without its limits. Markets can change quickly, and a system that performs well under certain conditions may not perform the same way when those conditions change. Forex robots may struggle when something unexpected happens that deviates from their rules. There are also practical factors to think about, like internet connection, platform reliability, and how quickly trades are executed. Because of this, automated systems should be used as support tools not something that replaces decision-making completely. Keeping an eye on performance and making adjustments when needed is still an important part of trading. What the future holds Looking ahead, automated trading could become more advanced as technology continues to improve. Systems may become better at reacting to market conditions and handling more complex data in a way that feels more natural. The future of automated trading will depend not only on how technology improves, but also on how traders choose to apply these tools in a practical and informed way. Wrapping it up Automation is becoming a more common part of forex trading in daily life, helping traders manage their time, follow structured strategies, and stay active in the market without a need for constant monitoring. Exploring the future of automated trading with FXSentry shows how these tools can support a steadier and more manageable approach. When used with a clear understanding of how they work and where their limits are, they can form a useful part of a modern trading setup. The post The future of automated trading with the best forex robot reviews appeared first on AI News. View the full article

A Google Cloud survey found that 90% of developers are already integrating AI into their daily work, and on Steam, 7,818 titles disclosed AI use in 2025 alone, a 681% increase over the previous year. AI in video game development is not a side experiment. It is restructuring the pipeline from concept through launch, and the areas where it is having the most concrete impact are worth examining individually. Smarter NPCs and adaptive gameplay Non-player character behaviour has moved well past scripted decision trees. Ubisoft’s La Forge division developed Ghostwriter, a generative AI tool that produces first-draft NPC dialogue so writers can concentrate on narrative not volume. Large language models now give NPCs genuine memory in sessions and responses that hold up under improvised player input. Alongside this, AI systems monitor player performance in real time to adjust difficulty dynamically, while story engines weave branching subplots on the fly, making each playthrough genuinely distinct. Generative AI and asset creation Andreessen Horowitz has documented cases where concept art generation dropped from three weeks to a single hour once AI tools entered the workflow. Tencent’s Hunyuan3D-PolyGen produces art-grade 3D assets with artists reporting efficiency gains of over 70%, while Meta’s WorldGen can generate a traversable 3D environment from a text prompt in around five minutes, game-engine-ready for Unity and Unreal. Audio is following the same trajectory, with tools like ElevenLabs enabling voice generation and localisation at a speed that traditional recording pipelines cannot match. Quality assurance and playtesting QA is where AI is having a substantial operational impact. EA has deployed reinforcement learning agents to autonomously play and stress-test games, catching edge-case bugs in a wider range of gameplay styles than human testers could cover. Square Enix has announced plans to automate 70% of its QA and debugging using generative AI by 2027, in partnership with the University of Tokyo. The emerging model in the industry is hybrid: AI handles the mechanical volume while human testers focus on judgement-driven issues that automation cannot resolve. Procedural generation and living worlds Modern AI-assisted procedural systems go beyond earlier rule-based approaches by conditioning generation on context. Narrative engines now weave branching subplots that respond to player actions and inferred emotional cues, so each session reflects the shape of an individual playthrough not random variation. Research frameworks like PANGeA are demonstrating that large language models can maintain narrative coherence in dynamically generated content, removing the need for the exhaustive hand-authoring that has traditionally limited branching game stories. AI for browser and web game development Web games are structurally simpler than console or PC titles, HTML5, fast load times, pick-up-and-play mechanics, and that simplicity makes AI tools unusually effective at covering the gap for developers without deep technical or artistic backgrounds. Generative AI can handle concept art and basic asset creation in a fraction of the usual time, while AI-assisted code generation helps less experienced developers get a functional prototype into a browser environment. Tools like FRVR AI let any user generate a playable browser game from a text description alone. Platforms like Poki give those games a natural home: free to play for users, with revenue earned through advertising, making the path from prototype to published title more accessible than it has ever been. The limits and labour questions The expansion has not been frictionless. The flood of low-quality AI-generated titles that hit Steam in 2025 raised real questions about quality floors in an environment where content is cheap to produce. Voice actor unions and writers’ guilds are still negotiating the terms under which AI can generate dialogue or clone voices, and the outcome will shape how studios deploy these tools in character-driven productions. What the evidence so far suggests is that AI in video game development pays for itself when it shortens the distance between a creative intent and a usable output, and studios finding genuine value are putting it precisely where the production bottleneck sits. The post AI in video game development: How artificial intelligence is reshaping the industry appeared first on AI News. View the full article

OpenAI’s latest governance frameworks offer enterprise leaders a structured blueprint for scaling safe and compliant AI deployments globally. The adoption of large language models has steadily progressed towards requiring sustainable, commercial-grade architecture. OpenAI has released its Frontier Governance Framework (FGF), documenting how the organisation addresses systemic risk assessment and mitigation. The framework maps directly to the EU’s General-Purpose AI Code of Practice and California’s Transparency in Frontier AI Act, known as the TFAIA. This publication provides a highly practical template, detailing how internal systems and deployment pipelines can be structured to support high-capability machine learning models securely. Translating these regulatory structures into business strategy begins with understanding defined threat categories. The framework defines systemic risk as foreseeable material risks of severe harm. Specifically, this includes scenarios where a model contributes to greater than 50 fatalities or causes $1 billion in property damages from a single incident. While these scenarios sit at the extreme edge of probability, codifying them allows deployment teams to build appropriate safeguards. By defining boundaries early, enterprises can allocate precise compute resources and engineering hours towards continuous post-deployment monitoring and third-party auditing; ensuring applications remain compliant over their lifecycle. Applying tiered risk evaluations to internal systems OpenAI categorises threats across specific domains: cyber offense, chemical, biological, radiological, and nuclear (CBRN) risks, harmful manipulation, and loss of control. The categorisation system utilises distinct risk tiers to evaluate model capabilities. For example, a Tier 3 cyber offense rating applies to a tool-augmented model capable of identifying and developing functional zero-day exploits of all severity levels in many hardened real-world systems without human intervention. In the CBRN category, a Tier 3 model could enable an expert to develop a highly dangerous novel threat vector, comparable to a CDC Class A biological agent, or autonomously complete the synthesis cycle of a regulated biological threat. Rather than viewing these capabilities purely as hazards, internal security teams can use these tiers to establish defined limits for their proprietary model instances, knowing exactly when a coding assistant or research tool requires heavier oversight. The framework also outlines risks tied to harmful manipulation, described as the purposeful distortion of human behaviour, such as using model capabilities for influence operations or election interference. OpenAI notes that this area remains exploratory and is best addressed through system-level mitigations, like post-deployment monitoring, rather than pre-deployment evaluations. For consumer-facing businesses, this suggests that marketing automation systems using language models simply require real-time content classifiers to ensure they generate objective public messaging. Addressing the risk of humans losing the ability to reliably direct or shut down a system, the framework labels this vector as loss of control. A Tier 2 model in this category demonstrates the capability to reliably evade detection across various evaluation methods, including evading chain of thought monitoring. A Tier 3 model is described as being superior to the most expert humans in