Microsoft shifts Windows focus to AI agents as Copilot+ PCs fade into background

Windows is back on the Microsoft menu

I can't remember the last time Microsoft kicked off a Build keynote with Windows front and center, but that's exactly what CEO Satya Nadella did this week. Nadella didn't address the issues Microsoft is trying to fix in Windows 11 but chose to woo the audience with Microsoft's slick Surface RTX Spark Dev Kit instead, calling it a "dream machine." Nadella unveiled the new Surface hardware just days after Nvidia officially returned to Windows on Arm with its new RTX Spark chips. Both companies are talking up these chips as some kind of new beginning for PCs, and it's clear that RTX Spark will drive local AI workloads in a way that Microsoft's previous Copilot Plus PCs haven't yet managed. Build really drove home that message this week, with Windows positioned as an all-important part of Microsoft's AI agent efforts. Microsoft's original mission under Bill Gates was a computer on every desk and in every home, and Nadella reframed that as "unmetered intelligence on every desk and in every home" within a few minutes of his keynote beginning. It set the stage for Microsoft and Nvidia to position their new Windows PCs as a potential solution for costly, usage-based pricing of cloud-based AI models. As local compute grows in capability, there's a clear gap that Microsoft and Nvidia can fill with powerful hardware you actually own. "I think we, as Microsoft, have the responsibility for building the best possible AI stack that we can on [Windows], and obviously drive the best AI stack that we can in the cloud," says Windows chief Pavan Davuluri in an interview with Notepad. Davuluri thinks that Microsoft is in a good position to capitalize on hybrid compute, where chips like the RTX Spark will handle a lot of local workloads and intelligently hand off to the cloud when they need something more powerful. Nvidia CEO Jensen Huang is even more bullish about local AI compute. He wants to turn PCs into devices that work for you, eliminating that idle time when PCs are switched off or you're not using them. "In the future, if I need my laptop to do something, I just text it with WhatsApp," said Huang earlier this week. "You don't want to necessarily run everything in the cloud, because if you can run it locally, it's free." Nadella seems to agree. "The amount of compute that there is at the edge is astounding," he said during his Build keynote. "Every PC, if you sort of aggregate that, that's a lot of compute power." That power is really on display with Nvidia's new RTX Spark chips, which will come to a variety of creator-focused laptops and miniature PCs later this year. RTX Spark is capable of running a 120 billion parameter large language model locally, allowing many AI workloads to run without ever touching the cloud. That's an appealing concept during a continued AI money squeeze for developers and consumers. Microsoft is targeting its own Surface Laptop Ultra at developers and creators and pairing it with ongoing improvements to Windows 11 performance and developer-friendly additions. While Microsoft's deeper embrace of Linux utilities inside Windows this week didn't generate the same gleeful audience reaction as the Windows Terminal announcement in 2019, developers I've spoken to are excited by the Coreutils and WSL containers additions. The Surface Laptop Ultra has also been generating some buzz, particularly among developers and power users. Microsoft isn't quite positioning this as a mainstream premium laptop, but there's certainly room for it to appeal far beyond developers. "I think you'll see us do well when it comes to STEM applications, and CAD apps running on the platform, because they take advantage of the same characteristic patterns of high-performance compute," explains Davuluri. All of this renewed focus on Windows at Microsoft seemed impossible only six months ago. Davuluri responded to the pressure on Microsoft to improve Windows 11 by laying out a plan to focus on performance, reliability, and overall experiences in the OS just a couple of months ago. I got to see some of the performance improvements at Build this week, with side-by-side comparisons of the Start menu and taskbar loading faster. Microsoft is putting in a lot of effort to turn Windows 11 around and listen to feedback from a variety of users. But I've been wondering why Microsoft doesn't just jump to Windows 12. It seems easier to just admit defeat on Windows 11 and then position Windows 12 as the remedy. Microsoft has done this many times in the past, particularly with the releases of Windows 7 and Windows 10. "There are a lot of considerations when you think about the versioning of an operating system itself, and I think for us, a lot of the core proposition with Windows 11, or quite frankly, with Windows 12, or any label we use, has to do with end users and how they use the product and the workflow that they're in," says Davuluri. "I think we are more focused on having the product experience be better in the context they're using it, and that I think is the most important thing for us." While we might not be getting a Windows 12 anytime soon, I'm curious how this Windows exists in a world of AI agents. Microsoft has been clear that it sees Windows as a home for AI agents and workloads, but it also unveiled Project Solara this week, a new platform for agent-first devices. Microsoft demonstrated a smart employee key card that could run an agent capable of transcribing and recognizing real world objects, and it also showed a reference design for an Amazon Echo Show-like device with an AI agent. It's clear that Microsoft wants to offer up a platform for dedicated AI devices of the future The big surprise is that Project Solara devices are powered by a version of Android, not Windows. Despite this, Davuluri expects to see Project Solara running on Windows devices too. "We are not hard bound to a device specific operating system," says Davuluri. "You should imagine a world where Solara will be great on a bunch of platforms, including Windows, both Windows 11 locally and Windows 365 instances in the cloud." Whether the future of AI agents runs on Windows, Android, or something else may not ultimately matter right now. For the first time in years, Microsoft seems determined to make Windows central to that conversation either way. Build 2026 wasn't about fixing Windows' past problems, it was about convincing developers that Windows still has a significant role to play in AI's future. I'm always keen to hear from readers, so please drop a comment here, or you can reach me at [email protected] if you want to discuss anything else. If you've heard about any of Microsoft's secret projects, you can reach me via email at [email protected] or speak to me confidentially on the Signal messaging app, where I'm tomwarren.01. I'm also tomwarren on Telegram, if you'd prefer to chat there. Thanks for subscribing to Notepad.

At Build 2026, Microsoft Sent a Clear Message: Copilot+ PCs No Longer Matter

Copilot+ PCs were once the key to Microsoft's AI ambitions in Windows, even if its vision never fully came to fruition. Microsoft still heavily promoted AI at this year's Build conference, but it neglected to mention its AI-first PC hardware brand. In fact, the only person who brought up Copilot+ PC features to me was a protester who pointed to the backlash Microsoft faced after introducing Recall. Now, the company is focusing on agents that run locally across a wider range of devices (especially new Nvidia-powered hardware). The message is clear: The agentic future of Windows won't require a Copilot+ PC. This change can't come soon enough. The Copilot+ PC Brand Was Missing in Action at Build Microsoft CEO Satya Nadella quickly dismissed Copilot+ PCs at the start of his keynote, telling developers that "you now have the full scope of GPUs that you can get to" when writing AI software targeting Windows ML. "I'm really thrilled that every developer out there can count on building for local onboard AI and then have it run across all of the install base," he said. That's a big change from Microsoft's previous approach to local AI features on Windows. Everything from AI-powered settings configuration to the controversial Recall tool to semantic search required a PC with a neural processing unit (NPU). That was a big part of what defined Copilot+ laptops. The vast majority of Windows 11 PCs, including even the most powerful desktop PCs, cannot support these AI features. The only way to get them was to buy new dedicated hardware. At Build, when the company showed me the Surface Laptop Ultra and Surface RTX Spark Dev Box, local AI was still the big focus. But the company wasn't interested in promoting the NPU or discussing any of the existing Copilot+ PC exclusives. OpenClaw-style AI agent experiences got all the attention. And those don't have the same arbitrary requirements. Microsoft Wants Local AI to Run Beyond Copilot+ Hardware Because most people likely can't afford the Surface Laptop Ultra or other Nvidia Spark-powered machines, I expect future versions of Windows 11 to deliver AI experiences using small on-device models. It would be a real mistake if the company went right back to restricting its best features to expensive hardware. Case in point, Microsoft announced its Aion-1.0-Instruct small language model on stage during the Build keynote. The company is integrating this model directly into the Microsoft Edge browser for summarization and other web browsing-related tasks. "This language model is smaller, faster, and more efficient," writes Sohum Chatterjee, Edge's web platform product manager, in a blog post. He notes that it runs on devices with less powerful GPUs and even on CPUs. The blog post doesn't mention even NPUs. Even the Copilot+ Hardware Rules Are Starting to Crumble When Microsoft launched Copilot+ exclusives in 2024, the company set a hard floor of 16GB of RAM for these experiences. Therefore, no laptop with less than that amount could earn the Copilot+ PC branding. More broadly, it seemed like an admission on Microsoft's part that you needed a minimum of 16GB to achieve good performance from a Windows machine. But the high price of RAM and external industry pressure are forcing Microsoft to adapt. To that last point, Apple launched the $599 MacBook Neo with 8GB of RAM and support for Apple Intelligence features. Other PC manufacturers are releasing affordable, capable laptops with 8GB of RAM to compete. Even Microsoft itself announced a version of the Intel Panther Lake-powered Surface Laptop for Business with the same amount of memory. I mentioned the MacBook Neo and its 8GB of RAM in the context of the 16GB requirement for Copilot+ PC's to a Microsoft representative at Build, but they didn't comment beyond saying that the company was always trying to make Windows more efficient. That's something Microsoft has been talking about a lot in recent Windows Insider builds, and I think it should remain an area of focus as users try to make their devices last as long as possible. The End of Copilot+ Exclusivity Is Good for Everyone The bottom line is that you should expect more AI features -- including local AI features -- to arrive on all your Windows 11 PCs going forward. Yes, Microsoft is offering some fancy new hardware that promises to get you the most out of AI features, but it doesn't seem intent on requiring you to buy specialized hardware to use them at all. NPUs will still be useful for running AI tasks efficiently, but they won't be the only way. The exclusivity of Copilot+ PCs (and anything like them) seems to be waning quickly, and I'm glad to see it. I was on the ground at Microsoft Build this week. For more coverage of the event, check out our Microsoft Build landing page.

