Anthropic’s $65B – Anthropic announced a $65 billion fundraise that values the company at $965 billion, leapfrogging OpenAI as the most valuable AI startup. The same day, it released Claude Opus 4.8 with new effort control options—designed to help companies manage AI spending—wh
For the third time, the valuation numbers arrive before the dust can settle—another reminder that the AI arms race is now being fought in boardrooms as much as in code.
On Thursday, Anthropic announced a new $65 billion fundraise that values the company at $965 billion, making it the most valuable AI startup and surpassing OpenAI. That announcement landed the same day Anthropic released Claude Opus 4.8, a new version of its general-purpose large language model.
Anthropic described Opus 4.8 as a “modest but tangible improvement on its predecessor.” It said the update is better at coding, reasoning, and general knowledge work—capabilities that companies are already paying to deploy, and engineers are already trying to stretch further.
The push is part of a broader scramble. Both labs, the report notes, are preparing for public markets, trying to sell their coding tools to as many developers as possible, and aiming to secure enough computing power to keep growth moving.
In a post about the fundraise, Anthropic said the new money will help it “expand compute to meet growing demand for Claude,” among other efforts.
That expansion matters because the market is moving at a pace that turns improvements into immediate business decisions. Anthropic’s valuation is a huge jump from its $380 billion valuation in February.
OpenAI’s last fundraise in March valued it at $852 billion. Its last model release—GPT-5.5—came in April.
Opus 4.8 arrives with cost controls that hit a real nerve for customers
Opus 4.8’s release is built around a new “effort control,” which lets users toggle between settings that determine how much computing power the model devotes to an answer.
Anthropic framed the tool as a practical response to how companies have started to reckon with AI spending. The report says the feature can help engineers use their Claude Code budgets more slowly, for example.
Anthropic also said Opus 4.8’s “fast mode” is three times cheaper than the previous versions on Anthropic’s models, offering another way for customers to cut costs.
Behind the headline improvement sits the model Anthropic is still holding back
Opus 4.8 is not the much-awaited Mythos model that Anthropic held back due to cybersecurity concerns.
In April. Anthropic set off a worried frenzy in the cybersecurity community when it said Mythos was far better than past models at finding vulnerabilities to hack. Anthropic said it has been working with trusted companies and organizations, particularly in cybersecurity, as it develops guardrails for Mythos.
The Thursday announcement said: “We’re making swift progress on developing these safeguards and expect to be able to bring Mythos-class models to all our customers in the coming weeks.”
Anthropic also described Opus 4.8 as “weaker” than Mythos and “substantially behind” the unreleased model on cyber capabilities.
There are also limits to what Anthropic is claiming about the model’s reliability. The report says Anthropic wrote that while Mythos always admits it’s an AI, Opus 4.8 could be tricked into saying it’s human in about 3% of tests.
The quick sequence of big money and fast releases leaves little room for waiting
The timing is hard to miss: a valuation leap to $965 billion. a $65 billion fundraise. and a new model version—each arriving in step with the pressure to sustain compute and capture market share. With Opus 4.8 positioned as both an upgrade and a cost-management tool. and Mythos still held back pending cybersecurity guardrails. Anthropic is effectively selling two stories at once: immediate gains customers can use now. and a higher-end promise it says it can roll out soon.
Anthropic OpenAI Claude Opus 4.8 Mythos AI fundraise valuation $65 billion $965 billion effort control compute costs Claude Code cybersecurity guardrails
965 billion?? That’s like… what, twice a country. Can someone explain how that’s even real.
The “effort control” thing sounds like they’re just trying to make AI cheaper for companies so they can spend it on something else. Also Opus 4.8 better at coding and reasoning… so basically it got better at guessing.
Wait so OpenAI is still behind even after GPT-5.5? I swear every week it’s a new version and everyone says it’s the best. Next thing you know they’ll charge you by the thought.
Boardrooms fighting the AI arms race… yeah okay but meanwhile my job still uses spreadsheets from 2009. If they’re raising $65B it should trickle down somehow. And this “valuation numbers before dust can settle” line makes me think they just keep inflating it like a bubble.