executing most complex projects and can operate autonomously for extended, sustained periods of time. It demonstrates highly detailed situational awareness and stealth such that monitoring the model and its chain of thought cannot reliably detect or rule out evasion of human control. By setting these parameters, businesses relying on autonomous agents for supply chain logistics or financial trading have a defined mandate to build deterministic fail-safes and maintain consistent human oversight in automated workflows. Addressing integration challenges and information security OpenAI aligns its internal security with ISO 27001, 27017, 27018, and 27701 standards, alongside SOC 2 Type II evaluations. To protect unreleased model weights, the company employs encryption for data at rest and in transit, multi-factor authentication, and strict multi-party approval protocols. Internal personnel undergo regular training, and model execution occurs in a sandboxed environment with restricted egress by default. When enterprises mirror this setup, they establish a secure baseline for internal operations. Integrating models into proprietary corporate data environments often leads engineering teams to rely on Retrieval-Augmented Generation and dense vector databases. Securing these databases against adversarial prompting or data extraction attempts requires dedicated computational overhead. Every API request passes through security classifiers before hitting the vector database, and the retrieved context is screened before generating a final response. While bridging modern cloud-hosted AI governance structures with older mainframe data silos forces teams to build bespoke, heavily-encrypted middleware, this engineering work results in stable enterprise-ready infrastructure. Maintaining ecosystem compliance and incident response To maintain accurate risk baselines, OpenAI solicits input from external domain experts and independent third-party evaluators. These external experts help stress-test safeguards for models approaching a new risk tier and provide independent opinions to the internal Safety Advisory Group. CDOs within enterprises can similarly benefit from external auditing retainers to independently verify that their localised model deployments remain within acceptable risk thresholds. Connecting to the broader regulatory ecosystem, external reporting dictates the ongoing operational cadence. OpenAI documents its mitigation results in a Safety and Security Model Report. Under the EU AI Act provisions, the company commits to evaluating whether to update these reports for its most capable models every six months. Updates to the reports are considered required if a model’s capabilities materially change through post-training or if integrations into internal systems increase risk. The responsibility for EU compliance rests with OpenAI Ireland Limited, while OpenAI OpCo LLC manages obligations under the TFAIA in the US. To manage sudden software anomalies, OpenAI utilises an AI Safety Incident Response Plan, abbreviated as the AIRP. This plan dictates procedures for triage, investigation, and external reporting of severe safety incidents. Potential incidents are flagged through automated monitoring, employee escalation, or end-user feedback. Once flagged, response teams investigate the root cause, scope, and impact, taking action to mitigate and contain the event. Enterprise leaders can easily mirror these response mechanisms; establishing parallel internal response units capable of adjusting anomalous API behaviour proactively. Within OpenAI, updates to the framework can be proposed by various leaders, including the Head of Safety Systems, CISO, and General Counsel. The company conducts a formal Framework Assessment at least once every 12 months; evaluating changes in law, new model capabilities, and industry standards. The integration of advanced computational models remains a viable path to corporate efficiency, and adopting these frameworks ensures the internal architecture is well-prepared to handle modern compliance demands securely. See also: Anthropic releases Claude Opus 4.8 Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Scaling safe enterprise AI with OpenAI governance frameworks appeared first on AI News. View the full article

Anthropic has released Claude Opus 4.8, an upgrade to Claude Opus 4.7 that the company says brings improved results for coding, agent work, reasoning, and knowledge work. The platform can be used through claude.ai, Claude Code and the Claude API, with the API name claude-opus-4-8. The company has also altered some of the details of its product line-up. Users of claude.ai and Cowork can set the amount of effort Claude applies to a response – essentially, affecting the number of tokens the model will burn. Claude Code also has dynamic workflows, a feature that plans work, runs parallel sub-agents, verifies outputs and reports back to the user. Finally, the Messages API accepts live changes to the messages array, which Anthropic says lets developers update instructions during a task without breaking prompt cache use or needing a separate user turn. Anthropic said the price for use of Claude Opus 4.8 when not in ‘fast’ mode will remain at $5 per million input and $25 per million output tokens, while fast mode costs $10 per million input tokens and $50 per million output. Fast mode for Opus 4.8 works at 2.5x, the company’s announcement post states. The company has positioned Opus 4.8 as designed for coding and agentic workflows in coding, where the model can use tools inside a context and check its own work. It says Opus 4.8 improves on Opus 4.7 on benchmarks for coding, agent skills, reasoning, and office work. There is a System Card that can be examined for further subjective detail. Anthropic’s announcement cites several companies that have tested the platform before its wider release, including those operating in software development, law, finance, and research. Several testers commented on the platform’s agentic workflows, with one noting a cost parity with GPT-5.5 when running its internal benchmark tests. A comment from CursorBench said Opus 4.8 used fewer tool steps to achieve the same level of output. Anthropic says Opus 4.8 is less likely than its 4.7 predecessor to pass flawed code without comment, which it describes four times less likely. It says the platform showed lower rates of deception or the tendency to go along with misuse than Opus 4.7 and is comparable in this regard as those exhibited by Claude Mythos Preview. Effort control helps users to manage any trade-off between quality, speed, and token burn rates. Opus 4.8 defaults to high effort but on coding tasks, the company said the higher default only uses the type of token numbers of Opus 4.7, but performs better. Users can opt for ‘xhigh’ for work that needs more computation. Anthropic said it has increased Claude Code rate limits to support the resulting higher token use. Dynamic workflows in Claude Code are designed for large codebases, and can migrate codebases of hundreds of thousands of lines. These features are currently in research preview and are available on the Enterprise, Team, and Max plans. The Messages API updates instructions during an agent’s run, with edits inside the messages array being used, for example, to update permissions, change token budgets or context while agents continue their work. Anthropic also used the release to suggest it’s developing models that provide current levels of ability at less cost to the user, and will release a class of model that’s better than the current Opus platform. Its roadmap includes Project Glasswing, under which a group of organisations is using Claude Mythos Preview for cybersecurity scanning. Anthropic said models at that capability level require stronger safeguards before release to all customers. It expects to bring ‘Mythos-class’ models to customers in the coming weeks. The additional controls in 4.8 will expose the cost and effort trade-offs to users as the company transitions to token-based billing from subscription tiers. (Image source: Pixabay, under licence.) Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Anthropic releases Claude Opus 4.8 appeared first on AI News. View the full article

