Anthropic Accounting Charge Drives 80% Net Loss

Anthropic’s net loss rose sharply to $80.7 billion in the first half of 2026, mainly because of an accounting charge.

According to a report from Reuters on Tuesday (Oct. 20), the news agency reviewed a draft prospectus for the company’s planned initial public offering (IPO), which is when a company first sells its shares to the public.

Anthropic is expected to go public next month. After its latest financial information was released, some investors began considering a lower company value of about $1.2 trillion, the report said.

The artificial intelligence (AI) startup lost much more money in the first half of 2026 than it did during the same period last year, when its loss was $13 billion. The latest loss included a $72 billion accounting charge. This charge was linked to an increase in the estimated value of bonds that could later be turned into Anthropic shares. It was not money the company spent on running its business, Reuters said.

Reuters reported last month that Anthropic had a net loss of $42 billion in 2025. Most of that loss was also linked to a similar accounting charge.

The report said these financial details show how expensive the race to develop AI has become. Investors are paying closer attention to whether the money being spent on data centers and computer chips will produce good results.

Anthropic has promised to spend as much as $518 billion on cloud services, computing power and infrastructure over the next 10 years, the report said.

The company spent nearly $9 billion on computing power and infrastructure in the first half of 2026. That was up from $2.3 billion during the same period a year earlier, according to the draft prospectus.

At an investor meeting earlier this month, Anthropic told people considering investing in its IPO that its revenue for the third quarter rose 45% to $16.8 billion compared with the previous quarter. Reuters reported this information, citing people familiar with the matter.

These figures eased some concerns about the high cost of developing AI systems. However, some investors were disappointed that revenue was growing more slowly from one quarter to the next. Revenue had more than doubled between the first and second quarters.

In other Anthropic news, PYMNTS reported on Wednesday (Oct. 21) on draft AI safety laws that the company has been proposing.

Under the proposed rules, an AI model could create a predictable risk of $900 million in damage and still not meet the requirements for the federal government to step in.

Instead, Anthropic’s draft proposal says the government should stop a highly advanced AI model only if it creates a foreseeable and immediate risk of more than $1 billion in damage, or of killing or seriously injuring more than 50 people. The danger must involve a biological weapon, a cyberattack or an AI model that gets out of its creator’s control.