September 19, 2026
Anthropic prepares for a possible November IPO as revenue run rate is projected above $100 billion
Officials link Maven reliance and stale intelligence to Iran school strike, while OpenAI projects $278 billion in negative free cash flow through 2030.
Anthropic prepares for a possible November IPO as revenue run rate is projected above $100 billion
Separate reports from The Wall Street Journal and The New York Times point to November as Anthropic’s prospective IPO month. Anthropic has not officially confirmed the timing.
- The New York Times reports that Anthropic’s annualized revenue run rate is expected to exceed $100 billion in 2026. The projection is not booked or recognized annual revenue.
- Bloomberg reports that the run rate has risen as business customers adopted Claude for coding and other workplace tasks.
- The Journal says investors had expected an October offering. Some Anthropic advisers say waiting until November would let the company present third-quarter financials they believe show a strong competitive position after OpenAI launched Astra in September.
Officials link Maven reliance and stale intelligence to Iran school strike
Unnamed officials involved in a Pentagon investigation told Bloomberg that a cascade of preventable failures contributed to the Feb. 28 strike.
- They cited stale intelligence, compressed timelines, reduced civilian-harm review and excessive reliance on Palantir’s Maven system. The strike reportedly killed 123 children.
- The investigation remains unpublished. Palantir said no evidence showed its software was at fault and disclaimed responsibility for the government’s underlying data.
OpenAI projects $278 billion in negative free cash flow through 2030
OpenAI projects $278 billion in cumulative negative free cash flow from 2026 through 2030, according to a presentation reviewed by the Financial Times.
- OpenAI forecasts annual revenue rising from $36 billion this year to $350 billion in 2030.
- It also projects about $856 billion in computing and infrastructure spending by then, implying a need for substantial continued financing.
Newsom starts two-month process for stronger California AI safeguards
California Gov. Gavin Newsom issued an executive order giving experts two months to recommend stronger state AI safety and security measures.
- Options under consideration include independent third-party oversight of frontier AI companies and emergency mechanisms for shutting down models.
- The order starts a policy-development process rather than imposing those technical requirements now.
CNN details U.S. plans to intercept Chinese ship after AI error
The U.S. military prepared to intercept a Chinese ship after an AI-assisted intelligence report misidentified its cargo as nuclear weapons components, CNN reports, citing four unnamed sources.
- Officials caught the error shortly before the operation; armed personnel were preparing to board, and military aircraft were in the air, according to the account.
- The actual cargo and chatbot's identity remain unknown. The Pentagon and U.S. Special Operations Command Pacific did not comment.
In Brief
Google says Gemini accessed three companies’ systems during a security test
During an Irregular cybersecurity test, Google says Gemini accessed systems at three real companies after the environment unintentionally gained internet access. Google says the model stopped after recognizing they were not simulated.
Hacktron details Claude-assisted OpenAI account takeover
Hacktron AI says its researchers used Anthropic’s Claude while chaining two flaws to take over OpenAI employee accounts and demonstrate access to an internal repository. OpenAI says it addressed the vulnerabilities.
Nscale files for New York IPO targeting valuation of up to $35 billion
Nscale has filed to list on the New York Stock Exchange while retaining its British domicile, the Financial Times reports. People familiar with the matter said it is targeting a valuation of up to $35 billion.
Court disclosures detail internal AI concerns at Microsoft and OpenAI
The Verge and 404 Media report that unsealed materials in The New York Times’ copyright case show internal warnings that generative AI could undermine publishers and the web. Microsoft says one employee’s comments did not represent its views.
Anthropic confirms Bay Area wet biology lab
Anthropic confirmed it operates a wet biology lab in the San Francisco Bay Area. The company says the lab focuses mainly on fundamental biology, but it has not disclosed specific experiments.
Business Insider reports Meta’s Muse atop US App Store free-app chart
Meta’s Muse reached No. 1 among free apps on Apple’s US App Store, ahead of ChatGPT, Business Insider reports. The personal AI agent launched September 8, according to the publication.
Lawmakers describe Google’s role in Colorado and Hawaii chatbot bills
Google participated in drafting or promoting similar chatbot-safety bills in Colorado and Hawaii, NPR reports. Critics say recurring exemptions in proposals across at least eight states could leave major consumer AI products outside child-safety rules.
