SAN FRANCISCO — Anthropic, which until recently was seen as OpenAI’s “less famous” competitor in artificial intelligence, has been on an almost uninterrupted climb in recent months.
The San Francisco company recently picked a fight with the Pentagon over the use of AI in warfare. It launched a powerful AI model, Mythos, which, according to the company, is frighteningly good at finding and exploiting hidden flaws in software. And he also advised Pope Leo XIV in the encyclical released on Monday, in which he warned of the need to protect humanity from the most disruptive effects of AI.
This Thursday, Anthropic crowned this trajectory by officially overtaking OpenAI and becoming the most valuable AI startup in the world. The company said it raised $65 billion in a round that valued the business at $900 billion before the new capital came in — enough to put it ahead of OpenAI’s last valuation of $730 billion. The company also presented its new “top of the line” model, the Claude Opus 4.8, much more advanced than the previous one in terms of code generation.
Continues after advertising
The new investment, led by Greenoaks Capital, Sequoia Capital, Altimeter Capital and Dragoneer Investment Group, practically multiplied Anthropic’s value by 2.5 in relation to the valuation of around US$380 billion made just three months ago.
The new round reinforces how the company has stood out in the development of AI focused on programming, a business that has become an important source of revenue. Since Anthropic turbocharged its code technology in November, hundreds of companies have started paying for the software. According to the company, its “revenue run rate” — the estimated annualized revenue based on the current pace — exceeded US$47 billion this month.
“This funding will help us meet the historic demand we’re seeing, continue at the frontier of research, and bring Claude to more places where work happens,” said Krishna Rao, Anthropic’s chief financial officer.
Anthropic’s speed draws attention even in an already heated environment around AI. Just 62 days ago, OpenAI — which sparked the current boom with the launch of ChatGPT in 2022 — announced that it had raised $122 billion, in a round that valued the company at $730 billion, a figure that took about a decade to build. Founded in 2021, Anthropic passed that mark in half the time.
The company’s momentum has gained even more momentum in recent months, as Anthropic CEO Dario Amodei has become more outspoken about the risks of AI and advocated for greater regulation. This ended up leading to a tough confrontation with the Pentagon over how the technology should be used in war scenarios, in addition to other public controversies.
The rise of Anthropic expands the dispute with OpenAI and Elon Musk’s SpaceX, in a race to reach the stock exchange.
Continues after advertising
Last week, SpaceX released its public offering prospectus and could debut on the market as early as next month. OpenAI also intends to confidentially file the IPO request in the coming weeks. People close to Anthropic say the company is also considering an IPO later this year, although the company does not comment on the matter.
The dispute, however, is far from over. Under pressure from Anthropic, OpenAI cut back on side projects to prioritize its own programming software, Codex. SpaceX reached an agreement to buy Cursor, a code interface that is very popular among developers, and decided to sell part of its computing capacity to Anthropic.
OpenAI declined to comment. Musk did not respond to interview requests.
Continues after advertising
In the most recent round, Anthropic brought on new strategic investors such as Samsung, Micron and SK Hynix — manufacturers of memory, storage and logic chips essential for training and running AI systems. In a post on the company’s blog, Anthropic stated that these partners will help increase the company’s computing capacity as demand for Claude Code grows.
Since it was created, Anthropic has raised more than US$130 billion, according to PitchBook, which tracks startups. Investors include houses such as Capital Group, Menlo Ventures, Iconiq Capital and Lightspeed Venture Partners, as well as technology giants such as Amazon and Google.
The new Claude Opus 4.8 surpasses all other public AI technologies in “vibe coding” — jargon to describe when AI writes software based on commands in colloquial English. According to Rayan Krishnan, CEO of Vals AI, a company that tracks the performance of the latest models, the Opus 4.8 was 10% better in this benchmark compared to Anthropic’s previous model.
Continues after advertising
The new model also outperformed the previous one in mathematics, another front on which key AI technologies have been advancing rapidly.
(O New York Times filed a lawsuit against OpenAI and its partner, Microsoft, accusing the companies of violating copyright when using journalistic content in AI systems. OpenAI and Microsoft deny the allegations.)
c.2026 The New York Times Company
