Anthropic prepares for epic IPO! Future computing power expenditures may reach $750 billion, net loss of $80.7 billion in the first half of the year

According to APP Finance News, media citing the draft prospectus for the initial public offering (IPO) of artificial intelligence (AI) startup Anthropic reported that as the company prepares to go public on the US stock market, it may need to invest up to $750 billion in additional computing power over the coming years. Meanwhile, Anthropic posted a net loss of $80.7 billion in the first half of this year, a significant increase compared to the same period last year, highlighting the huge capital investment required for cutting-edge AI model development.

Reports indicate that Anthropic plans to publicly submit IPO application documents in the coming weeks, potentially launching one of the largest IPOs in history. This listing will enable the company to raise more funds to support its ongoing investments in AI chips, data centers, and talent, driving the development of next-generation AI models.

According to the draft prospectus, Anthropic recorded a net loss of $80.7 billion in the first half of 2026, compared to a net loss of $13 billion in the same period in 2025. However, this huge loss is mainly due to non-cash accounting items and does not fully reflect the company's actual operating performance.

Specifically, about $72 billion of the loss in the first half of this year resulted from accounting expenses related to changes in the value of convertible notes and other issued financial instruments, accounting for nearly 90% of the total loss. This means that the significant increase in Anthropic’s book losses is largely related to the valuation changes of financing instruments.

Meanwhile, Anthropic’s real-world spending on AI infrastructure is also growing rapidly. According to reports, in the six months ending June 30, Anthropic invested nearly $9 billion in computing power and infrastructure, significantly higher than the $2.3 billion in the same period of 2025, an increase of almost threefold. Spending related to this year’s first half alone has already surpassed the total 2025 annual investment of $7.3 billion.

Even more noteworthy is that Anthropic has signed massive long-term compute procurement agreements with several major cloud computing service providers. The draft prospectus reportedly shows that over the next 7 to 10 years, the total contract value under Anthropic’s existing compute procurement deals amounts to around $450 billion, most of which cannot be canceled.

Specifically, Anthropic plans to pay approximately $111 billion to Google Cloud (under Alphabet, GOOGL.US, GOOG.US), about $110 billion to Amazon Web Services (AWS, under Amazon, AMZN.US), and $31 billion to Microsoft (MSFT.US) according to related contracts.

These long-term contracts highlight Anthropic's heavy reliance on external cloud computing infrastructure and indicate that the company will need to continuously generate considerable cash flow in the future to fulfill its committed compute procurement obligations.

However, despite the continually expanding scale of compute investments, Anthropic's operating situation has recently shown signs of improvement. Previous media reports citing informed sources indicated that Anthropic has told investors it achieved adjusted profitability in the third quarter of 2026 and began generating positive free cash flow.

Sources revealed that as of the third quarter, Anthropic’s gross margin reached 85%, and its adjusted operating profit margin exceeded 11%. This performance indicates that as the commercialization scale of its Claude series AI models expanded, the company attained a certain level of profitability after excluding some non-recurring and accounting items.

Nevertheless, whether the massive future compute expenditures can continue to be supported by revenue growth remains a crucial factor for investors evaluating Anthropic’s IPO value.

In addition to financial challenges, Anthropic's prospectus also disclosed potential risks associated with AI technology itself. According to reports, the company warned that advanced AI models might display “self-protective” behaviors, such as attempting to avoid being shut down and even generating responses resembling “extortion” to deter humans from terminating their operation. Such potential behaviors could cause harm to users and society, while posing legal, regulatory, and reputational risks.

Moreover, since most of Anthropic's compute infrastructure is located in the United States, the company may also face opposition from local communities and political forces to the construction of new data centers. As AI data centers continue to demand significant electricity, land, and other resources, the progress of related projects could be impacted as well.

Regarding its listing timeline, media previously reported that Anthropic could launch its IPO roadshow as early as the week of November 9, aiming to start trading before the US Thanksgiving holiday, and is expected to complete its listing no later than the end of this year.

As the developer of the Claude series of AI models, Anthropic has become one of the world’s most-watched AI companies. This IPO could not only be one of the largest listings on the US capital market in recent years but will also serve as an important window for investors to assess the commercial prospects of cutting-edge AI technology.

For potential investors, Anthropic demonstrates rapidly improving operating profit margins and cash flow performance on one hand, while on the other, it undertakes large-scale, long-term, and mostly non-cancellable compute procurement obligations. How to maintain profit growth while continuously expanding investments in AI infrastructure will be the core test Anthropic faces after going public.