{"id":1395,"date":"2026-06-03T14:42:42","date_gmt":"2026-06-03T14:42:42","guid":{"rendered":"https:\/\/silvybrand.com\/?p=1395"},"modified":"2026-06-03T14:42:42","modified_gmt":"2026-06-03T14:42:42","slug":"microsoft-build-ai-agents-openai-competition","status":"publish","type":"post","link":"https:\/\/silvybrand.com\/?p=1395","title":{"rendered":"Microsoft and OpenAI broke up \u2014\u00a0now they\u2019re ready to fight"},"content":{"rendered":"


\n<\/p>\n

\n
\n

At Microsoft\u2019s annual Build conference on Tuesday, the company announced a slew of new or expanded AI initiatives, including a super app, in-house reasoning models, a cybersecurity tool, and OpenClaw-esque AI agents. All this news added up to a clear message: Microsoft is positioned to be one of the biggest players in AI, and it\u2019s finally acting like it.<\/p>\n<\/div>\n

\n

For years, Microsoft\u2019s AI business leaned hard on its early and exclusive partnership with OpenAI. But the drama-filled marriage slowly devolved into a situationship, and the pair effectively separated<\/a> in late April (though Microsoft is still OpenAI\u2019s primary cloud partner \u2014 for now). This year\u2019s Build had the vibe of a freshly single divorc\u00e9e posting a thirst trap on Instagram. \u201cIt\u2019s always fun to be at developer conferences in times of great change,\u201d Microsoft CEO Satya Nadella said onstage Tuesday, adding that events like this are about \u201ccoming to grips with the new opportunity.\u201d<\/p>\n<\/div>\n

\n

AI chief Mustafa Suleyman, in an interview with The Verge<\/em>, put it even more bluntly.<\/p>\n<\/div>\n

\n

\u201cThe goal is to prove that we can become one of the top four labs in the world,\u201d Suleyman said. \u201cThere\u2019s three labs that matter, Google DeepMind, OpenAI, and Anthropic. We are not one of them at the moment, and that\u2019s always been my intention. It\u2019s why I came here. I want to build the very best frontier models in the world, fully multimodal, and in order to do that, we have to prove that we can do everything that we need to from the ground up, and we\u2019re not just going to take from others.\u201d<\/p>\n<\/div>\n

\n

One of Microsoft\u2019s first steps at Build was indeed to play catch-up on AI models. Suleyman unveiled MAI-Thinking-1, the company\u2019s first reasoning model, along with six other new models focused on image, voice, transcription, and coding. Microsoft said<\/a> the medium-size MAI-Thinking-1 model, which will likely be marketed to primarily enterprise clients, is \u201cbuilt from scratch for serious math, coding, and real-world enterprise deployment.\u201d Microsoft is years behind both OpenAI and Anthropic here; OpenAI began releasing reasoning models in the fall of 2024. But Suleyman emphasized its performance on benchmarks like coding and its price point, saying it was cheaper than OpenAI equivalents on some tasks \u2014 a big deal in the age of the AI money squeeze<\/a>, which has inspired a lot of complaints with customers.<\/p>\n<\/div>\n

\n

While Microsoft has had years to glean insights from OpenAI, Suleyman made sure to mention that its development did not involve any distillation, meaning that it wasn\u2019t trained using a different company\u2019s AI model. If MAI-Thinking-1 is good, Microsoft clearly doesn\u2019t want people thinking it\u2019s due to the influence of OpenAI.<\/p>\n<\/div>\n

\n

Suleyman told The Verge<\/em> that for Microsoft, \u201cthe pivotal moment was renegotiating our contract with OpenAI. That meant that we were allowed to train models at a larger scale and explicitly pursue superintelligence entirely with our own IP, with our own data, no distillation, training from scratch.\u201d<\/p>\n<\/div>\n

\n

Nadella also highlighted Microsoft\u2019s recently launched AI cybersecurity tool MDASH, which he said brings together 100 AI agents to find exploitable bugs \u201cbetter than any single model.\u201d It was clearly a dig at Claude Mythos Preview, which Anthropic introduced in April to much fear and fanfare<\/a> \u2014 and expanded access to<\/a> just before Build. OpenAI has its own cybersecurity-focused system<\/a> as well, and all three companies will likely use their offerings to jockey for position in the government and enterprise markets they desperately need to court.<\/p>\n<\/div>\n

\n

Microsoft is in a more complex situation with AI agents. The popular open-source platform<\/a> OpenClaw demonstrated a potential path forward for AI agents, and after OpenAI quickly hired<\/a> its creator, Peter Steinberger, Microsoft (among other companies<\/a>) is trying to catch up. One of its key strategies is making OpenClaw work well with Windows. At Build, Nadella said he was very committed to OpenClaw support, and Microsoft employees chatted with developers in the audience about how they were using it.<\/p>\n<\/div>\n