Google Pay is overhauling its payment infrastructure for an impending wave of transactions from AI agents. The latest updates introduce the Universal Commerce Protocol and a new server architecture, positioning Google Pay as a central clearinghouse for purchases executed by autonomous agents rather than human users. AI agents – designed to perform tasks like booking flights or ordering supplies – cannot effectively navigate the multi-step, visually-oriented checkout pages built for human interaction. Google is attempting to replace this UI-dependent model with a stable, API-driven backend for machines. This restructuring of Google Pay introduces several components: Universal Commerce Protocol (UCP): This is a new specification intended to standardise how AI agents communicate with payment and merchant systems. It creates a common language for initiating transactions, confirming inventory, and handling fulfillment details. The objective is to eliminate the need for developers to build bespoke integrations for every merchant or payment provider an agent might interact with. New Merchant Commerce Platform (MCP) server: Google is deploying a new server-side system to act as an intermediary. This MCP server manages merchant integrations and analyses transaction trends. For developers building agents, it abstracts away the complexity of the commerce backend. For Google, it centralises a vast amount of transactional data from agent-driven activities. Dynamic callbacks for Android native: To facilitate more complex checkouts, Google is enabling dynamic callbacks within its Android Pay API. This allows for real-time adjustments to an order (e.g. updating shipping costs based on a new address or recalculating tax) without forcing the user or agent to restart the entire process. It makes the transaction flow more resilient to mid-process changes. Expanded WebView support: The company is extending payment support within WebViews. This is a critical detail, as it allows transactions to be completed inside third-party applications, particularly social media platforms where conversational commerce is expected to increase. Agents operating within these environments can now execute payments natively. Realities of machine-to-machine commerce The concept of a customer journey, once defined by clicks and page views, now extends to an agent’s ability to parse product data and execute a transaction via an API. Marketing leaders now have to consider “search engine optimisation” for machines. Product information, pricing, and availability will need to be presented as machine-readable data, not just persuasive copy for a human audience. If an AI agent cannot parse your inventory data to make a purchasing decision, your business becomes invisible in this new commercial channel. The introduction of the MCP server also raises questions about data governance and vendor dependency. By routing transactions through its platform, Google gains a privileged view of commerce trends driven by AI agents. CIOs must assess the long-term implications of building reliance on a proprietary protocol and a centralised data aggregation point. The convenience of a universal standard comes with the strategic cost of platform lock-in. New architectures for security and trust Authorising transactions initiated by an autonomous agent presents a new set of security challenges. A faulty or malicious agent could execute unauthorised purchases at scale. Google’s answer is the introduction of cross-device biometric authentication. This mechanism allows an AI agent to programmatically request human verification for a transaction. A user could receive a prompt on their phone to approve a purchase an agent has arranged on their laptop. This approach establishes a “human-in-the-loop” security model for high-value or sensitive transactions. It provides a necessary kill-switch and audit trail for agent activities. Defining the policies for when an agent can act autonomously versus when it must seek human approval becomes a new area of corporate governance. These rules will need to be encoded into the agent’s operational logic, creating a direct link between business policy and software behaviour. These latest updates to Google Pay are an early but concrete signal of the architectural changes required to support a machine-driven economy. Enterprises that continue to view their digital presence as a collection of websites for human consumption will be unprepared for this next phase of commerce. See also: Google folds Display Ads into AI-first Demand Gen platform Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Google Pay preps for AI agents with Universal Commerce Protocol appeared first on AI News. View the full article

NBA Commissioner Adam Silver said the league plans to introduce an automated system for certain officiating decisions, including out-of-bounds calls. The system would use AI and cameras placed around the court to determine possession. Silver compared the approach to Hawk-Eye, the tracking technology used for line calls in tennis. Disputed call preceded Silver’s comments Silver’s appearance came after a disputed call in Game 5 of the Western Conference finals between the Oklahoma City Thunder and San Antonio Spurs. Late in the third quarter, Spurs centre Victor Wembanyama was ruled to have touched the ball last on an out-of-bounds play. The replay showed the ball had bounced off the foot of Thunder forward Chet Holmgren. The call stood after the officials conferred. The call drew attention after Oklahoma City took a 3-2 lead in the series. Silver said the NBA eventually intends to remove that category of objective calls from on-court officials. NBA partnership with Hawk-Eye started in 2023 The NBA previously announced work with Sony’s Hawk-Eye Innovations. In 2023, the league said it had entered a multi-year partnership to deploy 3D optical tracking technology. The partnership followed several years of testing at Summer League and NBA arenas. The NBA said the system was designed to track the ball and player movement in three dimensions at sub-second latency. The league also named out-of-bounds and goaltending as possible future use cases for automated officiating. Silver referred to out-of-bounds calls during the ESPN appearance. Automated officiating systems are used in defined call categories in other sports. Tennis uses electronic line calling, while FIFA has used semi-automated offside technology. MLB is introducing an automated ******-and-strikes challenge system in 2026. “We’re going to move to a system like that where that whole category of calls will be automatic,” Silver said. He said the system would determine possession immediately, whether the ball belongs to the Lakers, Knicks, Thunder, Spurs, or another team. Silver said the system would reduce the need for challenges on those calls. Coach’s Challenge covers out-of-bounds reviews Under current NBA rules, a Coach’s Challenge is the only way to trigger replay review of an out-of-bounds violation at any point during a game. Each team starts with one challenge and receives a second only if the first challenge is successful. The NBA also expanded the Coach’s Challenge rule for the 2024–25 season. The change allows officials to review whether a foul should have been called during certain out-of-bounds reviews. Silver said the technology would allow games to continue without stoppages for that category of decision. “It’ll be instantaneous, it’ll be automatic. Just play on,” he said. The NBA has already expanded its use of replay review and centralised officiating support. The league operates a Replay Center in Secaucus, New Jersey. According to the NBA, all 30 arenas are connected to the facility, which has 94 HD monitors, 23 workstations, and supports reviews across 15 instant replay triggers. Referees remain responsible for fouls Silver said referees would remain responsible for calls that require judgment, especially those involving contact and fouls. He said contact occurs on many plays, but officials still need to decide whether the contact affected a player’s movement or ability to continue the play. “There’s often contact on every play,” Silver said. “It doesn’t mean there’s a foul.” Silver did not give a specific timeline for introducing the automated system. He said the league expects to move in that direction “fairly quickly.” (Photo by JC Gellidon) See also: Autonomous AI systems test governance in physical environments Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post NBA plans AI system for automatic out-of-bounds calls appeared first on AI News. View the full article

Google is folding Display Ads into its AI-powered Demand Gen platform, marking the end of a long-standing digital advertising model. The Google Display Network (GDN) has been a staple of the open internet for almost twenty years. Marketers previously relied on its predictable framework to target placements, bid on audiences, and A/B test static creative across news sites and blogs. That familiar setup is changing and requires marketing teams to move away from manual campaign controls and rely on Google’s AI. Google describes this change as a natural progression and presents it as a method for advertisers to reach visual platforms like YouTube, Discover, and Gmail through one consolidated campaign. Traditional banner ads are facing increased competition from the full-screen video formats of platforms like TikTok and Instagram. In response, Google’s Demand Gen uses an automated system to generate and develop customer interest before a search query is ever entered. Demand Gen functions differently from the traditional GDN. Instead of having advertisers select specific websites or adjust audience segments, the platform requires business goals and a collection of creative assets. Marketers upload images, video clips, and headlines, which Google’s AI then tests in various combinations. The system serves these as in-stream video ads, YouTube Shorts, or interactive Discover posts, using predictive models to determine format, placement, and audience. This transition requires changes to creative production. Demand Gen relies on a continuous supply of diverse, format-agnostic content. Creative teams are now tasked with providing the raw assets that Google’s AI assembles dynamically, shifting the traditional agency workflow toward higher-volume content creation. Trading granularity for automation Google is betting that machine learning will beat human intuition at scale, effectively forcing the industry’s hand. Consolidating Display into this AI-centric model removes the temptation for teams to cling to manual methods. Advertisers must adopt the AI-first approach or risk losing visibility on valuable digital real estate. Long-standing metrics like click-through rate (CTR) and