Anthropic sold $81.3 billion of shares in what The New York Times calls a record IPO, valuing the Claude maker at more than $1.5 trillion, and its first trading day is today. Bloomberg adds the detail that the Japanese tranche raised only about $3 billion, the lower end of its planned range. The rest of the issue is about what happens around that number: chips, regulation, and who is exposed if the bet goes wrong.
Today’s Headlines
A note on sources: The New York Times, Bloomberg, Reuters and The Washington Post all blocked automated access today. Everything below comes from their headlines and the summary lines their feeds publish, not from the full articles, so treat it as a guide to what each piece says, not a substitute for reading it.
The Anthropic IPO
- The pricing. The Times says Anthropic sold $81.3 billion in shares at a valuation above $1.5 trillion, with trading starting Thursday. Bloomberg reports the Japan tranche brought in about $3 billion, at the low end of its range, per a regulatory filing.
- The winners. Bloomberg says several financial firms “bet the house” on Anthropic and are now set for billions of dollars in returns.
- The bubble question. The Washington Post’s headline says the offering “could test fears about an AI bubble,” and a second Post headline says Fed officials worry an AI bust could shake the financial system. Bloomberg also ran a video interview with GFTN’s Mohanty on investor sentiment around the debut.
Nvidia and China
- The Alibaba link. The Times reports that Alibaba, the giant behind the company at the center of what it calls the Nvidia spy saga, used its AI computing technology to aid a lab with links to China’s military.
- Inside the company. The Times says interim CEO Colette Kress used an all-hands meeting to try to put the “indiscretions” of co-founder Jensen Huang behind the company after his resignation. Reuters says Trump referred Huang’s situation briefly in a talk with Xi.
- Alibaba under pressure. Bloomberg says Alibaba faces mounting pressure to show its AI business plan can deliver despite new limits on its access to data center chips.
Rules and Risk
- Washington. Bloomberg reports Trump warned that “robust” regulation of AI would “stifle” the industry, even as the US adds new restrictions on access to advanced AI technology.
- Agents. Singapore’s central bank warned that AI agents could move deposits and other funds faster, raising liquidity and financial-stability risks (Bloomberg). Reuters reports Arizona is investigating OpenAI and other firms over possible rogue agents.
Infrastructure and Sovereignty
- Data center backlash. Ruth Porat says Alphabet and its peers need to invest in jobs where data centers are built, as projects face backlash. In Australia, Bloomberg says CBA is buying its own compute while data centers there face opposition and delays.
- Sovereign AI. The US and Malaysia plan to work on Malaysia’s sovereign AI cloud, with $50 million in US support. In India, Sarvam is selling access to DeepSeek and other Chinese models while championing national AI independence, which drew an online backlash.
- Also on the wire. Bloomberg video interviews with PayPal’s Enrique Lores on agentic commerce and with Porat on Google’s AI race, plus Reuters’ global markets wrapup.
The Throughline
Today’s stories all describe the same exposure from different sides. The Anthropic offering is the largest test yet of whether public markets will price AI companies the way private ones have been priced. A lower-end Japan tranche is a small signal, but it is the kind of detail that matters when the headline number is $1.5 trillion. The Post’s bubble framing and the Fed worry are the other side of the same bet, and they are already in the same morning’s news.
The second thread is that chips and policy are now one subject. Alibaba is squeezed by chip limits at the same moment the Times links its AI computing to a lab tied to China’s military, and Trump is arguing against regulation while the US tightens access to advanced AI. Those two positions can both be true, but they pull in different directions, and the companies caught between them (Nvidia, Alibaba, Sarvam reselling Chinese models) are the ones to watch.
The third thread is agents moving money. Singapore’s regulator and Arizona’s investigators are asking the same question from different angles: what happens when software acts for people at machine speed. PayPal’s CEO calls agentic commerce an opportunity. The regulators are listing the failure modes.
The Bigger Picture
An AI company is now being priced like a utility-scale bet on the future, and it is arriving at the exact moment regulators are starting to ask what a failure would cost. That does not mean a bust is coming. It means the sector has become large enough that its financing, its supply chain and its rules are a financial-stability topic, not just a technology one.
The same pattern shows up outside the US. Malaysia, India and Australia are each trying to secure their own compute and models, and each is running into the same constraints of chips, power and local opposition to data centers. Sovereignty is turning out to be an infrastructure problem before it is a policy one.
What to Watch
- The first trading day. Anthropic shares begin trading today, and the first sessions will show whether the lower-end Japan result was an outlier or an early read.
- Agent oversight. Arizona’s investigation and the Singapore warning suggest regulators are moving on agents before any federal rule exists. Watch whether other states or central banks follow.
- Nvidia leadership. Kress is interim CEO and the China questions are still open, so further disclosures about Alibaba and the Huang situation are likely.