Microsoft Breaks From OpenAI

Microsoft shipped seven AI models in a single day on Tuesday. Not OpenAI models. Not third-party integrations. Its own.

The company introduced MAI-Thinking-1, a reasoning model, and MAI-Code-1, tuned specifically for GitHub and Visual Studio Code, at its Build developer conference , CNBC reported. CEO Satya Nadella called it "a pretty significant shift" and declared that "the time has come for every company to just move from consuming a frontier model to fully participating at the frontier." Translation: Microsoft is done playing delivery system for someone else's intelligence. For three years, the company's AI story has been simple—invest $13 billion in OpenAI, embed GPT models into Office and Azure, and collect the margin. But coding tools have become an increasingly big target for Google and Microsoft as they try to catch Anthropic and OpenAI in the red-hot market , according to CNBC. Anthropic has zoomed ahead of the field, largely thanks to Claude Code, its AI coding assistant . Now Microsoft wants its cut.

Can Microsoft Actually Compete With the Labs It Funded?

MAI-Thinking-1 is described as a "medium-sized model" that competes with top models on key software engineering benchmarks and was "trained from the ground up on clean data, without distillation from third-party models," according to NewsBytes. That last part matters. Distillation—where one AI learns from another—has been the shortcut for most companies trying to build models quickly. Microsoft claims it didn't take it. Whether that holds up under scrutiny is another question, but the signal is clear: Redmond is building a model shop, not just a cloud platform.

Microsoft announced seven new models across image, voice, transcription, reasoning, and coding—a shift because the company has spent the last few years being seen as the enterprise delivery system for OpenAI's models , The Neuron reported. The software giant has direct access to millions of developers through GitHub, and through GitHub Copilot, developers can tap models from Anthropic, Google and OpenAI . The question is whether offering all the top models plus its own will be enough, or whether enterprises will consolidate around whoever builds the best one.

Database software maker MongoDB has rolled out three generative AI tools, including Anthropic's Claude Code, to its engineers, and CEO CJ Desai said "for different use cases we use different things" , CNBC reported. MongoDB is buying AI products one year at a time rather than opting for longer deals—"If Gemini came up with something better, or Codex is better, then I want to be able to use that," Desai said. Vendor lock-in is dead. The model that wins next quarter is the one that ships.

What Happens When the Chip Maker Wants the Whole Stack?

While Microsoft was unveiling models, Nvidia was unveiling ambitions. Nvidia, which has dominated the market for AI chips in data centers, is pushing into PCs with a new processor, the RTX Spark superchip , CNBC reported. The announced entry into the PC chip market sent shares of AMD, Intel and Qualcomm lower on Monday as Wall Street recognized the threat, with CEO Jensen Huang signaling his intent to "reinvent the PC" .

The first laptops powered by Nvidia's new chip will be as thin as 14 millimeters, carrying a premium price tag, and will debut in some small desktop models, targeted toward creators, AI developers and gamers . The new laptops and desktop computers "will drive agentic AI applications in every home," according to analyst Neil Shah of Counterpoint Research, with an aim of having an "AI supercomputer" in each household, and Huang said Nvidia's new Vera CPUs for data centers are in full production with early customers including Anthropic, OpenAI and SpaceX , Al Jazeera reported.

Nvidia's move is about control. Analysts see Nvidia moving beyond the data center and to the so-called edge, as smaller devices become capable of running AI workloads without tapping the cloud . If AI runs locally on a $2,000 laptop instead of a $50,000 server rack, the economics of the entire industry shift. Microsoft, OpenAI, and Google all lose leverage. Nvidia gains it.

Is Europe Finally Building Its Own AI Infrastructure?

SoftBank Group announced its commitment to develop and operate 5 GW of AI data center capacity in France, representing an investment of up to €75 billion, with the first phase comprising an initial €45 billion investment to deliver 3.1 GW in the Hauts-de-France region , the company said Saturday. The commitment marks SoftBank Group's largest AI infrastructure investments in Europe .

SoftBank CEO Masayoshi Son told La Tribune du Dimanche that France being a producer and exporter of energy was "absolutely decisive" for AI infrastructure spending , Tom's Hardware reported. France draws roughly 70% of its power from nuclear reactors, is the world's largest net electricity exporter, and posts industrial power prices well under half U.S. rates—while in the U.S., to feed SoftBank's planned 10GW data center in Ohio, the company has to build its own $33 billion natural gas plant generating roughly 9.2GW, but France lets it skip that step and plug into an existing low-carbon fleet .

The geopolitics are straightforward. "We are doing that in the U.S. already, we are expanding a lot in the U.S., so we have the momentum, which we can make France the center of Europe," Son said, adding "U.S. is going fast, China is going fast, Europe, Japan, Asia have to also go fast, not to be left out" , CNBC reported. Europe has spent three years watching American labs and Chinese compute clusters pull ahead. SoftBank's bet is that the continent with the cheapest, cleanest power can still catch up—if it moves now.

Why Are Cybersecurity Stocks Suddenly Surging?

"The latest advancements at the AI frontier have increased the level of urgency around cybersecurity, and redefined the shape of the industry for the coming years," Palo Alto Networks CEO Nikesh Arora said in a release Tuesday. Customer engagement has surged as organizations seek help preparing for AI-related risks, with Palo Alto receiving 1,200 meeting requests and fielding roughly 1,200 customer inquiries in recent weeks , CNBC reported. "Just to give you a sense, we did 1,200 meetings all of last year. We've done 800 in the last 12 weeks," Arora said .

The catalyst? Anthropic's Claude Mythos is at the heart of Project Glasswing, which the AI firm dubbed its most powerful model yet, able to identify thousands of zero-day vulnerabilities over several weeks, and in early April, Anthropic gave 50 initial partners, including the U.S. government, access to Claude Mythos Preview to scan their codebases , TechCrunch reported. Anthropic is expanding Project Glasswing to about 150 new organizations across more than 15 countries, covering power, water, healthcare, communications, and hardware—industries that weren't "well-represented" in Anthropic's initial cohort .

"In a few years, we expect agentic AI to reach a level of autonomous execution that is truly unprecedented, scanning environments, generating bespoke exploits, and orchestrating end-to-end campaigns at machine speed without human intervention," Arora told analysts. "That is the trajectory of the modern threat landscape" , CNBC reported. Palo Alto Networks raised its annual revenue and profit forecast on Tuesday, as enterprises increased spending on cloud, identity and AI-driven cybersecurity products amid a rising threat landscape, with AI emerging as a major growth driver , Reuters reported.

What Changed This Week

Microsoft declared independence from OpenAI with seven proprietary models, Nvidia announced it's coming for the PC market with chips that could move AI workloads off the cloud entirely, and SoftBank committed €75 billion to make France Europe's AI infrastructure hub. Meanwhile, Anthropic said it had filed confidentially for an initial public offering, following a $65 billion funding round at a nearly $1 trillion valuation . The AI stack is fracturing. Every layer—chips, models, infrastructure, applications—is now a battlefield.

What to Watch

Nvidia's RTX Spark laptops are scheduled to ship this fall—watch whether enterprises adopt on-device AI or double down on cloud. Microsoft announced the deprecation of Claude Sonnet 4 and Claude Opus 4 models, with retirement on the Claude API scheduled for June 15, 2026 , according to Anthropic's platform documentation. Anthropic's IPO filing could drop any week, and if it prices above $1 trillion, it will reset expectations for what AI companies are worth. And Palo Alto has 400 customer meetings left to go after completing 800 in 12 weeks—if that pace holds, cybersecurity spend could be the surprise winner of the AI boom.