\n

Steinberger himself made a surprise appearance to great audience reaction, taking the stage to boast about how OpenClaw had bolstered its security and earned user trust. \u201cWhat I kept hearing was, \u2018Peter, I love my Claw, but can I use it at work?\u2019\u201d Steinberger said. \u201cYou can totally run OpenClaw inside your company now, and we even made the harness itself a plug-in.\u201d Steinberger said that whether someone trusts Copilot, Codex, or another company\u2019s coding platform, users can now run OpenClaw on top of that via Windows.<\/p>\n<\/div>\n

\n

But Microsoft is also promoting its own separate Copilot \u201csuper app\u201d that integrates OpenClaw-esque agents. A super app is a major focal point for OpenAI right now \u2014 president Greg Brockman is leading development of one that will tie together ChatGPT, the Codex coding platform, and the Atlas web browser. Microsoft\u2019s strategy is similar, bringing together a variety of existing Copilot AI assistants. Its agents, called \u201cAutopilots,\u201d are designed to act as a helpful user interface. Cassidy Williams, GitHub\u2019s senior director of developer advocacy, called Copilot \u201cyour home base for development and operations on your computer,\u201d demonstrating how multiple agents could perform tasks like app-building. (In an extra flourish, Williams demonstrated how she could approve or deny code changes by flashing her computer camera a thumbs-up or thumbs-down.)<\/p>\n<\/div>\n

\n

Autopilots are designed specifically to appeal to business customers \u2014 Nadella called them \u201cautonomous, long-running agents with full enterprise compliance.\u201d The first one Microsoft will offer is \u201cScout,\u201d billed as<\/a> \u201cyour always-on personal agent,\u201d but clients can build and personalize their own. The Autopilot agents should be able to look through an email inbox, join group chats in Teams, check a calendar, and send daily briefings, among other things. Accordingly, employees on stage at Build repeatedly emphasized Copilot\u2019s security tools and guardrails \u2014 obviously aiming to calm enterprise clients who may have heard horror stories<\/a> about tools like OpenClaw.<\/p>\n<\/div>\n

\n

Suleyman made sure to emphasize, again and again, Microsoft\u2019s \u201chumanist superintelligence\u201d as an \u201cAI that prioritizes humanity first\u201d \u2014 part of AI companies\u2019 recent rebrand of AGI<\/a> to make it sound less frightening in an era when people are pushing back against AI<\/a> more than ever before.<\/p>\n<\/div>\n

\n

Nvidia CEO Jensen Huang, another speaker known for working closely with OpenAI, appeared via video call to tout how Nvidia\u2019s RTX Spark chip is fueling Microsoft\u2019s AI agent goals. \u201cI could be traveling and I\u2019m on the phone and I can text my PC \u2026 and it would fire up the tools on the PC,\u201d Huang said. \u201cThe idea that the PC evolved from a personal computer to a personal AI is just really exciting.\u201d<\/p>\n<\/div>\n

\n

Microsoft spent years betting on OpenAI, and in some ways, that\u2019s left it behind in the AI race. But as OpenAI and other competitors turn to enterprise to finally make money, it\u2019s got some obvious advantages. Microsoft already has a substantial client base and, compared with other AI companies, a reputation for safety and security. And like Google<\/a>, it also has deep pockets, considerable computing resources, and a diversified revenue stream, meaning it can take big bets without a ton of risk.<\/p>\n<\/div>\n

\n

Suleyman told The Verge<\/em>, \u201cThere\u2019s a lot of people who are either like chasing startup valuations or about to IPO, so we can operate with a little bit more humility and a little bit more long-term optimization.\u201d He added, \u201cWe\u2019ve got the money to be able to buy Anthropic [models] when we need to. We\u2019ve got the optionality in Azure with 11,000 models, so people can use literally whatever they want whenever they want, but that buys us the time to do it right from the start.\u201d<\/p>\n<\/div>\n

\n

At the same time, there are a lot of unanswered questions here. Microsoft called out a lot of benchmark wins and advancements for its seven new models, but that doesn\u2019t always translate to real-world adoption, and even a new model that pulls ahead for a week or two<\/a> can quickly fall behind. AI super apps are a mostly yet-untested idea. And Microsoft is entering a crowded yet still largely underwhelming AI agent marketplace with a product that we haven\u2019t seen in action. There\u2019s still plenty of room for its promises to fall flat.<\/p>\n<\/div>\n

Follow topics and authors<\/strong> from this story to see more like this in your personalized homepage feed and to receive email updates.<\/span><\/p>\n