Sam Altman used a Financial Times op-ed to call for a US-led international forum to set AI safety standards and control access to the most advanced models. Separately, FT reporting said OpenAI has held early talks about handing the US government a 5% equity stake, potentially worth tens of billions of dollars.
The proposal reportedly envisions rival labs such as Anthropic, Google and Meta contributing equity to a shared sovereign wealth fund, giving the public a financial stake in the AI boom. Any deal would likely require congressional approval.
What went down?
Altman used an FT op-ed to push US-led global AI safety standards
OpenAI held early talks on giving the US a 5% equity stake
The stake could be worth tens of billions of dollars
The plan floats a shared fund with Anthropic, Google and Meta
Any deal would likely need congressional approval
Upgrade to Absolutely Agentic Premium
Join our growing subscriber community and get access to:
Our daily Agentic Intelligence newsletter curates all the best stories from the top AI newsletters, development labs, news websites and social media into one email.
Ad-free podcast - premium ad free RSS feed that you can put into your podcast app of choice. We publish about 3 hours of video a month and now its available in audio form.
Subscriber area includes long form guides to the most important AI trends affecting content and marketing, along with automations and prompts that will save you a lot of time.
Why does it matter?
Washington's leverage over the labs, from export controls to procurement, is edging towards direct financial ownership.
A public stake answers the question of who benefits from AI, but it binds the largest labs far closer to the state.
Our take
This lands in a week when the state's hand was everywhere. Commerce lifted export controls that had frozen Anthropic's Fable and Mythos models for 18 days, with Anthropic agreeing to give the government pre-release access to future frontier models as a condition of the return. OpenAI's own GPT-5.6 launch stayed limited to around 20 vetted partners at the government's request. Offering equity is the logical next move in that drift, turning informal leverage into a formal seat at the table.
A wealth fund built on AI equity would answer a real grievance about who gets rich from this boom, while handing Washington a financial motive to keep its labs winning. That is the tension worth watching. A government that owns 5% of OpenAI is both referee and shareholder, and other capitals are already circling, with Austria lobbying the EU to host Anthropic. Expect the terms of any deal, rather than the next benchmark, to shape how frontier AI gets built.
And another big thing…Claude Sonnet 5 releases
Anthropic released Claude Sonnet 5 on 30 June, calling it its most agentic Sonnet model yet. It can make plans, uses tools like browsers and terminals, and runs autonomously at a level that required larger, more expensive models only months ago. Its performance sits close to the flagship Opus 4.8, and it is now the default model on Free and Pro plans.
The distinguishing behaviour is follow-through. Testers report it completes multi-step tasks where earlier models stalled halfway and verifies its own output unprompted, in one case writing a reproducing test, fixing the bug, then confirming the bug returned without the fix. The claims come from Anthropic and its launch partners, so independent testing will show how far they hold.