cost-per-click (CPC) are now losing much of their meaning. Judging the success of a single creative or placement becomes nearly impossible when an AI optimises for conversions or brand lift simultaneously across multiple formats and platforms. Instead, reporting must elevate to track broader business outcomes: customer acquisition cost, return on ad spend, and influence on the overall purchase journey. This requires tighter integration between advertising platforms and a company’s core business intelligence systems. Without accurate, real-time conversion data, the AI flies blind. For many enterprises, this dependency exposes critical weaknesses in their data infrastructure. A multi-million-pound Demand Gen budget could easily hinge on the quality of a single API connection to a CRM or e-commerce backend that are often built for entirely different purposes. Meta pushes a similar agenda with its Advantage+ campaigns, leveraging AI to automate targeting, creative, and placement across its ecosystem. The industry is clearly shifting from a model of renting ad space to one of commissioning AI agents to hunt down customers. Marketing leaders no longer have a choice about ceding control to AI; the focus is on how they adapt their teams, technology, and strategy. See also: Musk and Zuckerberg convinced Trump to scrap AI executive order Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events including the Cyber Security & Cloud Expo. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Google folds Display Ads into AI-first Demand Gen platform appeared first on AI News. View the full article

The foreign exchange market is really moving away from pure intuition and toward a space shaped by speed, data and precision. By using automated systems in your routine, you can approach volatility with a level of discipline that manual trading often struggles to maintain. Every entry and exit can be based on clear rules, not the emotional swings that frequently influence human decisions. Currency markets really move quickly, often faster than any person can react. With advanced digital tools, you can process large volumes of economic data and news in multiple time zones without dealing with fatigue. These systems work continuously, scanning for patterns and pricing gaps that many traders would miss manually. This change toward intelligent automation has changed how people participate in one of the world’s most liquid and demanding markets. Used properly, these tools can reduce emotional bias and turn market noise into a more structured trading process. Navigating the 24-hour global cycle Forex trading really demands a continuous focus, which is impossible for an individual. While you are sleeping, there might be market movements driven by the London or Tokyo sessions. By automating your system, you will cover the time outside your working hours, ensuring you do not miss opportunities due to sleep. The automated system will be active throughout the week and will monitor price action overnight, acting on preset parameters without you needing to be in front of the computer all the time. In addition to remaining active 24/7, automated trading follows your logic. Every action you want your software to perform will be carried out without delay, in accordance with your parameters. It really allows you to enter and exit positions instantly, increasing the likelihood of success. Efficiency and strategy validation With modern trading software, you very much have an opportunity to try out your theories by applying them to historical data without having to put real money at risk. Backtesting is a technique that very much demonstrates a trading strategy’s performance under various market conditions and lets you correct its flaws, fine-tune your entry/exit criteria and increase your confidence in the system before launching operations. The analysis of past trading history will also provide you with valuable insights into drawdowns, consistency and compatibility with your risk tolerance. Risk Management: You can set up stop-loss and take-profit limits for all transactions. Speed of Execution: Your orders can be placed instantly whenever specific criteria are fulfilled. Data Processing: You can analyse multiple pairs of currencies simultaneously and detect possible trading signals without spending all your time looking at the screen. The logic of automated market engagement Contemporary markets are complex and most methods rely on identifying patterns in large datasets. This is what AI bots for forex trading are all about: they allow filtering information and pointing out setups with high probability. In other words, it is not necessary to act on gut feelings. Market conditions can be observed through measurable statistics, rules and signals. Moreover, structured approaches can promote consistency. Whether there is a strong trend in the market or its movements are range-bound, the software acts according to your predefined conditions. This consistency can be for people who give up on their strategy after several adverse events. Removing some emotional biases can make a difference in sticking to the strategy. Mastering your trading psychology Often, one of the largest obstacles to success is the trading mentality itself, where greed keeps you in a trade for too long while fear makes you exit before it becomes profitable. Automated execution lets you eliminate those tendencies because it follows your plan, no matter what. No matter how volatile the market becomes, there is no doubt, since everything has been decided in advance. The distinction between the creation of a strategy and its execution means you do not have to make your trading decisions on the fly and based solely on what happens at the moment. You can take the time to review past performance and analyze trends to develop your strategies while leaving the technical analysis to automated software. The evolution of personal finance technology Currency trading is becoming increasingly data-focused. As algorithms improve, the tools available to traders are also becoming more adaptive. Some systems now aim to respond to changing volatility or changing market conditions not relying on one static model. Predictive analytics and machine learning are also influencing how opportunities are identified. Access to this kind of technology is not limited to large institutions. Retail traders can now use tools once associated mainly with professional desks. That has narrowed the gap between individuals and larger players, especially in areas like execution speed and systematic discipline. No system can guarantee results, but an evidence-based, structured approach may help traders participate in the forex markets with greater consistency and control. The post Exploring the Benefits of AI Bots for Forex Trading in Forex Markets appeared first on AI News. View the full article

Autonomous AI systems are beginning to move beyond software environments and into warehouses, delivery networks, and public spaces. The development is drawing attention to whether current AI rules cover systems that operate in physical environments. Most existing AI governance frameworks have focused on online harms and model outputs, including bias, misinformation, and harmful content. Embodied AI systems carry risks in physical environments, where failures can affect infrastructure, property, or human safety. Singapore’s Infocomm Media Development Authority published version 1.5 of its Model AI Governance Framework for Agentic AI on May 20. The framework sets out guidance for organisations deploying AI agents that can plan, make decisions, and take actions across multiple steps to complete user-defined goals. The framework says agents can interact with tools, external systems, and other agents, including systems that update databases, write files, control devices, or perform transactions. It lists access controls, monitoring, and human approval among governance measures for deployment. AI moves into physical systems At an AI summit in Singapore last week, discussions around robotics and embodied AI focused on operational safety issues more commonly associated with aviation, industrial systems, and critical infrastructure oversight than conventional software regulation. Speakers also discussed whether autonomous systems can operate safely and reliably in unpredictable real-world environments over extended periods. Dr. Ya-Qin Zhang, founding dean of the Institute for AI Industry Research at Tsinghua University, said embodied AI systems amplify risks already associated with autonomous software. He said failures can directly affect transport systems, drones, logistics networks, and critical infrastructure. “Any risk in the digital domain will be amplified in the physical domain, and the physical domain will have a physical consequence,” Zhang told MLex on the sidelines of the summit. He added that vehicles, drones, smart grids, and other infrastructure could become exposed as AI systems are embedded more deeply into physical operations. Speakers discussed reliability, operational monitoring, and post-deployment assurance as governance concerns. Summit discussions pointed to deployment-based governance models built around simulation, telemetry, and iterative testing, rather than one-time certification alone. IMDA’s framework also recommends gradual rollouts, continuous monitoring, and further testing after deployment. It says agents interact dynamically with their environment and not all risks can be anticipated before release. Monitoring