SpaceX pulls off the largest IPO in history

SpaceX priced its IPO at $135 a share Thursday night and opened at $150 on the Nasdaq Friday, closing at $160.95, up 19.2% on its first day under the ticker SPCX. The $75 billion raise values Elon Musk’s combined SpaceX and xAI entity near $1.77 trillion, making it the seventh-largest US company by market cap and placing it above Tesla.

The first-day pop matters less than what comes next. Anthropic and OpenAI are both preparing their own filings, and SpaceX has just shown how much investor money is ready to chase a large AI-linked listing.

Washington forces Anthropic to pull Fable 5 and Mythos 5

Anthropic disabled global access to its Fable 5 and Mythos 5 models on June 12 to comply with a US government export control directive citing national security authorities. The order suspended access for any foreign national, inside or outside the country, including Anthropic’s own foreign national employees, so the company says a worldwide shutoff was the only way to comply. All other Anthropic models stayed live.

Axios reported the Commerce Department moved after another company claimed it had jailbroken Mythos, and that the administration had tried and failed to get Anthropic to pause the launch first. Anthropic is complying while disputing the severity, calling the vulnerabilities minor and discoverable in other models. It appears to be the first time a leading lab has taken a publicly deployed model offline because of federal intervention, a precedent every enterprise relying on a single provider should note.

Anthropic warns its own models may soon be too powerful to control

In a rare public statement, Anthropic cautioned that capabilities in its Mythos tier are strong enough to cause serious damage without the right safeguards, with president Daniela Amodei describing the tier as highly capable in cyber warfare. The company plans to expand Mythos 5 access in phases, starting with biology researchers before a broader cybersecurity program.

When the company building the technology is the one raising the alarm, governance stops being a box to tick.

Zuckerberg admits Meta’s AI restructuring went wrong

In a June 12 internal memo, Mark Zuckerberg acknowledged Meta made mistakes in its aggressive AI restructuring, conceding the company will almost certainly make more. The reset follows May layoffs of roughly 8,000 staff, about 10% of the workforce, plus 7,000 employees reassigned into AI roles.

Zuckerberg pledged no further company-wide layoffs in 2026 and promised more stability, including a July hackathon and reduced manager spans. The admission lands awkwardly because Meta is spending more than ever on AI while conceding it mishandled the people side of the same bet.

ChatGPT crosses one billion monthly active users

OpenAI’s assistant reportedly passed one billion monthly active app users in May, the fastest app ever to hit the mark, per Sensor Tower data reported by Reuters and CNBC. A year ago the same product counted a fraction of that.

For most companies the question is no longer whether staff use these tools. They do. It is whether the company has any approved, governed way to manage work that is already happening.

Mistral chases €3 billion to keep Europe in the race

Mistral AI is in early talks to raise about €3 billion, roughly $3.5 billion, at a valuation near €20 billion, about $23 billion, Bloomberg reported Friday. That nearly doubles the €11.7 billion valuation from its September Series C, which was led by ASML. The raise lands as Mistral pushes into physics AI products for industrial engineers and builds data centers in France and Sweden.

Mistral has raised roughly $4 billion to date and would be valued near $23 billion. Anthropic, by comparison, closed a $65 billion round in May at a $965 billion valuation. Europe’s leading lab is well behind its US rivals on cash, and its bet is that open-weight models will keep it competitive anyway.

Apple ships a Gemini-powered Siri to the world, but not to Europe

At WWDC 2026 Apple unveiled Siri AI, a ground-up rebuild of its assistant running on a custom 1.2-trillion-parameter version of Google’s Gemini, in a licensing deal reported near $1 billion a year. There was a catch. The new Siri will not launch on iPhones or iPads in the EU at iOS 27 release, and it is blocked in China too, both on regulatory grounds.

Apple ties the EU exclusion to the Digital Markets Act, which would require giving rival assistants the same deep system access. Brussels disputes that framing, calling the block Apple’s own business decision rather than a legal necessity. Mac and Vision Pro users in the EU still get it, only iPhone and iPad are affected.

The more interesting part sits under the standoff. Apple will let third-party models including Claude and Gemini run as Siri extensions, so users pick which one answers, making Apple a router between rival AI rather than a single-model shop. The EU split, meanwhile, shows how the bloc’s rules can leave the same product available in one country and missing in the next.