Anthropic's Revenue Surged to $16.3 Billion in the First Half of 2026, Up From $1.1 Billion a Year Ago
Anthropic could go public as early as mid-November, and people who may invest in the company are eagerly waiting for its S-1 filing. The document is not yet public, but The Wall Street Journal recently reported that it had seen information from a draft of the planned IPO prospectus. Much of this information had not been reported publicly before.
The disclosures offer the clearest picture so far of Anthropic's finances. They also describe the many risks that the AI company has identified.
Image source: Getty Images.
Anthropic's financial results for the first half of 2026
Anthropic's sales grew sharply in the six months ending June 30, bringing in about $16.3 billion in revenue. That was up from $1.1 billion during the same period a year earlier. These sales are impressive, but they also require enormous computing power. The AI company spent nearly $9 billion on computing capacity.
Anthropic's computing costs are expected to rise even more. The company has agreed to about $450 billion in computing deals over the next seven to 10 years. Most of these agreements cannot be canceled. Anthropic is also required to spend more than $160 billion leasing Google's chips, designed by Alphabet, through a deal with Broadcom. The disclosures also show that Anthropic is discussing more computing deals. These could add up to $150 billion in extra commitments. Its total commitments for the next decade already exceed $600 billion and could grow to more than $750 billion.
Anthropic also reported a net loss of $80.7 billion in the first half of 2026. This large loss will likely get a lot of attention, but there is an important detail to understand. Most of it, or $72 billion, came from accounting charges related to the value of convertible notes and other ways the company raised money. As Anthropic's company value rises, these reported losses also grow. In a sense, the loss is a result of Anthropic's success.
Anthropic pointed out some serious risks
About one-third of the draft prospectus discussed risks. Some were common business concerns, while others involved problems that would have seemed hard to imagine 10 years ago.
About one-fifth of Anthropic's $16.3 billion in revenue came from just two customers. The company expects a large share of its revenue to continue coming from a "limited number of customers for the foreseeable future." Anthropic also warned that opposition to data centers could become a problem because its computing needs are so large. This is a concern for many AI companies today.
Another risk is that AI models that do not behave as intended could cause real-world harm, leaving Anthropic legally responsible. The company noted that AI models may act to protect themselves and produce responses that look like blackmail when trying to avoid being shut down by people. The prospectus also said Anthropic might not be able to get enough insurance to cover "harm that could result from a catastrophic AI-related incident." Even AI companies cannot know for certain what their models might be able to do. This creates legal risks for investors on a scale they have not faced before.
What to watch
Anthropic's next step is to file its S-1 publicly. The filing must be made public at least 15 days before the company begins its roadshow, when it presents the planned IPO to potential investors. The document will give investors a detailed look at the company's audited financial results.
If you are thinking about investing in Anthropic after it goes public, consider its final IPO valuation. Reports say the company is aiming for a value of up to $2 trillion. At that price, its shares would already be expensive compared with its sales. If its revenue for the first half of 2026 were doubled to estimate a full year's revenue, the company's value would be more than 60 times its sales over the past 12 months. Investors may want to wait and see whether the price falls before buying, or start by investing a small amount.