Daniela Amodei, co-founder and president of Anthropic, in San Francisco, California on May 9, 2024.David Paul Morris/Bloomberg/Getty Images

Anthropic’s S-1 filing on Monday was confidential, but there’s much we already know about the $965 billion AI juggernaut.

The nearly trillion-dollar startup has gotten so big that it’s hard to hide in the markets the way a growth-spurting elephant is hard to hide in a zoo. 

Back when our AI editor Jeremy Kahn wrote his Fortune cover story on Anthropic in December 2025, the company—founded in 2021 and helmed by siblings Dario and Daniela Amodei—was valued at $183 billion. No longer.

Anthropic’s annualized revenue run rate hit $47 billion in May, and while that’s no guarantee of the future, it is astonishing. Competition for Anthropic secondary shares has been rabid, as I’ve previously reported. There is some speculation that if Anthropic makes it to the public markets before OpenAI, it could dampen demand for its less profitable rival.

A vital caveat: A confidential S-1 is still a maybe. We have no real numbers yet, and timing remains touch-and-go. Anthropic could go public this summer, this fall, or never. 

Nevertheless, it’s very much a signal the company wants to send loud and clear: Anthropic intends to beat OpenAI to the public markets.

Now it’s OpenAI’s move. That company could motor to file and try to leapfrog Anthropic this summer. How SpaceX fares may be a factor, too.

We may already know quite a bit about Anthropic. If the company continues following the IPO path, we will soon learn a lot more. —Allie Garfinkle

Artificial intelligence is expensive, it turns out—so expensive that one of the world’s largest corporations could use some extra dough to pay for it.

Google’s parent company, Alphabet, said Monday that it would raise $80 billion in equity offerings to fund its AI spending plans.

The new effort comprises a $40 billion “at the market” program to sell shares, $30 billion in underwritten public offerings of shares and convertible preferred stock, and a $10 billion investment deal with Berkshire Hathaway.

“AI is driving an expansionary moment for Alphabet,” the company said in a statement. “The company is experiencing strong demand for its AI solutions and services from enterprises and consumers, at levels that are exceeding the company’s available supply. By scaling its investments, the company seeks to expand its foundational infrastructure to support the significant growth opportunity ahead.”

Alphabet plans to spend between $180 billion and $190 billion on capital expenditures this year, double what it spent in 2025. It similarly expects 2027 capex “to significantly increase.” 

Shares were down 2% to $365 in after-hours trading. —AN

It’s all patched up now. But for a while Monday, hackers were able to ask Meta’s AI to give them access to high-profile Instagram accounts…and it worked.

According to a new 404 report, the hackers simply asked Meta’s support bot to change the email address associated with the target account. Presto change-o.

“Over the last several days,” the report reads, “Telegram groups for security researchers and hacking groups have been sharing videos and screenshots of the steps taken to steal an account.”

It was easier than expected. All one needed to do was ask the Meta AI bot to link the target account with a new email address—no two-factor authentication necessary.

In March, Meta said it would deploy AI support to all accounts on Facebook and Instagram. That support would be able to perform critical account maintenance functions including account security and recovery, seemingly without a human in the loop.

The discovery (and fix) coincides with a spate of account takeovers on Meta’s platforms. 

The Obama White House account, dormant since 2017, rose from the dead to post AI-generated images and messages written in Arabic. The Chief Master Sergeant of the Space Force, John Bentivegna, saw his account share anti-American messages and pro-Iranian propaganda. And Sephora’s Instagram account posted Iranian messages as well as graphic sexual content. All were remedied soon after. —AN

—Florida sues OpenAI and Sam Altman. Allegations of deceptive trade practices, negligence, and public nuisance for marketing ChatGPT as safe.

—Hewlett Packard Enterprise shares soar 33%. A billion dollars more in Q2 revenue than analysts expected and sunny forecasts for fiscal 2026 and 2027.

—Binance makes a “super app” push by launching trading for 7,000+ U.S. stocks.

—The EU's cyber agency, ENISA, will reportedly gain access to Anthropic’s powerful Mythos Preview AI model.

—Nvidia says it will work with humanoid robotics firms in Europe, South Korea, and the U.S.—and not just Unitree in China—to build research ‘bots.

—Did you know Salesforce owns a $5 billion stake in Anthropic? Neither did we!

—Quantinuum sweetens its IPO. The Honeywell-backed quantum computing startup now aims to raise nearly $1.5 billion at a valuation of more than $14 billion.