I Saw the Future of Windows at Microsoft Build, and It's Unrecognizable

SAN FRANCISCO -- From the front row at the Build 2026 keynote, I watched Microsoft CEO Satya Nadella promote Windows as the best platform for AI agents. Various Microsoft leaders then took the stage to demonstrate a new Windows companion app for the OpenClaw AI system. They proudly demonstrated the AI agent trying and failing to delete a folder of user files (thanks to new guardrails), before declaring OpenClaw safe to run on business PCs. Microsoft clearly thinks OpenClaw-style AI agents are the future of all computing, but it's first making its case to business users and developers by prioritizing security and hardware that's capable of running local AI models. To that end, it announced Microsoft Execution Containers (MXC) for securely running OpenClaw on Windows and championed its Nvidia RTX Spark-powered Surface Laptop Ultra. And the agentic experience is coming to regular Windows users soon, too, with the launch of the OpenClaw-based Microsoft Scout imminent. The company isn't talking about Windows 12 at all yet because it believes Windows 11 can already usher in an age of computing in which agents are more important than human users. That vision might sound like it's straight out of a science-fiction novel, but it's exactly what Microsoft sold at Build. Below are my main takeaways from the keynote. Microsoft Wants AI Agents to Take the Wheel Nvidia CEO Jensen Huang appeared during the Build keynote live from Computex in Taipei to enthusiastically discuss the future of AI agents with Nadella. "The PC evolved from being an incredible tool to now being a tool that is used autonomously by an AI assistant," said Huang. "The idea: I could be traveling and on the phone, and text my PC, and ask my PC to get coding done," he said. "My PC became an assistant," Huang continued. "The idea that the PC evolved from a personal computer to a personal AI is really exciting." In fact, Nadella and other Microsoft leaders hardly spoke about AI outside the context of agents. That extends to the hardware side, too. With Project Solara, Microsoft imagines end-user computing focusing on new agent-first devices that don't run traditional applications at all. For Windows PCs, the company envisions AI agents taking actions on your behalf. "We want Windows to be a fantastic place to run and scale agents," said Nadella on stage. "We are very deeply engaged with the team to make OpenClaw run super well on Windows." I see no reason for this focus on optimization unless agents are going to become a major part of all Windows experiences, not just the developer-focused ones. It's not a leap to reimagine Jensen's example in the context of everyday computing tasks. Microsoft seems to be on a mission to develop 'calm' experiences, and outsourcing your busywork to a local machine with a personalized agent fits that narrative. How long it takes a practical version of that concept to trickle down to the consumer experience is an open question. Why Microsoft Showed an AI Agent Failing on Stage Earlier this year, the open-source OpenClaw AI agent system transfixed the tech industry, with OpenAI going so far as to hire its creator, Peter Steinberger. However, OpenClaw was an experimental piece of software that required a dangerous level of access to a computer's operating system. The demand for hardware dedicated to AI agents even led to a shortage of Mac minis. As mentioned, Microsoft is tackling the securing concerns of such AI agents with MXCs. In these restricted environments, a developer or IT administrator decides what resources they can access. The idea is to run AI agents on your primary Windows PC, while relying on Windows to keep them under control. Microsoft's Samantha Song and Scott Hanselman demonstrated the new OpenClaw Windows companion app on stage, which lets you configure the claw agent's permissions in a few clicks. They showed how to set the Desktop folder to read-only, and then, in one of the keynote's most memorable moments, asked the OpenClaw agent to delete everything on the desktop. It failed to do so. Steinberger (the so-called "clawfather") took the stage moments after. "I'm so excited to see OpenClaw native on Windows," he said. "You know, watching a claw try to delete all your desktop files and just fail makes me really happy. Because six months ago, that totally would've worked," he said to a laughing audience. Other companies are also on board, so MXC seems poised to quickly become the standard for securing AI agents for deployment on Windows PCs. "Continuously-running local agents, like Hermes Agent, require intentional isolation. Developers need control over what an agent can access and trust that those controls will hold," says Dillon Rolnick, CEO of Nous Research. Microsoft says the new Hermes Agent application for Windows will integrate MXC. I can imagine a lot of regular Windows users wanting to wait to see how well MXCs work since an uncontrolled OpenClaw agent can absolutely wreak havoc on your digital life. After the problematic rollout of Recall, I expect at least some initial hesitation from anyone who cares about privacy and security. The Vision Is Clear, But the Use Case Isn't Grandiose promises are nothing new from tech companies, but the advent of AI has caused such claims to inflate exponentially. Microsoft has made some concrete progress on its development of an agentic Windows: It has real hardware, such as the RTX Spark Dev Box and the Surface Laptop Pro, that can run local AI models with some guarantee of security. It hasn't solved all the problems with AI agents, but it seems committed to helping businesses and developers overcome roadblocks and create compelling experiences. Microsoft might be able to sell its vision to the technical crowd at Build, but it needs to start doing the same for regular users if it intends to make people want to use Windows again. I appreciate that Microsoft continues to introduce long-awaited changes and scale back some AI features, but the value of AI agents isn't obvious and could spur even more AI pushback if it isn't careful. Unless the company can provide a clearer idea of how ordinary Windows 11 users can easily use AI agents that improve their lives, I suspect the reception will be unenthusiastic at best. Catch Up on Everything You Missed at Build I was on the ground at Microsoft Build for the entire event, attending demos and experiencing the future of Windows. Check out our Build 2026 recap for all the details.