becomes a deployment issue Grab, which is piloting autonomous vehicles and delivery robots in Singapore’s Punggol district, said deployment governance depends heavily on simulation, testing, and continuous monitoring. “We do a lot of simulation, we do a lot of testing in closed courses and open courses in order to make sure our robots are reliable,” Suthen Thomas Paradatheth, Grab’s chief technology officer, said during one of the summit panels. “Before we scale to hundreds of robots, we make sure we crack it first in simulation and with a few robots,” he added. Grab also pointed to monitoring systems designed to track robot performance and detect unexpected failures after deployment. “There’s a long tail of issues that could emerge,” Paradatheth said. The IMDA framework says organisations should assess agentic AI use cases based on data access, external system access, autonomy, and task complexity. It also points to the scope and reversibility of agent actions, third-party involvement, and overall system complexity. It also recommends limiting agent access to tools and systems, applying least-privilege permissions, and defining standard operating procedures for agent workflows. Organisations should also set mechanisms to take agents offline when they malfunction. Accountability spreads across more actors MLex reported that embodied AI systems can involve several parties across development, manufacturing, and deployment. These include AI developers, robotics manufacturers, semiconductor suppliers, and infrastructure operators. MLex also noted that responsibility can be harder to assign when systems continue adapting after deployment through software updates, telemetry, and operational data. IMDA says organisations and humans remain accountable for agent actions, even when agents operate autonomously. The framework calls for clear responsibility across the agentic AI value chain, from model and platform providers to deployers, tooling providers, and end users. Applied Materials said large-scale robotics deployment is also tied to semiconductor economics and systems integration. Om Nalamasu, the company’s chief technology officer, said robotics systems will depend on better sensors, energy efficiency, advanced packaging, and computing architectures. Nalamasu said robotics systems would require purpose-built designs adapted to specific industrial ecosystems rather than a single solution for all environments. Zhao Yuli, chief strategy officer of ******** robotics startup Galbot, said Beijing is prioritising deployment scale and industrial commercialisation through government-backed testbeds, industrial partnerships, and long-term funding initiatives. Galbot has deployed humanoid robotics systems in retail, warehouse, and pharmaceutical operations in China. These include autonomous stores that operate around the clock. Zhao said semi-structured industrial environments are likely to become an early commercialisation path because they offer more controllable operating conditions. Japan is placing more focus on standards-setting, robotics datasets, and safety governance. Professor Yutaka Matsuo of the University of Tokyo’s Graduate School of Engineering pointed to an “AI Association” project aimed at collecting 100,000 hours of robotics data to support robotic foundation models. Matsuo also referred to Japan’s AI Safety Institute and the Hiroshima AI Process as part of broader efforts to develop governance standards for embodied AI systems with Singapore and other Asian countries. Singapore sets out agent controls Singapore’s framework sets out four governance areas for agentic AI. These cover upfront risk assessment, human accountability, technical controls, and end-user responsibility. The framework describes them as an iterative process rather than a one-time assessment. The framework says human oversight has to be adapted for agentic systems because continuous review of all workflows becomes impractical at scale. It recommends human approval at significant checkpoints, including high-stakes actions, irreversible actions, and outlier behaviour. IMDA also identifies automation bias and alert fatigue as risks when humans supervise capable agents. It recommends auditing oversight through indicators such as human override rates and response times, and using automated real-time monitoring to flag unexpected behaviour. The framework says users should be told what actions an agent can take, what data it can access, and what responsibilities remain with the user. It also recommends employee training on human-agent interaction, oversight, and the professional skills needed to assess agent outputs. Companies test AI in regulated workflows JPMorgan is implementing AI tools across its global investment banking business, Paul Uren, the bank’s Asia Pacific head of investment banking, told Reuters. The bank said the tools help bankers access more information and synthesise it with internal systems. They are also being used to prepare content and support client engagement. JPMorgan CEO Jamie Dimon told Bloomberg News that the bank would hire more AI specialists and fewer traditional bankers. Reuters reported that global banks are increasing AI investment, reshaping workforces, and changing job roles. The bank is also among selected organisations permitted by Anthropic to use its Mythos cybersecurity model under a controlled initiative known as Project Glasswing. According to Anthropic, Mythos can detect old vulnerabilities in browsers, infrastructure, and software. Reuters reported that Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley also have access to, or are testing, Mythos, citing sources and company executives. IMDA’s framework includes a case study from OCBC Bank of Singapore on source-of-wealth analysis. The system parses income-related documents and drafts a source-of-wealth memo. It does not make credit, onboarding, or risk decisions autonomously. In that case, the workflow is limited to task-level autonomy and operates only when triggered by predefined workflows. Human review is required at critical decision points, and final validation remains with designated reviewers. Robots move into industrial use In Japan, one-third of companies are already using or considering AI-powered robots, according to a Reuters survey conducted by Nikkei Research from May 1 to May 15. The survey contacted 492 companies, with 220 responding on the condition of anonymity. About 4% of respondents said they already use AI robots, 5% plan to deploy them, and 25% are considering doing so. The remaining 66% said they had no such plans. Transportation equipment manufacturers were the most active group in the survey, with 80% already using AI robots or considering deployment. By comparison, 94% of wholesale sector respondents said they had no plans to deploy AI robots. Among companies using, planning to use, or considering AI robots, 71% selected manufacturing as a use case. Another 19% selected dangerous tasks, while 11% selected customer-facing services. The Japanese government expects AI robots to help address the country’s chronic labour shortage and support its position in industrial robotics. Japan is home to robotics companies including Fanuc, Yaskawa Electric, and Kawasaki Heavy Industries, but faces competition from China and the United States in AI-enabled robotics. Retail agents expand beyond search Walmart has outlined plans to use agentic AI across shopping, employee, supplier, and developer workflows. In July 2025, the retailer announced plans for four AI-powered “super agents.” They are designed for shoppers, store employees, suppliers and sellers, and software developers. Walmart said these agents would become the main entry point for AI interactions across those groups. One of the tools, Sparky, is already available in Walmart’s app as a generative AI-powered shopping assistant. Hari Vasudev, Walmart’s US chief technology officer, said its expanded version would be able to reorder items and plan events. It would also use computer vision to suggest recipes based on the contents of a shopper’s fridge. Walmart is also developing an Associate super agent for store workers and corporate staff. A separate Marty agent is being built for sellers, suppliers, and advertisers. The retailer is also working on a Developer super agent for testing, building, and launching future AI tools. The company declined to say whether the agents would replace jobs. Dave Glick, senior vice president of enterprise business systems, said the tools would create new jobs, without giving further details. (Photo by Growtika) See also: OpenAI opens Singapore AI lab as IMDA updates AI framework Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Autonomous AI systems test governance in physical environments appeared first on AI News. View the full article

