Anthropic surpasses OpenAI, becoming the most valuable AI startup in the world

On May 28, Anthropic announced that it had successfully raised $65 billion, bringing the company’s valuation to $965 billion and becoming the most valuable AI startup today. Anthropic said the new investment will help increase computing capacity to meet the growing demand of the Claude model, among other goals.

Theo Business Insiderthis is an enormous jump for Anthropic compared to the figure of 380 billion USD in February. Meanwhile, rival OpenAI was valued at 852 billion USD at its most recent capital call in March. This startup also launched its latest AI model GPT-5.5 last month.

“Claude’s advances have driven widespread adoption of the model by the world’s most demanding organizations. This momentum positions Anthropic to lead the next phase of artificial intelligence innovation,” Brad Gerstner, founder and CEO of investment firm Altimeter Capital, told Altimeter Capital. Financial Times.

Also on May 28, Anthropic released Opus 4.8, a new version of the large language model Claude. The company calls this a “modest but noticeable improvement over its predecessor,” with better programming, reasoning, and general knowledge task handling.

With Opus 4.8, Anthropic also introduced Effort control, which allows users to change the amount of computational power the model invests for each answer. Many companies are starting to rethink AI costs, so this feature helps engineers use their budgets more effectively. Anthropic also said that Opus 4.8’s fast mode is 1/3 cheaper than previous versions, providing more options for customers to save costs.

Opus 4.8 is not the Mythos model, but Anthropic is postponing its release due to cybersecurity concerns. The company said that in tests, Mythos was highly skilled at cybersecurity and hacking tasks, even surpassing humans. Last month, the company only granted limited access to a few partners instead of commercializing it. The startup has also partnered with trusted organizations, especially in the field of cybersecurity, to develop protection mechanisms for Mythos.

Anthropic evaluates Opus 4.8 as weaker than Mythos and “significantly lagging behind” in terms of cybersecurity capabilities. Additionally, while Mythos always claims to be AI, Opus 4.8 can be fooled into identifying itself as human in about 3% of tests.

Anthropic was founded in 2021 by a group of leaders and researchers who worked at OpenAI. The company’s products, especially the Claude Code assistant, have become popular over the past year. This startup’s annual revenue is estimated to exceed $30 billion.