My Big Build Takeaway: Microsoft Is Pushing For AI You Control on Your Terms

SAN FRANCISCO -- What stood out most for me at this year's Microsoft Build conference was the emphasis Microsoft put on how anyone can "participate fully" in the new era of AI and agentic computing. As Build is a developers conference, they were the intended audience. However, there was a similar message for enterprises. CEO Satya Nadella stressed that organizations could use their own data to fine-tune models and create and manage their own agent ecosystems, while keeping costs in check. The company even applied that concept to individuals in the communities where Microsoft wants to build data centers. Nadella said Microsoft has to prove it won't raise electricity costs or use a significant amount of water, while contributing to the tax base, and helping local organizations. There were lots of Windows-specific announcements, including new ways to get a clean desktop and move the taskbar, as well as a new Intelligent Terminal designed for developers who would put the regular terminal in one window and an agent in the other. Perhaps the most important of these announcements is the ability to run agents in containers that can be sandboxed and have their own permissions, all controlled by the individual developer and the organization. These Microsoft Execution Containers (MXC) isolate agents so they can't do damage to other systems or resources (such as a rogue agent accidentally deleting a database). Within these containers, you can run agents, including an OpenClaw. (Running OpenClaw on Windows was a big theme throughout the show.) It's a powerful tool that can do many things on your behalf, but organizations have been reluctant to allow users to run it on their machines because of the potential for mischief. The new container framework should mitigate many of those concerns. The company also described other ways of creating long-running agents that work on your behalf as "autopilots." 'This Is a New Era in AI' Microsoft AI Head Mustafa Suleyman also announced seven new Microsoft AI models, including a general model, the company's first reasoning model, and new models for creating images, transcriptions, speech, and code. Microsoft made a big deal about having a "clean lineage" with lots of transparency about how they were trained. Interestingly, the company didn't claim its new models are the best, just the most cost-effective for many tasks. "This is a new era in AI...that you control on your terms," Suleyman said. This concept of grounding AI within the organization is something Microsoft stressed at its Ignite conference last fall, where it introduced WorkIQ, a layer that includes information stored in Microsoft systems such as email, Teams, OneNote, and SharePoint. At Build, it announced a new component, WebIQ, which it touted as the fastest way to get real-time web data. Combining this with information from the AI tools and agents (Foundry IQ) and its data warehouse (Fabric IQ), users can create a comprehensive context for AI systems and agents. This seems to be a major way Microsoft wants to differentiate itself from other agentic platforms. More to the point, Nadella described how organizations could fine-tune these models using their own internal data to create a "hill-climbing" AI tailored to their knowledge and ways of working. I doubt that smaller organizations or individuals will do much of this, but I can see where it would be attractive to larger enterprises. Microsoft devoted a lot of time and floor space to its new hardware, specifically a Surface Laptop Ultra and an RTX Spark Dev Box, both running a new Nvidia RTX Spark processor that the company says can support up to 128GB of shared memory and run up to 120-billion-parameter local models. These are due in the fall, and look very powerful, especially for AI developers. But of course, the changes to Windows and the development tools should work on any machine, and Microsoft continues to push the use of Windows 365 cloud instances for development as well. The big benefit of running local models is that they incur no additional costs because they use the device's processors. Nadella called this "unmetered intelligence," and given the costs many organizations have run into with cloud-based tools, I can see how this could be very attractive. Of course, it's not just local PCs that are changing in the move to "agentic computing." Nadella discussed how the move to agents was affecting the entire computing stack, from infrastructure to underlying models and tools to security. In another session, Microsoft EVP for Cloud and AI Scott Guthrie described the significant changes in building data centers today, from new networking approaches to more automated, managed services -- all necessary because new data centers are much larger and being built at a faster pace than ever. Nadella said Microsoft has added more data center capacity in the last 18 months than it did in the first 10 years of Azure. Nadella ended his keynote by reminding the audience that, in general, there are two possible stories we can tell about the move to AI and agentic computing. In one, technology concentrates power and reduces human agency; in the other, we use the new technology to unlock more opportunities for everyone. Making the second alternative a reality is a north star for Microsoft. Taken together, these announcements should make AI development and deployment easier for developers and safer for organizations. All this is crucial, and we've heard similar concepts from Google and AWS in recent months. Still, I would argue there's still a lot more to do. Figuring out how we'll implement all this so we get real value is going to be the biggest issue for organizations for at least the rest of the decade.

The future of Windows 11 -- what is Microsoft building next?

Now that Microsoft Build 2026 is officially in the can, a clearer picture of the future of Windows 11 has appeared. Rather than announcing Windows 12, Microsoft is utterly dedicated to its current OS. That doesn't mean the operating system you're currently using is going to remain the same, though. Far from it. After CEO Satya Nadella wrapped his keynote, it was hard to shake the feeling that the traditional desktop OS may soon be a thing of the past. Microsoft's focus seems to be squarely aimed on delivering an AI-centred platform, where agentic systems are cooked into every corner of the Windows 11 experience. Perhaps the main takeaway from Build concerning Windows 11's future is that the OS could well make your day job easier going forward... well, unless you're a trapeze artist or a lion tamer. Using Microsoft IQ and MAI-Thinking-1, always-on AI agents will deploy real workplace knowledge to help users structure their daily workflow. The first of these to be announced was Microsoft Scout, and it could be a legit lifesaver for folks who struggle to organize hectic work schedules. Microsoft is giving developers more tools to assist users with day-to-day tasks. That's where the GitHub Copilot app (that uses the Big M's new reasoning model) enters the equation. It's available in preview to Microsoft Insider Program members now, and it's just one of a number of productivity-focused features that could transform Windows 11 going forward. Let's get into what that shiny new future might look like. Scout and about Following the latest Microsoft Build keynote, there can be little doubt as to where Microsoft's focus is on when it comes to the future of Windows 11. Productivity-focused AI agents are going to play a key part in the evolution of the operating system, which is where Microsoft Scout enters the picture. Available now to Microsoft Frontier customers, Scout is Microsoft's first Autopilot agent. Focused on full Microsoft 365 app integration, this "always on" agentic AI can operate across the cloud, web and desktop, allowing users to connect to Outlook, Teams and OneDrive. To put that in more digestible terms, think of Scout as your very own personal assistant; one that constantly organizes your day as it learns and adapts to how you work. Whereas traditional agents constantly needed to be prompted, Scout has been designed to work autonomously, allowing it to take proactive actions. This means it can help shape your working day by coordinating your schedule through 365 apps, be it organizing meetings across time zones or monitoring your inbox and Teams to flag any outstanding responses that need to be made. In theory, this should mean you no longer have to constantly flit between emails, docs and browser tabs, as much of this plate-spinning busywork can be delegated to Scout. Powered by OpenClaw, Scout monitors your working habits through Work IQ, in turn becoming more efficient at helping you complete tasks the more it observes your daily practices. Clearly, Scout is going to be more useful for business environments than Windows 11 users lounging around on the couch. Yet if this Autopilot agent lives up to its virtual co-worker potential, it could genuinely change the way people go about their jobs. Windows is thinking on its feet Another huge announcement at Build -- one that is directly tied to Scout -- was the unveiling of MAI-Thinking-1. Microsoft's new reasoning model helps the Autopilot AI out with planning tasks, and senior product manager Tanaya Yadav briefly demoed the tech during the keynote. While the task of training Scout could take hours, the endgame benefit should hopefully be AI agents that prove hugely useful for workflow purposes. MAI Thinking could have a big and broad impact across Windows 11 going forward. With the future of Copilot and 365 tied to agentic systems, this reasoning model may shape Microsoft's AI-focused goals going forward. The pivot to MAI is also interesting in that it looks to free Microsoft from being reliant on OpenAI. As Windows 11 continues to evolve, Copilot should primarily take instructions from the Redmond giant's internal reasoning tech. MAI is just the starting point too, as Microsoft's AI Superintelligence Team (that must be one seriously cool business card) is also developing a further six in-house reasoning models. Say hello to Solara One of the most pleasant surprises to come out of Build this year was the announcement of Project Solara. This chip-to-cloud platform has been built from the ground up to ensure interacting with AI agents works in seamless fashion. Probably the coolest feature of Solara is its eye-catching form factor. Because it's been designed for spaces where traditional PCs and laptops might not be around, Microsoft has focused on delivering products that don't take up much space. Enter devices which resemble a smart clock, and even more niftily, a work ID badge. The clock-aping gizmo is intended to remain stationary, whereas the badge has been designed to be used while you're on the move. On the latter, it can be unlocked with your fingerprint. The badge can even record video and take voice instructions, and it will subsequently clean up said audio so that you can send the note to colleagues in easily decipherable chunks. Engineered with a new era of AI agents in mind, Project Solara takes cues from Android devices. The daring design of this new agent-focused OS seems to suggest Microsoft is fully committed to a new era of AI that thumbs its nose at traditional hardware by embracing imaginative new forms. Security is being geared for AI agents With Microsoft increasingly pivoting towards AI as the future of Windows 11 takes shape, the matter of security becomes ever more vital. While the prospect of Scout mapping out your workday as you sip on your morning mocha is undoubtedly an appealing one for folks who get easily flustered, protections do need to be in place. This is where the announcement of Microsoft Execution Containers (MXC) becomes crucial. Now in preview, this security framework allows developers to run containment boundaries that are enforced by Windows 11. Essentially, a dev describes the security parameters it needs its agents to adhere to, then the OS enforces them everywhere these agentic systems run. Now, while this probably sounds roughly as exciting to you as the prospect of dining in a restaurant that only serves steamed celery, MXC is important to the future of AI agents. As they become more autonomous, it's crucial that these agents can't accidentally access sensitive info or make decisions they're not permitted to. With MXC in place, admins should be able to rest safe in the knowledge that this agent-native runtime is operating under safe security parameters. GitHub Copilot shifts the focus to devs Now available in preview to developers, the GitHub Copilot app is an AI-assisted engineering platform that brings agentic workflows to native desktop experiences. One of the major Build takeaways concerning the app is how it will work autonomously, though not in an unfettered manner (devs will most definitely still be in control). GitHub Copilot will use the newly revealed MAI-Code-1-Flash, an internal coding model that should improve coding performance through latency reduction. The deployment of this MAI model is to ensure the GitHub Copilot app is more aware of good workspace practices. In turn, this allows it to function like a true agentic assistant, rather than a glorified chatbot that continually needs to be prompted. Going forward, GitHub Copilot should make developers' lives easier, taking some coding pressures off their shoulders as it tests and deploys code with genuine autonomy thanks to AI agents. In an ideal world, this app will decrease developer workloads, rather than physically replacing them. Again, none of this is particularly focused on the Average Joe/Jenny Windows 11 experience. Then again, Build has always been focused on developers over consumers. Ultimately, the agentic benefits of GitHub Copilot that developers receive will ideally lead to smoother OS experiences for all of us. The future of Windows 11 centers around AI If Windows 11 is just a gateway that allows you to jump into the best PC games, there likely wasn't much for you to get excited about coming out of Build 2026. Yet if you're a dev or someone who is looking to use AI tools to coordinate workflows, the latest keynote definitely provided a peek into the future of Microsoft's OS. Rather than revealing Windows 12, Microsoft doubled down on its commitment to its current operating system. Moving forward, AI is clearly going to be the cornerstone of everyday Windows functions. With agentic systems like Microsoft Scout leaning on reasoning models to help business-minded folks map out their day by coordinating schedules, and Project Solara devices that have been crafted to replace traditional hardware, the focus is on agents that have real autonomy in order to assist users. The Redmond Giant is going hard on a vision for Windows 11 where agentic AI is baked into its very fabric. Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Subscribe to Tom's Guide on YouTube and follow us on TikTok. Finally, you can visit our dedicated Tom's Guide Savings Squad hub for expert help on getting the best products for less.

