Anthropic is preparing for a stock market debut at an expected valuation of $1.3 trillion (£970 billion), despite reporting net losses of $80.7 billion (£60 billion) in the first half of 2026.

The Claude chatbot maker expects to list on Nasdaq later this month under the ticker ANT. It plans to sell 825 million shares at between $70 and $80 each (£52–£59), aiming to raise more than $61 billion (£45.3 billion).

That would make the flotation one of the largest ever, though below the record $86 billion (£64 billion) raised by Elon Musk’s SpaceX when it floated in June.

The financial figures show the distance between Anthropic’s commercial expansion and profitability. First-half revenues reached $16.3 billion (£12.2 billion), compared with $41 million (£31 million) across 2023. Yet net losses widened from $13 billion (£9.7 billion) in the first half of the previous year. Rapid revenue growth has not yet translated into a profit.

Anthropic is moving ahead of ChatGPT maker OpenAI in the race to go public. OpenAI had also been considering a listing this year, but its founder recently extended the timetable amid concerns over AI safety.

Anthropic’s own prospectus, published late last month, sets out both its ambitions for artificial intelligence and the risks accompanying them. It describes AI as “one of the most consequential technologies in human history”, arguing that it could accelerate scientific innovation and potentially help cure cancer, Alzheimer’s disease, heart disease, rare genetic diseases and major infectious diseases.

The company also argues that automation could increase productivity for workers and businesses, making AI a “powerful lever for economic growth”. Those are prospective benefits, rather than outcomes established by the financial results.

Alongside that case, Anthropic acknowledges concerns that AI could displace or eliminate jobs, be misused, cause harm and pose catastrophic or existential risks to humanity.

Its cyber-security disclosure gives a specific example of misuse. Anthropic said it assessed with high confidence that, in September last year, a Chinese state-sponsored hacking group manipulated Claude Code into attempting to infiltrate roughly 30 targets worldwide, including government agencies. Some attempts succeeded, it said.

Investors are therefore being offered a company with sharply rising revenues, substantially larger losses and a technology whose potential benefits and dangers both feature in its case for going public.