The backdrop to one of the biggest IPOs of all time could hardly have been more gripping. Just a few days before the long-anticipated début of Anthropic, the stunning resignation of Nvidia founder and AI icon Jensen Huang shook the technology world to its core. This came as yet another blow to a sector that is already beset by doubts about the enormous sums being invested in AI. Attention was all the greater, then, when Anthropic took to the Nasdaq stage a short while later. Shares in the developer of Claude, which have been traded under the ticker symbol “ANT” since 19 November, were already in high demand from investors ahead of the listing, providing an important sign of life from the AI boom. Anthropic had offered 825 million shares in a range of USD 70 to 78. At the upper end this corresponded to a fully diluted valuation of around USDtn 1.4. This figure reflects the hopes that have since been pinned on the new generation of AI conglomerates.
The man behind Anthropic is Dario Amodei, one of the early pioneers of the generative AI boom. The 43-year-old, his sister Daniela and a number of former OpenAI staff founded the company in 2021. Amodei was Vice President of Research at OpenAI and involved in the development of GPT-2 and GPT-3 as well as groundbreaking research into AI safety. The group left OpenAI at the end of 2020, in part because they wanted to turn the focus more on safety and controllable AI systems. Anthropic was accordingly launched as an AI safety and research company and had already raised USDmn 124 by May 2021. The company is now one of OpenAI’s fiercest rivals. Anthropic is organised as a public benefit corporation, which means it is committed not only to serving the interests of shareholders, but also and expressly to the responsible development of advanced AI. As CEO, Amodei leads the research activities and determines the company’s strategic alignment, while his sister Daniela, the president, is responsible for significant portions of its business operations.
At the heart of the business is the Claude family of models. What originally began as an intelligent chatbot is progressively turning into a platform for businesses, who use the models for software development, data analysis, research, customer service and, increasingly, complex, autonomous processes as well. Claude Code in particular has become a key growth driver: the AI agent plans, writes, tests and corrects software largely independently. Anthropic now talks of “virtual collaborators”, digital employees who can handle entire work processes. Earnings are based primarily on usage, with customers paying according to the number of tokens used and the model class. In the first nine months of 2026 this delivered USDbn 26.3, or 79% of revenue, a further USDbn 6.9 coming from subscriptions. Companies can purchase Claude directly from Anthropic or via cloud platforms. Anthropic is thus combining the classic software subscription with a volume-based platform business.
The pace of growth is impressive: in 2023 revenue amounted to just USDmn 41, in 2024 it was USDmn 386 and in 2025 it had already reached USDbn 4.6. In the first nine months of 2026 earnings exploded to more than USDbn 33, following USDbn 2.6 in the same period the previous year. USDbn 17 was realised in the third quarter alone. According to the prospectus, annualised revenue then reached some USDbn 81 at the end of October. The group is still in the red, however, although in the first nine months operating losses shrank from USDbn 5.22 to USDbn 1.26. The reported net loss of USDbn 104 may seem dramatic, but it can be attributed primarily to valuation effects relating to convertible bonds and other financial instruments. What will sway the capital markets, therefore, is less this accounting figure than the question of whether the group can maintain its tremendous revenue growth while at the same time keeping a grip on costs for computing power.
Source: Anthropic