Microsoft AI chief says company was "set free" from OpenAI to pursue superintelligence

For three years, Microsoft's artificial intelligence story has been inseparable from OpenAI. The partnership -- cemented by a cumulative investment exceeding $13 billion -- gave Microsoft early access to the most advanced AI models on the planet, catapulting its Copilot products into the enterprise mainstream and adding hundreds of billions of dollars to its market capitalization. To the outside world, Microsoft's AI strategy was OpenAI. Mustafa Suleyman wants to change that narrative. In an exclusive sit-down interview with VentureBeat at Microsoft Build 2026, the CEO of Microsoft AI disclosed that a contractual change with OpenAI roughly six months ago granted his division the formal authority to pursue what he openly calls "superintelligence" -- using Microsoft's own researchers, its own data pipelines, and its own custom silicon. "We were only sort of set free from our contract with OpenAI about six months ago to formally pursue superintelligence," Suleyman said. "So this is very early days." The comment, delivered matter-of-factly backstage at the Fort Mason Center here, offers the clearest signal yet of a strategic inflection point unfolding inside the world's most valuable public company. Microsoft is not abandoning OpenAI. But it is building something alongside it -- and, eventually, something that could stand entirely on its own. Microsoft's first in-house model family signals a new level of AI ambition The most tangible evidence of that shift arrived the same day. Microsoft announced a family of seven new AI models developed entirely in-house by its AI Superintelligence Team, spanning reasoning, code generation, image creation, transcription, and voice synthesis. The models -- branded under the "MAI" family name -- are Microsoft's most ambitious first-party AI release to date. The flagship, MAI-Thinking-1, is a 35-billion-active-parameter reasoning model that Microsoft says matches leading models in its weight class on key software engineering benchmarks and demonstrates advanced mathematical reasoning. Suleyman emphasized one point repeatedly: the model was trained from scratch on clean, commercially licensed data, without distillation from third-party frontier models -- a direct, if unstated, contrast to the widespread industry practice of using outputs from competitors' systems to train cheaper alternatives. "We train our reasoning models from scratch," Suleyman wrote in a blog post accompanying the announcement. "We don't distill from other labs and we don't rely on unlicensed or opaque data." The rest of the family fills out a multimodal portfolio designed for enterprise deployment: MAI-Code-1-Flash, a lightweight coding model built specifically for GitHub Copilot and VS Code; MAI-Image-2.5, which supports both text-to-image and image editing; MAI-Transcribe-1.5, which Microsoft claims is the most accurate transcription model available, operating across 43 languages; and MAI-Voice-2, a multilingual speech-generation system. All of the models ship through Microsoft Foundry, the company's model-hosting and deployment infrastructure, and for the first time, developers can tune model weights themselves through third-party platforms including OpenRouter, Fireworks, and Baseten. But Suleyman made clear in the interview that the seven models are a proof of concept, not a finished product. The real project is the lab itself. "Our job is to make sure that when we look out to 2030 and beyond, we have the capacity not just to buy models from third parties, but to build the absolute frontier, the best models in the world," he said. "That's a long transition." What "set free" from OpenAI actually means for Microsoft's AI future To understand what Suleyman means by "set free," you need to understand the unusual contractual architecture that has governed Microsoft's AI efforts for years. When Microsoft invested billions into OpenAI beginning in 2019, the partnership came with a specific arrangement: OpenAI would build the frontier models, and Microsoft would serve as the exclusive cloud provider, integrating those models into its products and reselling them through Azure. The deal gave Microsoft extraordinary commercial leverage -- access to the world's most advanced AI without having to build it -- but it also created a dependency. Microsoft was explicitly barred from pursuing its own AGI research, and the agreement even capped how large a model the company could train, restricting it from building systems beyond a certain computing threshold measured in FLOPS. That arrangement was formally renegotiated. As Fortune and Axios reported in November, a revised deal with OpenAI removed those restrictions, clearing the way for Suleyman to launch the MAI Superintelligence Team and pursue what he calls "humanist superintelligence." The result, in Suleyman's telling at the time, was a "best-of-both environment, where we're free to pursue our own superintelligence and also work closely with them." By the time he sat down with VentureBeat at Build 2026, roughly six months had passed since that self-sufficiency effort formally began. Microsoft had already started shipping in-house models -- including MAI-Image-2-Efficient, a lighter-weight image generation model released in April -- but the seven MAI models announced at Build are the team's most ambitious release yet: a full multimodal family spanning reasoning, code, image generation, transcription, and voice. Even so, Suleyman does not view the shift as a rupture with OpenAI. He described Microsoft's current position as one of abundance, not scarcity. "There's no immediate urgent need to fill a gap in three months' time or six months' time," he said. "We have OpenAI, we have Anthropic, we have thousands of models inside Foundry. So there's already a huge amount of optionality available to us." The framing is telling. Microsoft's push into first-party frontier models is not born out of a crisis in the OpenAI relationship but out of a strategic calculation: as AI becomes the most consequential technology layer in enterprise computing, the company cannot afford to depend entirely on partners for the foundational capability. "Over the next five years, we have to be able to produce state-of-the-art frontier-scale models," Suleyman said. "That's our mission." Suleyman says the shift from chatbots to autonomous AI agents has already begun If the seven MAI models represent the technical ambition, a new capability called Frontier Tuning represents the commercial logic. Announced alongside the models at Build, Frontier Tuning allows enterprise customers to customize MAI models using their own proprietary data, workflows, and domain terminology, all within their own secure compliance boundary. The system uses reinforcement learning environments -- what Microsoft calls "training gyms for AI" -- that let agents learn directly from real workplace tasks without affecting production systems. The results Microsoft shared are striking. An MAI model tuned for Excel reportedly matches GPT 5.4 performance while operating at up to ten times greater efficiency. Early enterprise adopters are seeing similar gains: when tuned for one unnamed organization's exacting standards, the MAI model achieved the highest win rate of any model tested at roughly one-tenth the cost. Suleyman framed Frontier Tuning as part of a broader evolutionary stage -- a move from intelligence to action. "We've basically moved beyond just conversation," he told VentureBeat. "Now we're moving to action." He introduced a new framework for thinking about that progression: the shift from IQ (factual intelligence) to EQ (emotional intelligence, or the ability to follow tone and style instructions) to what he calls AQ -- the "Actions Quotient." Future AI agents, in Suleyman's telling, won't just answer questions. They will log into enterprise software, navigate complex multi-application workflows, and execute tasks across Excel, Word, Teams, Jira, Adobe InDesign, and customer relationship management systems -- just as a human employee would. "You should be able to show up on day one and almost provision credentials to a new AI agent," he said. "The model needs to be able to move across all of these different environments, and that's actually the great strength of Microsoft." The Build 2026 announcements bore this out in concrete product terms. Microsoft Scout, the company's first "Autopilot" agent, operates as an always-on background assistant built on the open-source OpenClaw technology. It runs with its own governed identity inside Microsoft Entra, so its actions are auditable and attributable. Windows 365 for Agents gives AI agents their own managed Cloud PCs, allowing them to interact directly with applications and browsers inside enterprise environments. And the Foundry platform received major updates -- including hosted agents with