The AI Industry's Big Week: Microsoft Goes Independent, Anthropic Eyes an IPO

AI News

The first week of June 2026 delivered one of the most eventful stretches in recent AI history. Three major storylines converged almost simultaneously: Microsoft made its boldest push yet toward AI self-sufficiency, Anthropic crossed a stunning valuation milestone while filing IPO paperwork, and the White House signed a new executive order reshaping how the government engages with AI safety. Here's what happened and what it means for creators and builders watching this space.

Microsoft Builds Its Own Models — and Means It

For years, Microsoft's role in the AI boom was essentially that of a well-capitalized distributor: invest in OpenAI, wrap those models in Azure and Copilot, and sell access to enterprises. That picture changed significantly at Microsoft Build 2026, held June 2–3 in San Francisco.

Microsoft used its Build 2026 keynote to make its clearest case yet for owning the models beneath its products, rolling out seven new MAI systems across reasoning, coding, image, voice, and transcription. The family was developed entirely in-house. These models were developed by the Microsoft AI Superintelligence Team, led by Mustafa Suleiman, and represent the company's most significant push to build its own AI capabilities — all seven were trained from scratch with zero distillation, meaning none of them relied on outputs or architecture from third-party models such as OpenAI's GPT series.

The headline model is MAI-Thinking-1. The Microsoft AI Superintelligence Team released MAI-Thinking-1 as Microsoft AI's first reasoning model, trained from scratch with zero distillation on enterprise-grade, clean, and commercially licensed data. It's a mid-sized, 35-billion active parameter model with a 256K context window built for high efficiency and performance at a low token cost. It achieved 97% on AIME 25, and sits at 53% on SWE Bench Pro, placing it alongside Opus 4.6 on one of the toughest coding benchmarks.

On the creative-tools side, MAI-Image-2.5 and its Flash variant are Microsoft's first models to serve both text-to-image and image-to-image workloads, ranking #3 and #2 respectively on the Arena AI leaderboard. MAI-Image-2.5 is already live in PowerPoint and reaching OneDrive. For creators using Microsoft's productivity suite, that's a meaningful integration — not just a demo.

Microsoft also announced MAI-Code-1-Flash, its inaugural model that takes written descriptions from people and spits out source code for applications and websites. MAI-Code-1-Flash started rolling out in VS Code through the GitHub Copilot model picker, tuned for fast, low-cost coding.

After refining its models for the needs of consulting firm McKinsey, Microsoft was able to outperform OpenAI's GPT-5.5 with 10 times better cost efficiency, according to Microsoft AI CEO Mustafa Suleyman. That's a striking claim — and one that signals a new phase of competition at the model layer, not just the infrastructure layer.

Microsoft also announced a partnership with Mayo Clinic to jointly develop a new frontier model for health and deploy it in their hospital system — an early signal of how the MAI family is being positioned for specialized, high-stakes domains.

Anthropic Files for an IPO Near the $1 Trillion Mark

The same week Microsoft was announcing models, Anthropic was making moves of an entirely different kind on Wall Street.

Anthropic, the AI lab behind Claude, filed confidentially for an initial public offering on June 1. The company, which is valued at close to $1 trillion, submitted a draft registration statement to the U.S. Securities and Exchange Commission for a proposed initial public offering. The filing came less than a week after Anthropic raised $65 billion in a Series H funding round that pushed its valuation to $965 billion.

The company has experienced explosive growth, announcing in May that its revenue run rate has ballooned to $47 billion, up from $10 billion in annual revenue last year. That kind of trajectory is what turns a filing into a genuine market event.

Anthropic is one of three highly anticipated AI IPOs expected this year, along with OpenAI and SpaceX. Analysts at Wedbush Securities wrote that this "represents an opening of the floodgates for the IPO market, which has been relatively dormant for a few years."

For AI creators, this matters beyond the headlines. Anthropic going public would mean unprecedented financial transparency about the business of frontier AI — what it actually costs to run Claude at scale, which products generate real revenue, and where enterprise AI spending is concentrating. Wall Street would get a much deeper look at the company's business segments — including which products are generating the most revenue — through earnings reports if Anthropic were to go public.

Anthropic is best known for its family of AI models called Claude, which power products like its popular coding assistant, Claude Code. Fintech firm Ramp reported that more businesses used Anthropic than OpenAI for the first time in May. That's a remarkable data point for a company that was widely considered a distant runner-up just a year ago.

A New Executive Order on AI Safety

Rounding out the week, the Trump administration signed a new AI safety executive order that asks AI companies to voluntarily submit their most powerful models for government review up to 30 days before releasing them to the public. The emphasis on voluntary review is the key word — the order steers away from mandatory pre-release testing requirements while still creating a formal channel for government engagement with frontier models.

The executive order's stated rationale is that the United States continues to lead the world in AI "because we refuse to stifle this innovation with overly burdensome regulation," and that the administration has encouraged AI innovation by slashing bureaucratic constraints placed by the prior administration.

The order acknowledges that advanced AI capabilities make the nation stronger, but also "introduce new national security considerations that require coordinated action across executive departments and agencies."

What This All Adds Up To

Take a step back and the picture is striking: in a single week, the industry's top infrastructure partner began competing directly on models, its most safety-focused lab filed for a near-trillion-dollar public offering, and the federal government formalized (however gently) its desire to be in the loop before frontier models ship.

New AI models currently arrive roughly every three days across the industry. The pace isn't slowing — it's compressing. For creators building with these tools, the practical upshot is that the competitive dynamics between platforms are shifting fast, costs are moving, and the models available to you six months from now will look meaningfully different from what you're using today.

The most grounded advice for any creator navigating this: watch which models are getting embedded into the tools you already use (MAI-Image-2.5 in PowerPoint is a good example of how this happens quietly), pay attention to pricing shifts as providers compete more aggressively, and don't build too deep a dependency on any single provider's ecosystem when the landscape is reshaping this quickly.

Sources

ai newsmicrosoftanthropicai policyfoundation models
← Back to Blog