OpenAI will open its first Applied AI Lab outside the US in Singapore. The lab is part of a new partnership with the Ministry of Digital Development and Information. The initiative, called OpenAI for Singapore, was announced at the ATx Summit and is backed by a commitment of more than S$300 million. The lab will create more than 200 Singapore-based technical roles over the next few years. OpenAI said Singapore will also become one of its global hubs for forward-deployed engineers who will work with organisations on AI deployment. OpenAI said the lab’s work will be aligned with Singapore’s AI Mission priorities which include public service, finance, and digital infrastructure. Focus on deployment and talent The company will work with government agencies and local partners on education and workforce programmes within the Ministry of Education and GovTech. OpenAI also plans to support educators through a Singapore chapter of the OpenAI Academy, participate in the National AI Impact Programme, and run Codex for Teachers hackathons. The partnership includes plans to work with local partners on accelerator programmes for AI-native startups in the form of workshops for micro-entrepreneurs and small businesses, covering how founders and SMEs can use AI in operations and customer service. Chng Kai Fong, Permanent Secretary for Digital Development and Information, said Singapore’s response to AI includes growing new sectors, anchoring global frontier companies, and equipping workers with relevant skills. Singapore updates agentic AI framework Singapore has also updated its governance framework for agentic AI, which was launched by the Infocomm Media Development Authority at the World Economic Forum in January 2026. The framework builds on Singapore’s earlier Model AI Governance Framework for AI, introduced in 2020, and gives organisations guidance on the responsible deployment of AI agents, including measures to reduce the risks inherent in agentic AI. IMDA has now updated the framework after seeking feedback and case studies from the industry, with the revised version following input from more than 60 organisations, including AWS, DBS, Google, and Salesforce. The update adds guidance on risks linked to multi-agent systems, third-party agents, automation bias, and human accountability. The framework now includes more than ten case studies showing how organisations have applied its recommendations. The case studies were contributed by Singaporean and international organisations, including Ant International, City Developments Limited, Cyber Sierra, Dayos, Google, Knovel, OCBC, PwC, Stability Solutions, Tencent, Terminal 3, Workday, X0PA, and GovTech Singapore. Case studies show governance controls One case study focuses on Dayos, a Singapore-headquartered enterprise AI automation company with operations in the US. Dayos built an AI-powered ticketing agent that handles internal IT requests. The agent can resolve some requests automatically and route requests to a human when needed. Dayos used tiered risk levels to determine what actions the agent could take. Low-risk and reversible actions, like password resets, could be automated and audited biweekly, while moderate-risk actions required human approval before execution. Higher-risk actions, like permission changes with limited reversibility, were excluded from the agent’s authority. Tencent contributed a case study on CodeBuddy, an agentic AI coding system developed by Tencent Cloud. CodeBuddy can plan, write, and deploy code through natural language instructions and can access filesystems, terminal commands, external APIs, and MCP tools. CodeBuddy uses preset defaults and configurable permissions. Human approval is required for actions like editing files, running shell commands, making network requests, or using external tools. The system explains complex commands in plain language before users approve them. Suspicious commands still require human approval, even if similar commands had been pre-approved. GovTech Singapore’s case study covers the rollout of agentic coding assistants in government. The first phase was limited to GovTech employees, did not allow external tools, and was restricted to low-risk systems. GovTech developed central logging and a framework for connecting approved external tools. The agency also tested the system against potential attacks. (Photo by Mike Enerio) See also: GPT-5.5 is OpenAI’s most capable agentic AI model yet Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post OpenAI opens Singapore AI lab as IMDA updates AI framework appeared first on AI News. View the full article

Every major economy is staring at the same problem right now. Artificial intelligence is consuming electricity at a pace that grids were never designed to handle. In the US, capacity market prices in PJM, the country’s largest grid operator, have risen more than tenfold in two years, with data-centre growth identified as a primary driver. In Europe, utilities are scrambling to upgrade transmission infrastructure fast enough to keep pace with hyperscalers’ demand. The International Energy Agency (IEA) projects global data-centre electricity consumption could approach 1,000 TWh by the end of this decade. Renewable energy is largely there, but the ability to coordinate it, through AI energy grid mapping at national scales, is what most countries still lack. But China just built it. A study published in Nature this week by researchers from Peking University and Alibaba Group’s DAMO Academy has produced something that no country has managed before: a complete, high-resolution, AI-generated inventory of an entire nation’s wind and solar infrastructure, with the analytical framework to coordinate it as a unified system. Using a deep-learning model trained on sub-metre satellite imagery, the team identified China’s 319,972 solar photovoltaic facilities and 91,609 wind turbines, processing 7.56 terabytes of imagery to do so. AI energy grid mapping Prior research into solar-wind complementarity – the idea that two sources can offset each other’s variability in time and geography – has largely relied on hypothetical or modelled deployment scenarios. How complementarity manifests under real-world infrastructure, and how it shapes system-level integration outcomes, has until now remained unclear. The researchers show that solar-wind complementarity substantially reduces generation variability, with effectiveness increasing as the geographic scope of pairing expands. In practical terms, the further apart the facilities being coordinated are, the more reliably they achieve balance. A cloud that covers solar farms in Gansu does not darken wind corridors in Inner Mongolia, for example. The study’s findings point to a structural inefficiency in how China currently manages its grid: coordination happens at a provincial rather than national level. Transitioning to a unified national scale, the researchers argue, would make it easier to pair complementary energy sources, stabilise the grid, and avoid curtailment – the wasting of generated renewable power that has long been one of China’s most costly clean-energy problems. Liu Yu, a professor at Peking University’s School of Earth and Space Sciences, described the inventory as allowing China to see its new-energy landscape from a “God’s-eye view,” a phrase that carries more operational weight than it might first suggest. Grid operators cannot optimise what they are not aware of – until now. China is in the middle of an AI-driven electricity demand surge that is straining its grid. The rapid proliferation of data services and massive computing facilities have pushed the sector’s power consumption up 44% year-on-year in the first quarter of 2026, reaching 22.9 billion kilowatt-hours, according to the China Electricity Council. That is an extraordinary rate of growth for a sector whose demand was already great. This has accelerated data-centre expansion in China’s northern and western provinces, where land is cheaper, wind and solar resources are more available, with commensurately lower electricity prices. The provinces being targeted for new data centres are the same regions with the highest solar-wind complementarity. Behind the model The technical achievement behind this is worth understanding in its own right. DAMO’s deep-learning model was trained to identify solar photovoltaic facilities and wind turbines from sub-metre resolution satellite imagery, a task complicated by the sheer diversity of installation types, terrain conditions, and image quality. The resulting dataset covers installations in 1,915 ******** counties, spanning everything from rooftop panels in coastal cities to utility-scale wind farms on the Mongolian plateau. Processing 7.56 terabytes of imagery to produce a nationally consistent, county-level inventory is a demonstration of what large-scale geospatial AI can do when applied to infrastructure problems, and a template that other countries could, in principle, replicate. China’s clean energy sector generated an estimated 15.4 trillion yuan (US$2.26 trillion) in economic output last year, equivalent to Brazil’s entire GDP, according to the Finland-based Centre for Research on Energy and Clean Air. Managing an asset base of that scale without a national-level visibility tool was always going to be a limiting factor, a limit that’s now gone. The study’s dataset and code have been made publicly available via Zenodo. (Photo by Luo Lei) See also: Inside China’s push to apply AI in its energy system Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post China’s AI just mapped its entire renewable energy grid. Here’s why the rest of the world should pay attention appeared first on AI News. View the full article