sub-100-millisecond cold starts, a new Microsoft Agent Framework, and one-click publishing to Teams and Microsoft 365 Copilot. Why Microsoft believes enterprise data is the next AI training frontier Suleyman also articulated why he believes Microsoft's position is uniquely defensible -- and the argument has less to do with model architecture than with where work actually happens. "We've sort of hoovered up all of the obvious pools of training data," he said, referring to the industry's early scramble to ingest the open web. "In the next phase, we actually want to be able to give these agents to companies to train on their specific tasks with the data that they have inside of their own big workflows." The claim is subtle but consequential. The first wave of generative AI was trained on publicly available text -- books, websites, Reddit posts, code repositories. That data is now largely exhausted, and its use is increasingly contested in court. The next wave, Suleyman argues, will be trained on enterprise-specific data: the internal workflows, decision traces, and institutional knowledge that define how real organizations operate. Microsoft, which serves 493 of the Fortune 500 through Azure according to Suleyman, is already embedded inside those workflows through Microsoft 365, Teams, Dynamics 365, and the broader Azure ecosystem. Frontier Tuning is the mechanism that converts that positional advantage into model performance. "People underappreciate that that's going to be the next domain," Suleyman said. The early partner list for Frontier Tuning reflects the ambition: Mayo Clinic, where Microsoft is co-creating a frontier AI model for healthcare using de-identified clinical data; EY, which is tuning a tax-advisory agent for deployment to 75,000 professionals globally; Land O'Lakes, where Frontier Tuning delivered what the company's product development scientist called "meaningful improvements in grounded outputs and style compliance"; and Pearson, which is using tuned models to provide learning-science-aligned feedback in its Communication Coach product. The Mayo Clinic partnership may be the most significant. Microsoft and Mayo Clinic are collaborating to build a healthcare-specific frontier model that combines Mayo's clinical expertise and longitudinal patient insights with Microsoft's AI capabilities. The model will be owned by Mayo Clinic and deployed first within Mayo's own environment before being made available to other organizations through Foundry. Microsoft's custom AI chips and GPU buying spree reveal the scale of its compute advantage None of this works without an industrial-scale compute infrastructure, and Suleyman was unusually candid about the hardware economics underlying Microsoft's strategy. "We are the largest buyer of GPUs on the planet," he said. "We're the largest buyer of GB200s and GB300s in the world." Microsoft will continue purchasing Nvidia accelerators "for many, many years to come," Suleyman said. But the company is simultaneously building its own custom silicon. Maia 200, Microsoft's second-generation AI accelerator, is already running in production across data centers in Iowa and Arizona, with deployments planned for Italy, Australia, and South Korea. According to Microsoft, Maia 200 delivers the best tokens-per-dollar-per-watt in the company's fleet. Suleyman put a finer point on the economics in the interview: Maia 200 is 30 percent more cost-efficient than Nvidia's GB200, he said. And when Microsoft co-optimizes its own MAI models to run natively on Maia silicon, the company sees an additional 1.4x improvement in performance per watt. "It is going to be cheaper in years to come to build on MAI models with Maia 200 and Maia 300 inside of Azure," he said. That claim -- if it holds at scale -- has profound implications for the competitive landscape. It means Microsoft is not merely buying its way to AI dominance through Nvidia; it is building a vertically integrated stack in which its own models, running on its own chips, inside its own cloud, tuned on its customers' own data, could offer performance and cost characteristics that no competitor can replicate. Suleyman rejects the idea that AI models are becoming commodities Suleyman also pushed back sharply against one of the most popular narratives in Silicon Valley: that AI models are rapidly commoditizing. "A lot of people are saying models are commoditizing," he said. "I don't think that's true." His argument hinges on what he calls "quality tokens" -- the proposition that the composition, curation, licensing, and deduplication of training data matter at least as much as raw scale. Microsoft's new MAI models, he said, were trained on a pre-training mix composed of approximately 50 percent high-quality code, with the remainder drawn from commercially licensed and carefully curated sources. The result, he argued, is a distinct "lineage" of models optimized for coding, reasoning, and agentic behavior -- fundamentally different from models optimized for consumer chat, cultural content, or multilingual breadth. "We're going to see very distinct lineages that reflect different training objectives of different companies," he said. "Quality tokens matter more than just brute-force scale." This is a strategically important argument for Microsoft to make. If models are commodities -- if any lab can match the frontier within months using cheaper compute and distilled training data -- then the model layer becomes a race to the bottom, and Microsoft's billions in compute investment offer no durable advantage. But if model quality is a function of data discipline, research depth, and institutional patience, then the lab-building approach Suleyman is pursuing becomes a genuine competitive moat. He used a specific metaphor to describe that approach, one borrowed from optimization theory: the "hill-climbing machine." The phrase describes a system that continuously improves -- cycle after cycle -- by applying more compute, better data, and sharper evaluation. "The goal here is to build what we think of as a hill-climbing machine," he wrote in his blog post. "An organization that can continuously improve, cycle after cycle." The metaphor is revealing because it describes a process, not a destination. Suleyman is not promising that Microsoft will build the world's best model next quarter. He is arguing that Microsoft is building the system -- the research culture, the data pipelines, the silicon co-optimization, the evaluation infrastructure -- that will produce progressively better models over years. Inside Microsoft's five-year plan to become a self-sufficient AI superpower The strategic picture that emerges from Suleyman's comments -- and from the full scope of the Build 2026 announcements -- is of a company preparing for a future in which AI capability is not rented from a partner but generated internally, at scale, across every layer of the stack. Microsoft still needs OpenAI. The partnership continues to power Copilot, Azure AI services, and ChatGPT's infrastructure. Suleyman acknowledged as much, describing Microsoft's portfolio of model providers as a source of strength, not a problem to be solved. But the direction of travel is unmistakable. With its own frontier models, its own custom silicon, its own reinforcement learning environments for enterprise tuning, and its own autonomous agent infrastructure, Microsoft is constructing a parallel path -- one that, by 2030, could make the company a fully self-sufficient frontier AI lab embedded inside the world's largest enterprise software platform. "Our ultimate goal is what we call Humanist Superintelligence," Suleyman wrote in his blog post. "That means advanced AI systems designed to serve people and organizations, not replace them." Whether that goal is achievable -- or even clearly definable -- remains one of the great open questions in technology. And Suleyman expressed more confidence than caution when asked about the trajectory of progress. "I really think we're at the tip of the iceberg," he said. "The models are so much more powerful than we know how to extract intelligence from them." But confidence and execution are different things. Building a frontier lab is not an announcement; it is a decade-long commitment that requires retaining elite researchers, maintaining scientific rigor under commercial pressure, and producing results that justify the staggering capital expenditure. Google learned this with DeepMind -- which Suleyman himself co-founded in 2010, before joining Microsoft -- and even that lab, widely regarded as one of the best in the world, spent years navigating the tension between pure research and product delivery. Suleyman seemed aware of the contradiction. "If you rush it, you'll screw it up," he said. The sticker on his laptop reads: "Patience and urgency." It is a paradox that Microsoft now has five years -- and several hundred billion dollars -- to resolve.

