Daily AI Briefing

Anthropic priced its IPO at $83 a share, valuing the Claude maker at about $1.5 trillion. Demand blew past the shares available. Now the public-benefit mission, the safety commitments and the push to expand computing capacity all get a lot more scrutiny, with employees and advocates pressing for guardrails. OpenAI’s Sam Altman and Greg Brockman make a similar case: keep building, but add stronger safeguards and monitoring.

Anthropic prices IPO at $83, valuing company at about $1.5 trillion

Anthropic priced its shares at $83, and investor demand ran several times higher than the shares on offer, the Wall Street Journal said. That’s above the planned $70-to-$78 range, and the company sold more shares than it first intended. Trading is expected to start on Nasdaq under the ticker ANT on Thursday.

It gives the public its first chance to buy into a leading AI model developer, and it arrives amid market worries about AI spending and tougher regulatory scrutiny. The prospectus shows $33 billion in revenue and a $104 billion net loss in the first nine months of 2026.

Employees and safety advocates are calling for slower AI development, too. In an opinion column, the Times notes that Anthropic plans to use IPO money to expand computing capacity, and argues the company should publish a concrete safety plan as its influence grows. Meanwhile, Business Wire reports the Center for AI Safety joined a coalition challenging Anthropic’s safety branding ahead of the IPO.

Google expands Gemini 4 with Neon and Hydrogen models

Google introduced Gemini 4 Neon for software, business and scientific work. It has a one-million-token context window, it’s rolling out across Google’s AI products (with paid access in some), and Google says it ranks highly on coding and other benchmarks. Google also launched Hydrogen, a faster, cheaper model that replaces Gemini 3.8 Flash and has broader access than earlier Gemini 4 releases.

In Antigravity, a new Orchestrator shifts work between Hydrogen and Carbon. Google says it approached Carbon’s success rate while using 24% less quota and finishing 19% faster. Separately, Kernel found in a test that check-ins shortened Carbon’s delay before editing a bug fix.

Quote of the Day

[The] only way that we have known how to make this technology safe is to confront reality.

Sam Altman, OpenAI Ceo, NY Times

Leave a comment

Your email address won’t be published. Name, email and comment are required.