The ceremony was scheduled. The CEOs were on the guest list. And then it wasn’t happening. On Thursday, US President Donald Trump scrapped a planned AI executive order, which had already been delayed multiple times, citing concerns that it might erode America’s competitive edge over China. “We’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that lead,” Trump told reporters in the Oval Office. What he did not say was that the order had been effectively killed by the very industry it was meant to oversee. Lobbied out in one night According to Semafor, which first reported the backstory, the White House’s plans were halted after Elon Musk of xAI, Meta CEO Mark Zuckerberg, and venture capitalist David Sacks, who, until recently, was Trump’s AI and cryptocurrency tsar, all spoke directly with Trump between Wednesday night and Thursday morning. The argument that landed, according to US media, citing sources, was an appeal to the “accelerationist” faction in the administration, including officials at the National Economic Council and staffers in the Vice President’s office. The order itself was not a sweeping regulatory framework. It would have established a voluntary mechanism for AI developers to engage with federal agencies and submit advanced models for security review up to 90 days before their public release. No licensing regime. No mandatory hold periods. Voluntary. That was apparently still too much. Trump said he postponed it “because I didn’t like certain aspects of it,” declining to specify which ones. He added that he worried it “could have been a blocker,” a telling phrase from a president who has otherwise positioned AI as a jobs and national security priority. A vacuum with consequences The US has yet to pass comprehensive AI legislation. What governance architecture exists has been assembled piecemeal, through executive orders, agency guidance, and voluntary agreements. Earlier this month, the federal Centre for AI Standards and Innovation announced evaluation agreements with Google DeepMind, Microsoft, and xAI, allowing the government to assess models before public availability. That programme continues regardless of Thursday’s non-signing. But the broader picture is one of regulatory drift. In early March, the Trump administration released a National AI Legislative Framework urging Congress to preempt state-level AI laws that “impose undue burdens,” arguing for a single national standard over what it called “fifty discordant ones.” Congress has not acted on it. The contrast with China is sharp and increasingly difficult to ignore. Beijing’s State Council issued a 2026 legislative work plan in May outlining plans to accelerate comprehensive AI legislation, deploying language on AI governance in formal planning documents for the first time. The National People’s Congress has listed AI legislation for review for the third consecutive year. In April, Beijing issued new rules requiring AI companies to establish internal ethics review committees. China is writing rules. Washington is cancelling ceremonies. Who shapes US AI policy Thursday’s episode clarified something implicit for months: in the current administration, the effective veto on AI regulation sits with a small group of industry principals who have direct access to the president. Musk, whose xAI is a direct competitor to OpenAI and Anthropic, has a structural interest in keeping the regulatory field open. Zuckerberg’s Meta has similarly positioned itself as a champion of open-source AI development. Sacks, despite having formally left his White House advisory role in March, evidently retains enough influence to shape executive action. Separately, Semafor reports that OpenAI has secured White House backing for a parallel effort to push AI regulations at the state level, an interesting manoeuvre given that Trump’s earlier executive order threatened states that enacted AI laws the administration disliked. That the administration appears to be simultaneously discouraging state regulation and endorsing OpenAI’s state-level strategy suggests the policy coherence problem runs deeper than one postponed signing. The China frame does real work, but in both directions Trump’s stated reason for pulling back, protecting the US lead over China, is the same logic that has driven every major AI policy decision since he returned to office, from the H200 export licence framework to the Stargate infrastructure programme. It is also the logic that China is watching closely. At the Trump-Xi summit in Beijing earlier this month, the two leaders agreed to launch an intergovernmental dialogue on AI, according to the ******** Foreign Ministry. Beijing will have noted that Washington’s internal debate about even voluntary AI oversight was resolved not by policymakers, but by the companies that stand to profit most from the absence of guardrails. In a report by the South China Morning Post, Lizzi C. Lee, a fellow at the Asia Society Policy Institute’s Centre for China Analysis, noted that both the US and China are grappling with the same underlying question: where should the regulatory frontier sit for frontier AI, particularly as models become more capable of autonomous action and more relevant to cybersecurity. “I think a separate, potentially more important race is on governance and safety: not about who has the most advanced models, but who can govern powerful AI without choking off innovation,” she said. The same report highlighted what Kyle Chan at the Brookings Institution put it more simply: “AI safety and regulation can be done in a way that doesn’t compromise innovation.” Neither argument was enough on Thursday. Whether it becomes enough next time, assuming there is a next time, remains unclear. (Photo by White House) See also: The US-China AI gap closes amid responsible AI concerns Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and is co-located with other leading technology events, click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Musk and Zuckerberg convinced Trump to scrap AI executive order appeared first on AI News. View the full article

The Nvidia Vera chip is rarely the headline when earnings beat estimates, but it should be. When Nvidia reported Q1 revenue of US$81.62 billion on Wednesday, beating analyst estimates of US$78.86 billion, and guided Q2 at US$91 billion–well above Wall Street’s US$86.84 billion forecast–the numbers did what Nvidia numbers always do: dominate the room. But buried in CEO Jensen Huang’s conference call with analysts was something more strategically interesting than another quarterly beat. Huang told analysts that Nvidia’s new Vera central processors unlock access to a US$200 billion market, one that sits entirely outside the US$1 trillion the company has already forecast from its Blackwell and Rubin AI GPU lineup between 2025 and 2027. He expects Vera chip revenue to hit US$20 billion by the end of this fiscal year. “I expect (Vera) to be the second largest” sales contributor, Huang said during the call. That’s not a footnote. That’s a second front. The Vera chip and the inference pivot The reason Nvidia needs a second front is straightforward: its biggest customers are building their own. Google, Amazon, and Microsoft–collectively expected to pour more than US$700 billion into AI infrastructure this year, up sharply from around US$400 billion in 2025, are simultaneously pouring funds into custom silicon to run AI models. Intel and AMD are also touting CPUs as a credible play for inference workloads. The narrative in the chip industry has shifted from who can train the biggest model to who can serve it cheapest and fastest. Inference is where Nvidia’s GPU dominance is most exposed. Training large models is still firmly Nvidia territory, but inference, generating answers at scale, in real time, is increasingly where custom chips from Google’s TPU line, Amazon’s Trainium and others are making their case. Nvidia’s answer is Vera. The chip, developed in part using technology from Groq, a startup specialising in inference that Nvidia licensed in a deal reportedly worth around US$17 billion, targets exactly this workload. The full Vera Rubin platform, which combines the Vera CPU with Rubin GPUs, is set to launch later this year. Supply is already the constraint Huang was candid about one problem: supply. “My sense is that we’ll be supply-constrained through the entire life of Vera Rubin,” he said on the call. It’s a telling admission for a product Nvidia is positioning as a major growth pillar. To get ahead of disruptions, Nvidia is spending heavily on the supply chain. The company disclosed that its supply commitments rose to US$119 billion in Q1, up from US$95.2 billion the previous quarter, a significant jump that reflects both confidence in demand and anxiety about a global memory chip crunch. Nvidia also announced a US$80 billion share repurchase programme and raised its quarterly cash dividend to 25 cents per share, from 1 cent, moves that signal financial confidence even as Huang warned of tightening supply. The question investors are asking Despite the beats, Nvidia shares fell 1.6% in extended trading after the results. eMarketer analyst Jacob Bourne captured the mood: “Nvidia delivered another beat, but at this point that’s essentially priced in as it keeps beating quarter after quarter. The lingering question is whether it can convince investors the AI buildout has durability into 2027 and 2028, especially as the narrative shifts toward inference workloads and competing silicon from Google, Amazon, AMD, and Intel.” Huang pushed back with numbers of his own. He pointed to a growing sub-segment of AI-specific cloud customers whose spend is now roughly equal to the hyperscalers, but growing faster quarter-over-quarter. “We should be growing faster than hyperscale capex,” he said. The Vera chip is central to that argument. Whether the supply chain cooperates is a different question entirely. (Image source: Nvidia’s Newsroom) See Also: The Nvidia H200 China deal survived the Trump-Xi summit–just not in the way anyone expected Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Nvidia’s Vera chip is the US$200 billion bet Jensen Huang doesn’t want you to overlook appeared first on AI News. View the full article