The Future Microsoft Showed at Build 2026 Barely Looks Like Windows

SAN FRANCISCO -- At Build 2026, I watched from the front row as Microsoft CEO Satya Nadella extolled the utility of AI agents for everything from business application development to scientific research in his keynote. But it was the OpenClaw announcements that drew the loudest applause. The highlight of the presentation was a demo in which Microsoft proudly showed off a sandboxed local AI agent repeatedly trying and failing to delete a bunch of user files. The takeaway, of course, is that Microsoft wants you to want OpenClaw-style AI agents on your PC. And it thinks the way to convince you (starting with developers) of that vision is to emphasize safety and dedicated hardware platforms. To that end, it announced Microsoft Execution Containers (MXC) for securely running OpenClaw on Windows, alongside a dedicated companion app. The company also championed its Nvidia RTX Spark-powered Surface Laptop Ultra, which can run powerful AI agents locally -- no data centers or internet connection required. What this all means for regular Windows users remains unclear. Is the goal to give them the ability to spin up AI agents that perform basic tasks, convincing them this is the future of personal computing? It seems ambitious, but I can't see anywhere else Microsoft would take Windows 11 based on what I've seen thus far at Build. Below are my main takeaways from the keynote. Microsoft Wants AI Agents to Take the Wheel Nvidia CEO Jensen Huang appeared during the Build keynote live from Computex in Taipei to enthusiastically discuss the future of AI agents with Nadella. "The PC evolved from being an incredible tool to now being a tool that is used autonomously by an AI assistant," said Huang. "The idea: I could be traveling and on the phone, and text my PC, and ask my PC to get coding done," he said. "My PC became an assistant," Huang continued. "The idea that the PC evolved from a personal computer to a personal AI is really exciting." In fact, Nadella and other Microsoft leaders hardly spoke about AI outside the context of agents. That extends to the hardware side, too. With Project Solara, Microsoft imagines end-user computing focusing on new agent-first devices that don't run traditional applications at all. For Windows PCs, the company envisions AI agents taking actions on your behalf. "We want Windows to be a fantastic place to run and scale agents," said Nadella on stage. "We are very deeply engaged with the team to make OpenClaw run super well on Windows." I see no reason for this focus on optimization unless agents are going to become a major part of all Windows experiences, not just the developer-focused ones. It's not a leap to reimagine Jensen's example in the context of everyday computing tasks. Microsoft seems to be on a mission to develop 'calm' experiences, and outsourcing your busywork to a local machine with a personalized agent fits that narrative. How long it takes a practical version of that concept to trickle down to the consumer experience is an open question. Why Microsoft Showed an AI Agent Failing on Stage Earlier this year, the open-source OpenClaw AI agent system transfixed the tech industry, with OpenAI going so far as to hire its creator, Peter Steinberger. However, OpenClaw was an experimental piece of software that required a dangerous level of access to a computer's operating system. The demand for hardware dedicated to AI agents even led to a shortage of Mac minis. As mentioned, Microsoft is tackling the securing concerns of such AI agents with MXCs. In these restricted environments, a developer or IT administrator decides what resources they can access. The idea is to run AI agents on your primary Windows PC, while relying on Windows to keep them under control. Microsoft's Samantha Song and Scott Hanselman demonstrated the new OpenClaw Windows companion app on stage, which lets you configure the claw agent's permissions in a few clicks. They showed how to set the Desktop folder to read-only, and then, in one of the keynote's most memorable moments, asked the OpenClaw agent to delete everything on the desktop. It failed to do so. Steinberger (the so-called "clawfather") took the stage moments after. "I'm so excited to see OpenClaw native on Windows," he said. "You know, watching a claw try to delete all your desktop files and just fail makes me really happy. Because six months ago, that totally would've worked," he said to a laughing audience. Other companies are also on board, so MXC seems poised to quickly become the standard for securing AI agents for deployment on Windows PCs. "Continuously-running local agents, like Hermes Agent, require intentional isolation. Developers need control over what an agent can access and trust that those controls will hold," says Dillon Rolnick, CEO of Nous Research. Microsoft says the new Hermes Agent application for Windows will integrate MXC. I can imagine a lot of regular Windows users wanting to wait to see how well MXCs work since an uncontrolled OpenClaw agent can absolutely wreak havoc on your digital life. After the problematic rollout of Recall, I expect at least some initial hesitation from anyone who cares about privacy and security. The Vision Is Clear, But the Use Case Isn't Grandiose promises are nothing new from tech companies, but the advent of AI has caused such claims to inflate exponentially. Microsoft has made some concrete progress on its development of an agentic Windows: It has real hardware, such as the RTX Spark Dev Box and the Surface Laptop Pro, that can run local AI models with some guarantee of security. It hasn't solved all the problems with AI agents, but it seems committed to helping businesses and developers overcome roadblocks and create compelling experiences. Microsoft might be able to sell its vision to the technical crowd at Build, but it needs to start doing the same for regular users if it intends to make people want to use Windows again. I appreciate that Microsoft continues to introduce long-awaited changes and scale back some AI features, but the value of AI agents isn't obvious and could spur even more AI pushback if it isn't careful. Unless the company can provide a clearer idea of how ordinary Windows 11 users can easily use AI agents that improve their lives, I suspect the reception will be unenthusiastic at best. Stay Tuned for More Build Coverage I'll be on the ground here at Microsoft Build for the entire event, attending demos and experiencing the future of Windows. Follow our Build 2026 live blog for all the latest news.

