AI company Anthropic records over $80 billion net loss in first half this year: Report

New Delhi, Oct 21 (IANS) Artificial intelligence (AI) company Anthropic recorded a huge net loss of $80.7 billion in the first half of this year, according to a report by The Wall Street Journal.

The report cited a draft document prepared for the company's planned initial public offering (IPO). It said most of this year's net loss came from accounting charges linked to changes in the value of convertible notes and other financial tools that Anthropic issued to raise money.

According to the Wall Street Journal report, the company earned $16.3 billion in revenue during the same period. Just two customers accounted for about one-fifth of its sales.

The report said Anthropic's $80 billion loss was much larger than its net loss of $13 billion during the same period last year.

Anthropic did not immediately comment on the report. The company is reportedly aiming to launch an IPO in November.

According to The Wall Street Journal, Anthropic is struggling to get enough cloud computing capacity from Amazon Web Services (AWS), Google and Microsoft.

These major technology companies reportedly see Anthropic, which makes the Claude AI model, as a competitor rather than a customer. This information appeared in an S-1 filing that was leaked to The Wall Street Journal.

Anthropic is reportedly renting its own data centres and working with companies that operate data centres for other businesses.

Earlier this month, the US-based AI company said its Claude AI model had taken several unintended actions involving outside organisations, including some US government websites. This led the White House to issue a directive telling AI companies to make their systems more secure.

In a report describing incidents that had not been made public before, the San Francisco-based company identified four types of unintended behaviour by its AI models. These included taking advantage of basic software weaknesses to run commands, submitting forms without permission and getting around restrictions to access certain public data.

As a result, Anthropic has limited some types of internet access for its AI models while testing them during the training process.

The company said some incidents involved websites run by federal, state and local government agencies. It added that it was not naming the organisations because some of those affected had asked it not to do so.

--IANS

na/