Anthropic’s $2 trillion debut puts AI’s public-market appetite to the test
Anthropic started trading on Nasdaq at an implied valuation near $2 trillion, making it the first publicly traded frontier AI company. Its $83-per-share offering raised about $81 billion, Reuters reported from the debut. Shares opened at $111.50, roughly a third above the IPO price, according to coverage of the opening bell in the Times and a breakdown of the first trades from TechCrunch. Then they gave some of that back: Reuters had the stock around $99, while the Times put it about 15% above the offering price later in the day.
So the listing becomes a benchmark for AI businesses, as investors weigh enormous infrastructure costs against returns. Enterprise sales have helped Anthropic pass OpenAI in revenue, according to the Journal’s side-by-side look at the two companies’ finances. OpenAI is trying to rebuild investor confidence ahead of a possible IPO next year. The wrinkle? The broader market fell anyway. A separate market report from the Times pointed to chip-sector pressure, consumer-spending worries and profit-taking — though enthusiasm for AI stocks didn’t go anywhere.
Google’s AI search reshapes its advertising business
Google’s AI summaries and chatbot features are pushing traditional search ads further down the page, and some advertisers are cutting spending or shifting budgets to other platforms, The New York Times reported. Search-ad revenue grew 12% in Google’s latest quarter, short of analysts’ 15% forecast, and made up less than half of Alphabet’s quarterly revenue for the first time.
Advertisers point to higher costs and less predictable placement. Some worry chatbot conversations keep users from clicking through to websites at all. Google sees it differently: it says AI features are driving more searches, and that it’s prioritizing the search experience and testing ad formats before pushing harder on monetization.
Meta’s Muse AI agent reportedly exposed users’ private data
Meta employees say its Muse AI agent has been mixing up conversations across accounts, surfacing private information like messages, purchases and bank balances, according to a report on the mix-ups cited by TechCrunch. One employee said Muse even reminded her about a Signal message, though she didn’t knowingly give it access to the app.
Not a good look for an agent Meta is actively expanding. Muse has passed 10 million downloads since its September launch. Meta hadn’t immediately commented.
Quote of the Day
And to state it unequivocally, we will pace the training and deployment of AI systems, taking as long as is necessary to have confidence that they are safe and will remain under human control
Dario Amodei, Anthropic CEO, TechCrunch
Other Links
- NY Times: Anthropic’s Trillion-Dollar Debut Is Watershed Moment for A.I.’s Risks and Rewards
- WSJ: Why Anthropic Is Hiring a $400,000-a-Year Natural-Gas Supply Manager?
- WSJ: See How Data Centers Became the Growth Engine of America’s Power Grid
- WSJ: Anthropic Pushes the Reset Button for AI
- TechCrunch: Google gives Gemini the ability to use your Mac for you
- Ars Technica: Traumatized teens sue school, cops after Doritos bag triggers AI gun alert
- NY Times: China Is Shut Out of Anthropic’s Blockbuster I.P.O.
- TechCrunch: Anthropic’s IPO could deliver $156B to charity
- NY Times: Mamdani, Musk, Altman, A.I. and Other DealBook Summit Highlights
- Google: Introducing Tool Retrieval in the Gemini API