Microsoft's AI Futurist explains how he uses Copilot -- and the real-world problems enterprises are solving with agents

Microsoft used its Build 2026 conference this week to push a clear message: agents are rapidly moving into production throughout enterprise systems, and the winning platform will be the one that gives them reliable context, governance, identity, memory -- and secure access to enterprise data. The company announced Microsoft IQ as a context layer across GitHub Copilot, Microsoft Foundry and Copilot Studio; Work IQ APIs coming June 16; Fabric IQ for structured business data; Foundry IQ for retrieval across enterprise knowledge and the live web; and Web IQ as a new agent-facing web search stack. Microsoft also introduced Scout, a personal work agent, and a whopping seven new in-house AI models in its growing MAI family across modalities and use cases, including MAI-Thinking-1. Those announcements sit directly in Marco Casalaina's lane. Casalaina is Microsoft's VP Products, Core AI and AI Futurist. He leads Microsoft's AI Futures team and previously led teams across Azure AI, including Azure OpenAI, Vision, Speech, Decision, Language, Responsible AI and AI Studio. Before Microsoft, he led Salesforce's Einstein AI team and earned a computer science degree from Cornell University. CRN reported that he joined Microsoft in early 2022 as vice president of products for Azure Cognitive Services, meaning he has now been at the company for more than four years. VentureBeat spoke with Casalaina ahead of Build about Microsoft's agent strategy, the company's model-choice philosophy, how Microsoft IQ fits with MCP, and why he believes enterprises need far more than just access to powerful models. The interview below has been edited for clarity and condensed from the transcript. VentureBeat (VB): To start, can you explain your role at Microsoft and what "AI Futurist" means in practice? Marco Casalaina (MC): I am VP Products of what we call Core AI. Core AI is our set of tools for AI developers, and that includes Foundry, Visual Studio, VS Code, GitHub and GitHub Copilot. That's our overall group. My Silicon Valley title is AI Futurist, and that has a very concrete meaning here. I've worked with other folks who are considered futurists, like Peter Schwartz, and that can be a little bit more fuzzy. For me, what it means concretely is that I am the first person to try anything new here. I am constantly getting things from all over Microsoft, not even just Foundry, because I work with really everybody across the company. Pretty much everybody sends me the new things at all times. Even today, I got something brand new just before this call. I'm usually the first person to try anything new here, which is pretty cool. I get to see a lot of really cool stuff. A friend of mine, who is head of AI at Intuit, calls me an "adjacent possiblist." I consider my futurist concept to be about a year out from now -- the immediate future of what's about to happen next. That's what I focus on. VB: Where are you looking at the agentic state of things, and in particular Microsoft's position as enterprises and individuals rush to adopt agentic AI? MC: We can look at it from bottom to top. At the very base of the stack is our commitment to model choice. All along, we've had the OpenAI GPT frontier models. Now we have a really solid partnership with Anthropic, where we're offering the Claude models. We just launched Claude Opus 4.8 on Azure -- on Foundry, I should say -- and at Build, we are introducing our new MAI model. The MAI models are a set of frontier models that we're building in-house. They are made for token efficiency, optimization and customization. We are specifically making them for our customers to customize on their own data sets. One level above that, we are announcing hosted agents in Foundry. That is our managed agent capability in Foundry. It automatically handles scaling, containerization and those kinds of things. It is an environment where you can manage agents. One level above that is the Foundry control plane. At least for the agents you build, you want to have control over them. This gives you observability into their cost, tokens and correctness. You can do continuous evaluations and sample interactions with those agents, run evals and make sure they are continuing to work and not drifting. The big news is going to be the GA of what we call the IQs here at Microsoft. There are currently three, and there will be four. There is Foundry IQ, which is basically for knowledge -- largely unstructured knowledge. There is Fabric IQ. We have a ton of customers who have entrusted a lot of data to the Microsoft Cloud in Fabric, Power BI and related technologies. Fabric IQ is about making an agent-facing interface for this data, so agents can get to it without literally going through a Power BI report. That's ridiculous. Work IQ is about the Microsoft ecosystem. You can look at Work IQ as the agentic face of all the Microsoft apps: Outlook, Teams, Word, SharePoint and all those kinds of things. How does an agent interact with those things? That is Work IQ. And finally, the fourth IQ is Web IQ. We are releasing our new agent-facing web search capability. It can search the web, search through videos and even do some kinds of browsing tasks automatically. It is super fast, and it kind of has no face. It's headless. The interface is intended for agents. We will also be announcing Agent Optimizer. That includes a new type of evaluation that allows you to evaluate much more granularly whether an agent is actually working and working correctly. The optimization step can go back in and make modifications to the prompt, obviously with your consent, and modify your agent so it works more correctly going forward. Effectively, it creates a feedback loop to make agents work better. VB: Microsoft has sometimes been criticized for murky and clunky product naming. Where do these IQ products sit? Are enterprise users supposed to go to IQ first, or is IQ more for developers to connect to? MC: All of the IQs are headless. The concept of IQ is that each one provides a different type of context to an agent specifically. Largely, it will be developers interacting with the various IQs -- developers and the agents they build. The IQ brand is really about agent context. End users largely won't interact with the IQs. It is true that if you use Microsoft 365 Copilot today, you'll notice a little thing that says it is using Work IQ. So it is a little bit visible, but the customer or end user doesn't have to go find the IQ. Their system or developers hook that up. VB: Is the IQ family essentially Microsoft's version of MCP? Is it using MCP, or is it something different? MC: All of the IQs are indeed exposed as MCP servers. You have correctly characterized MCP as basically an agent-facing or self-describing API. It's not that fancy. That's really what it is, with some authentication layers and capabilities built in, which is super useful. Something like Work IQ -- really all the IQs -- have to be authenticated. In order for Work IQ to see my email, Teams messages, documents and stuff like that, I have to be able to authenticate it on behalf of me. That gets us to another core differentiator that we will be announcing at Build, which is agent identity. We have this Entra system, and Entra is, I believe, the world's largest used identity system for human users. For some time now, you have been able to declare an agent to have an identity in there. Now, agents will be able to have their own identity, their own Teams box, their own email inbox and stuff like that. These agents will use Work IQ to check their own email, check their own documents and that sort of thing. VB: Enterprises are not one-size-fits-all on models. Microsoft supports many leading models through Foundry and Azure, while also building its own. Is Microsoft a model company, an infrastructure company or a connector between models and work products? MC: The answer is yes. We are obviously the hyperscaler. We are absolutely committed to model choice, and we will continue to offer the frontier models from all of the major players: OpenAI, Anthropic, Mistral, Black Forest, xAI -- you name it. They are all going to be represented in there. At the same time, we have what is now called our Microsoft AI Superintelligence Team, formed by Mustafa Suleyman, and we are building our own frontier models as well. Like I said earlier, we are really gearing these models toward optimization -- token efficiency, bang for the buck and customization. These are things our customers have been asking for: the ability to more finely customize models, whether that is fine-tuning or continued pre-training. Continued pre-training is literally changing the weights of the model, whereas fine-tuning is adding a little layer on top. We have these capabilities in Foundry: fine-tuning, distillation and those kinds of things. I would note, by the way, that our MAI models are not distilled. Some model providers, especially some of the less scrupulous ones, will distill other models into theirs, and that can have unusual effects. We don't do that. The data provenance of our models is of primary importance to us. When we come out with these models, we want our customers to know that the data provenance is clean in terms of the rights to the data, where it came from and all that kind of stuff. The choice thing also goes above the model layer. When we talk about Foundry hosted agents, we have the Microsoft Agent Framework. You talk about agent orchestration -- how you make agents work together when you have multiple agents -- and Microsoft Agent Framework is an excellent framework for that. However, I can make a LangGraph or LangChain Foundry hosted agent. I can make a CrewAI Foundry hosted agent. I can use any number of orchestration frameworks and put that up as a Foundry hosted agent, and it becomes a first-class Foundry agent. That means I get the observability. It shows up in the Foundry control plane. I can do evaluations on it. I can do traces on it. I can get all those things from the Foundry control plane with an agent built in really any framework I choose. VB: Some companies are interested in Chinese and open-source