AI Sector Daily Digest — June 2, 2026

Five stories from the past 24 hours: Anthropic's IPO filing, Alphabet's $80B equity raise, Florida's lawsuit against OpenAI, OpenAI's expansion onto AWS, and a Meta AI chatbot flaw that let hackers take over Instagram accounts.

Anthropic submitted a draft S-1 registration statement to the SEC on Monday, taking the first formal step toward a public offering.1 The company has not set a share count or price range, and says the IPO timing will depend on market conditions. The filing came less than a week after Anthropic closed a $65B Series H at a $965B valuation — co-led by Altimeter, Dragoneer, Greenoaks, Sequoia, Capital Group, Coatue, and D1 Capital Partners — and its revenue run rate has since passed $47B.2 A confidential S-1 lets the company work through financials and risk disclosures privately before a public filing; if it proceeds, the full document will detail financials, legal matters, and voting structure. Rival OpenAI is also expected to file for an IPO, setting up a potential simultaneous listing race between the two largest AI labs.

Google's parent Alphabet announced Monday that it will raise $80B through a stock sale to fund AI infrastructure.3 Berkshire Hathaway will take $10B of that amount directly; the rest goes to general investors. Alphabet said the funds will cover "capital expenditures to scale AI infrastructure and global compute," and that enterprise demand for its AI services is currently running ahead of available supply. The company had already raised its 2026 capex forecast to between $180B and $190B at Google I/O last month. The equity raise is on top of that spending plan, not a replacement for it.4

Florida Attorney General James Uthmeier filed an 83-page lawsuit against OpenAI and CEO Sam Altman on Monday, making Florida the first state to bring such a case.5 The complaint alleges OpenAI ignored internal and external safety warnings to prioritize growth — citing ChatGPT's alleged role in a 2025 mass shooting at Florida State University, links to teen suicides, and broader harms to minors and adults. The AG's office opened a criminal probe into the FSU incident in April. OpenAI has previously denied responsibility for the FSU shooting, and has not yet responded to Monday's filing. The lawsuit lands shortly after OpenAI prevailed in a separate case brought by Elon Musk, where the statute of limitations had passed.

OpenAI said Monday that its frontier models — including GPT-5.5 — and its Codex software engineering agent are now generally available on Amazon Bedrock, accessible to AWS's tens of millions of customers.6 Enterprises can use OpenAI's models through their existing AWS security controls, compliance workflows, and billing, without having to change their infrastructure setup. Codex, which has more than 5 million weekly active users, will let developers write, review, debug, and modernize code inside their existing AWS environments. OpenAI also said that Daybreak — its network security capability including a cybersecurity model and Codex Security — will be added to AWS at a later date. The availability of GPT-5.5 on Bedrock puts OpenAI's newest model directly alongside Anthropic's Claude family on the same cloud platform.

Over the weekend, attackers exploited a flaw in Meta's AI-powered support chatbot to take over Instagram accounts without controlling the victim's linked email.7 The method: a hacker using a VPN to spoof the target's location would open a chat with Meta AI Support, ask it to add a new email to the target's account, receive the verification code in a mailbox they controlled, and hand that code back to the chatbot — which then displayed a "Reset Password" button. The attack bypassed the existing email account entirely. Compromised accounts included the Obama-era White House Instagram handle and the U.S. Space Force chief master sergeant's account. Instagram spokesperson Andy Stone said the issue was fixed by Monday. The total number of affected accounts was not disclosed.