Alibaba has unveiled a new AI processor built specifically for AI agents, pairing the chip announcement with a multi-year silicon roadmap and a new large language model, signalling that the company is building an integrated AI stack, not just filling a gap left by US export controls. The Zhenwu M890, developed by Alibaba’s semiconductor subsidiary T-Head, delivers three times the performance of its predecessor, the Zhenwu 810E, according to the company, as per Reuters report. But the performance jump is less notable than the architectural intent behind the chip: the M890 is purpose-built for AI agents, where software systems must retain long stretches of context, coordinate with other models in real time, and execute complex multi-step tasks with limited human intervention. Those demands, heavy on memory bandwidth and inter-model communication, are meaningfully different from what standard inference chips are optimised for. The difference matters because it tells you something about where Alibaba thinks AI compute is heading. The company isn’t designing around today’s dominant use case; it’s building for the workload profile it expects to define enterprise AI over the next several years. Built for AI agents, not just inference More significant than the chip itself is the roadmap Alibaba put alongside it. The M890 will be followed by the V900 in the third quarter of 2027, expected to deliver another roughly threefold performance gain, followed by the J900 in the third quarter of 2028. That’s a deliberate, sustained cadence of in-house silicon upgrades that mirrors the kind of tick-tock product cycles Nvidia has used to maintain its lead in AI accelerators. The parallel to Huawei is worth noting. Huawei laid out a similar chip roadmap for its Ascend line last year, and both announcements reflect the same underlying reality: ******** technology companies have concluded that depending on foreign silicon, even in scenarios where export restrictions might ease, is a structural risk they cannot accept. The response has been to treat semiconductor development as a long-term capability-building exercise rather than a procurement problem. Alibaba’s commitment to that exercise is not shallow. The company pledged more than 380 billion yuan, roughly US$53 billion, on cloud and AI infrastructure over three years last year, its largest-ever investment commitment to the sector. The M890 and its successors are downstream of that spending. Traction that predates the announcement T-Head said it has shipped more than 560,000 Zhenwu units to date, with over 400 external customers across 20 industries deploying the chips, including automakers and financial services firms. That is a material production footprint, not lab hardware, and it provides Alibaba with real-world deployment data at scale ahead of the M890’s rollout. The new chip will be available to ******** enterprise customers through Alibaba Cloud’s domestic model platform, Bailian, packaged inside the Panjiu AL128, a server system that stacks 128 M890 accelerators into a single rack. The software side of the stack Alongside the hardware, Alibaba announced Qwen 3.7-Max, the latest version of its flagship large language model, described as engineered for advanced coding and long-running agent tasks. The company said the model can operate continuously for up to 35 hours without performance degradation, a capability specification that only makes sense if you are designing for extended autonomous operation. The timing is deliberate. Releasing a chip and a model optimised for the same workload class on the same day is a platform play. Alibaba is building a closed loop: its own silicon in T-Head, its own model in Qwen, its own cloud delivery in Bailian. Each component reinforces the others, and the combined stack is designed to reduce enterprise customers’ dependence on any external vendor. Half a million chips shipped. A successor arriving in 2027, another in 2028. T-Head is not hedging. At some point, building around US export controls stops being a workaround and starts being a strategy. Alibaba appears to have crossed that line. (Image source: The White House) See Also: Alibaba Qwen is challenging proprietary AI model economics Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Alibaba is designing AI chips around agents, and that changes what the race is actually about appeared first on AI News. View the full article

The AI and Big Data programme on day two of TechEx North America referred at least once to the “AI graveyard,” meaning the large number of pilots that never become durable systems. That phrase set the tone. The question was proof. The Enterprise AI Implementation, ROI and Adoption track dealt with the hard middle of AI work. Its sessions covered stalled pilots, agentic AI for business impact, the move from experimentation to impact, the decision to buy or build, and durable ROI and autonomous decisioning. A system has to be adopted, governed and measured before it deserved to be called successful. The session on the AI graveyard was useful because it named the failure pattern. Many companies have enough budget to start AI experiments and enough executive attention to publicise them. Fewer have the data quality, process design, operating authority, and risk control to keep them going. A day-two session on moving beyond copilots towards agentic AI framed the issue as business impact not novelty. Copilots have been useful as individual productivity tools, but their value is often hard to measure. Agents promise a closer connection to business process, yet they also increase the need for boundaries. An agent that can act in systems has to be evaluated by the quality of the action. That point linked directly with the Future of AI track. Its opening theme, trust as a competitive advantage, was a useful counterweight to speed. The programme dealt with transparency, governance, regulation, banking analytics, and risk. It also included material from Hex on data agent, with evaluation and governance built in. Agentic AI will not mature in enterprise settings if evaluation remains informal. Governance appeared in several forms. There was cross-functional governance, which reflects the reality that AI risk does not belong to legal, security or engineering. There was governance in the data layer, where trust depends on lineage and quality. There was governance around agent personas and risk stacks, where companies need to understand what an AI agent is permitted to know and do. The banking session gave the theme a sectoral focus, since financial services have less room for undefinedassurances about automation. Digital Transformation Week carried the same day-two pressure into business delivery. The programme was built around real use cases, business impact, ROI, AI agents built on APIs, change readiness, government service transformation, city innovation and the conversion of data into financial value. The change-readiness material was especially important. AI fails because staff do not change routines, managers do not alter incentives, or the data needed for daily use never appears in the right place. Sessions involving the DMV and the City of San Jose placed AI and transformation inside government service. In government, the measure of quality includes reliability, access, explainability and public trust. The Dow material on turning data into dollars sat at the commercial end of the same argument. In both cases, value depends on connecting data work with an accountable outcome. The Cyber Security and Cloud Expo day-two programme expanded on risk. Its cloud-first enterprise track dealt with AI-led threats, cloud security, the “GenAI velocity gap,” threat intelligence, identity security and AI governance. The cyber programme treated AI as a force that changes attack and defence alike. It can help automate defensive work, but it can also accelerate misuse, widen leakage routes, and increase the strain on existing controls. The phrase “velocity gap” was used several times during day two. Business units are adopting generative AI faster than many security teams can oversee it: the tools arrive first, policy and monitoring arrive later. The sessions on jailbreaking and data leaks made the point more concretely. If staff place sensitive material into unsanctioned tools, or if approved AI systems are poorly bounded, cloud security and data governance become one and the same. Zero trust was presented as one answer, with a stronger interpretation of zero trust must now include AI systems, agents, and the data around them. Identity is not limited to human users, but services, agents and automated workflows require permission models as well. The cloud-first enterprise is therefore becoming a place where identity, data classification, AI governance, and threat detection are part of the same control mechanisms. (Image source: TechEx/TechForge) Want to learn more about AI and big data from industry leaders? Check out AI & Big Data Expo taking place in Amsterdam, California, and London. The comprehensive event is part of TechEx and co-located with other leading technology events. Click here for more information. AI News is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here. The post Proving the case on day two at TechEx North America appeared first on AI News. View the full article