models. How much of Microsoft offering its own models is about giving customers an American version of that? MC: I can't speak to that exactly. Of course, we offer DeepSeek models and Qwen models in Foundry, so we offer all of these choices today, and our customers can make that choice. The MAI models are really focused on token efficiency and customizability. That is what our customers are demanding, and that is the gap we are filling. VB: As agents take on longer tasks and more specialized work, will enterprises keep expanding the number of models they use, or will there be a winnowing? MC: I do see it expanding. We are not just focused on tokens per se. A token is not a token is not a token. One token is not necessarily equivalent across these things. It is all about what you are doing with each token and the efficiency of that. It comes back to what kind of value you are getting for the cost. That is a lot of the rationale behind why we are developing our own MAI models. Part of my job is to travel all around the world. I've been all over the place. For example, I've been working with Bayer. One of the things we are measuring is not just token usage, but number of users -- monthly active users and daily active users -- because we have a lot of first-party capabilities like Microsoft 365 Copilot. Over the last year, we've seen a 6x increase in monthly active users. We have over 20 million users of Microsoft 365 Copilot alone. That is on the agents you use. In terms of the agents you build, Bayer put up its own agent system on Foundry, and now it has 20,000 of its own employees on it. A few weeks ago, I was in Sydney, Australia, hanging out with AEMO, the Australian Energy Market Operator. They operate the electrical grid of Australia. They showed me that they had built agents to manage grid operations. This is a human-centered thing. They have grid operators sitting in centers in West Sydney, Brisbane and places like that, and they are bombarded with alerts. I wouldn't believe it if I hadn't seen it myself. The alerts are constant. They built a system to triage those alerts. Is this alert a super major thing, or is it just that a transformer is getting a little hot? It also says, here is when we had this problem last time, and here is how we resolved it last time. Maybe now we need to replace this component, or whatever. Ultimately, it is the grid operators making the choice. A lot of our philosophy here is human empowerment. These human-centered agents are the ones that are working best among our customers. What I saw at AEMO and Bayer is this notion of human empowerment: taking away some of the grunt work, or in the case of AEMO, taking billions of alerts and reducing them to something much more manageable and actionable for the people involved. We are moving past the era where agents are just answering questions. AI in general is moving past that. We are not just answering questions anymore. We are moving toward a place where AI can really meaningfully help you do your work. VB: How do observability, tokenomics, ROI analysis and agent governance fit into Microsoft Foundry? MC: That is what the Foundry control plane is all about. We introduced it in November of last year. If you looked at my own Foundry control plane -- I've built a ton of these agents, and I am a developer by background -- you would see all of my agents that are running and the ones that are paused. I can see how many tokens they've used over the last day, week or month. I can look at trends. I can look at costs, because the cost will be different depending on what underlying model I'm using. If I'm using our model router, it can route to different models depending on the complexity of the inbound prompt. We also have Azure cost management overall. Azure has had cost management for over a decade, before the AI thing even happened. This integrates with overall Azure cost management. It is not just narrowly about what your AI is doing. Your AI will be using storage resources, data resources and other compute resources around that AI. You can get a complete picture of not just the cost and token usage of the AI itself, but everything around it. When you think about governance, that also extends to evaluation. One of the things we are releasing in preview is rubric-based evaluation. Rubric-based evaluation is much more granular. Let's say you have built a restaurant reservation agent. The things you want to test about that agent are not really groundedness. Groundedness is the opposite of hallucination, and that is very question-answering. For a restaurant reservation agent, you want to test very granular things. If you say, "Make me a table for two tomorrow," did it come back and ask, "What time would you like the table?" Before it gave you a table for two tomorrow at 6 p.m., did it actually check that the table was available, or did it randomly give you a table without checking first? There are very granular things you want to test about that specific use case. You don't just want to test whether the agent works. You want to test whether the agent works right. That is what we are approaching with our new rubric-based evaluation system. You will see that in Satya's keynote. I have been using it myself lately, and I'm very happy about it. I've been waiting for this. VB: Microsoft is also partnering with companies like Anthropic and allowing Claude to work with Microsoft 365. How important is Copilot to this story? Why would someone turn to Copilot over other options? MC: Microsoft 365 Copilot is a huge advantage for us. As I mentioned, we crossed the 20 million user mark on Copilot relatively recently. The great thing about that is that it is the face. When you go into Foundry and make an agent, there is a button that says "publish to Copilot" -- actually, it says "publish to Copilot in Teams," because you can put it in Teams too. The idea is that you want to put these agents where your users are. A lot of people who use the Microsoft ecosystem are in Teams, or they are using Copilot. I can create a custom agent, as many of my colleagues have, and now it is in Copilot, which I use maybe 50 times a day. Since January, Copilot has become more and more capable. I now use it to draft my email. I am not just using it for question answering. I'm starting to use it to manage my calendar and draft emails. I really do this every day now. When I want to use a custom agent -- for example, to file my expenses, because we have a custom agent for that now -- I can access that agent not in some random standalone interface, but in Copilot or Teams, where I already am. That surface area that people are already engaging with is a major advantage. VB: As people offload more repetitive work to AI, what are they able to spend more time doing? MC: Let's consider something I did yesterday. I got an email from a customer named Frankie, and he asked me a question about Foundry hosted agents. I knew the answer because I had talked to my colleague Jeff Holland, who is the head of our hosted agents product management. I had asked Jeff the same question two weeks ago. Where or how I asked him, I don't remember. Was it in Teams? Was it email? Was it a meeting? I don't really remember. But I knew the answer to the question Frankie was asking. So I went into Copilot and said, "Answer Frankie's question about how hosted agents scale, and reference the conversation I had with Jeff a couple of weeks ago on this same topic." And it did it. It drafted the email. Over time, I have taught Copilot my style. I don't do the bold-print thing. I tell it: don't use em dashes and that kind of stuff. I have a certain style in the way I write emails. It's a little terse, to be perfectly honest, but I want it to be the way I write. It drafted this thing. It searched through my Teams messages, my emails and the transcripts of my meetings with Jeff. It used Work IQ, as a matter of fact. It found the answer, drafted the email and provided a link to the documentation that specifically covered the question Frankie was asking. I looked at the draft and thought, yep, that's it. Yes, I could have composed this email myself. I knew the answer to the question. I could have looked up the documentation. If I dug around, I'm sure I could have found the conversation I had with Jeff in whatever medium that was. I could have done that stuff. It probably would have taken me, I don't know, an hour to find all the information and compose it. Instead, I did it in about a minute. I had a draft, I looked at it, I was happy with it, I pressed send, and that was the end of that. It really is about giving people time back. It is not even just grunt work. It is all this time you spend looking things up and finding things. Now, I can make it take an action. It didn't just answer the question. It fully drafted the email and copied Jeff. VB: Do you fear for your job? How has AI changed your own work? MC: I don't fear for my job. My job has changed. For one thing, I do a lot more now, both in my business life and personal life. This weekend I was using Web IQ, the new Web IQ. I've been car shopping. My car's lease is coming up, and there is a very specific car I'm trying to find, which is hard to find. It's a Hyundai Ioniq 6, which Hyundai, for whatever reason, has stopped offering in the United States. I'm going to get one, though. I set my agent to the task, using Web IQ, of finding all the Hyundai Ioniq 6s available in the entire Bay Area -- everywhere, all the way out to Sacramento, all the way as far south as Gilroy. I set it to this task, and then I went on a hike. When I got back, I had a big long list of all the Hyundai Ioniq 6s, at least the 2024 and 2025 models, available in the entire Bay Area. From that, I started calling down these dealers. Even in my personal life, I'm using it constantly. It saves me a ton of time. That would have taken me hours, to go through every single dealer's inventory like this. But Web IQ could do that, and it was super quick. VB: Any final thought for developers around this news? MC: Foundry is really the place. This is the place where you can build your agents, scale your agents, test your agents and improve your agents. That's what it's all about, and it's happening.