AI firm Anthropic posts US$104bil losses ahead of IPO

SAN JOSE: The AI company Anthropic has accumulated losses of US$104 billion since the start of the year, according to its updated prospectus filed on Monday, reported German Press Agency (dpa).

Meanwhile, its revenue soared to US$33.2 billion in the first nine months of 2026, up from US$2.5 billion in the same period last year.

In the first three quarters of 2025, losses totalled US$25 billion.

Anthropic is expected to go public in the coming weeks, overtaking ChatGPT developer OpenAI, which no longer plans to list its shares this year.

In the updated prospectus, Anthropic included its financial results for the third quarter, reporting a loss of US$23.3 billion on revenue of US$17 billion.

The losses were largely attributable to an accounting adjustment, while the operating loss stood at US$567 million.

Artificial intelligence developers are currently investing heavily in developing their AI models and expanding their computing capacity.

In the last quarter, Anthropic spent around US$5 billion on computing resources.

The updated prospectus also revealed that the company had already budgeted US$895 billion for data centre capacity and infrastructure expansion over the coming years.

Of this amount, just under US$450 billion is due before the end of 2030.

Anthropic's AI assistant, Claude, competes not only with ChatGPT but also with technology heavyweights such as Google, Microsoft and Meta.

Investors' long-standing concerns about whether massive investments in AI can generate sufficient returns have recently resurfaced.

There is also growing debate about the risks posed by the technology following independent hacking attacks carried out using AI software.