Anthropic reports operating loss and massive compute commitments ahead of IPO
The AI developer's latest filing reveals a swing to losses and nearly $900 billion in future spending commitments as it prepares for a Nasdaq debut.
The short version
- Anthropic's revenue jumped 47% to $17 billion in the third quarter, more than twelve times its sales from a year ago.
- The company reported an operating loss of $567 million, down from a $974 million profit in the second quarter.
- Anthropic disclosed approximately $895.4 billion in purchase and lease commitments, mostly for computing capacity.
- About $70 billion of those commitments fall due within the next 12 months.
- Investors may have lowered valuation expectations to around $1.5 trillion following concerns over high AI spending.
Why it matters
This disclosure highlights the extreme capital intensity required to build and maintain leading AI models. For investors, it underscores the tension between rapid revenue growth and the massive infrastructure costs needed to sustain that growth.
A swing to operating losses
Anthropic recently filed an updated prospectus with US regulators, revealing a shift in its financial health. While the company saw its revenue jump 47% to $17 billion in the third quarter, it simultaneously swung back into an operating loss. This loss of $567 million stands in sharp contrast to the $974 million profit the company reported in the second quarter.
The company attributed part of this shift to the high cost of securing immediate computing power. When demand for its services outstripped its long-term contracts, Anthropic spent $5 billion on short-term capacity. The filing notes that these short-term arrangements currently cost the company significantly more than its planned long-term deals.
Other factors also impacted the bottom line during this period. Anthropic booked a $2 billion non-cash charge related to a program for matching staff donations. Additionally, the company reported a net loss of $23.3 billion, which was significantly higher than the $12.1 billion loss it recorded a year earlier.
The price of computing power
One of the most striking figures in the new filing is the scale of Anthropic's future obligations. The company disclosed about $895.4 billion in purchase and lease commitments, the vast majority of which are for computing resources. These commitments represent the physical infrastructure, such as chips and data centers, required to run large-scale AI models.
To put these numbers in perspective, the commitments are nearly seven times the amount of cash and short-term investments Anthropic held at the end of September. Of that massive total, about $70 billion is due within the next year, while roughly $450 billion is expected to be paid by the end of 2030. However, some of these obligations, totaling about $214 billion, can be canceled or are subject to specific conditions.
The high cost of these resources is a primary focus for potential investors. As the company prepares for its Nasdaq listing, the sheer scale of these commitments will be a major point of scrutiny. The filing suggests that the infrastructure required to stay competitive in the AI space requires a level of capital expenditure that is difficult to ignore.
Growth and regulatory hurdles
While revenue has grown significantly, the pace of that growth appears to be cooling. Anthropic noted that sales more than doubled between the first and second quarters, but preliminary figures shared in October suggested a slowdown. The company also warned that its crackdown on customers breaking usage rules might reduce overall revenue and growth rates in the future.
The filing estimates that between 5% and 7% of the revenue from the first nine months came from customers violating commercial terms or usage rules. This indicates that Anthropic is taking steps to ensure its platform is used appropriately. However, these enforcement actions could create a headwind for the company's top-line growth as it moves toward its public debut.
The company also touched on the technical progress of its models. Anthropic stated it may soon meet its own threshold for "dramatically accelerated progress attributable to AI." This refers to a process where AI is used to build even more capable AI. The company warned that this specific type of progress could pose a considerable safety risk, though it has not yet labeled any specific model as having crossed that line.
The road to the Nasdaq
Anthropic is currently moving through the final stages of preparing for its stock market debut. The company plans to trade on the Nasdaq under the ticker ANT, though it has not yet announced a specific price range for its shares. Because the prospectus was made public on October 26, the company is positioned to begin its marketing roadshow as early as next week.
Market sentiment regarding the IPO has been influenced by broader trends in the technology sector. Some reports suggest that investors may have lowered their valuation expectations to around $1.5 trillion. This is a decrease from the $2 trillion figure that was floated earlier, largely due to concerns over the high costs of AI spending. Despite this shift, the valuation remains well above the $965 billion set during its private funding round in May.
The company's listing is expected to occur during the week of November 16. Investors will be looking closely at the balance between the company's massive revenue growth and its heavy debt and spending commitments. The upcoming debut will be a significant test of whether the market believes Anthropic's business model can sustain its high infrastructure costs while remaining profitable.
What to watch
- Whether the $1.5 trillion valuation holds steady during the official roadshow.
- How the company manages the $70 billion in computing commitments due over the next 12 months.
Try it
Check the Nasdaq website for the 'ANT' ticker to see if the company has finalized its listing date or price range.
Sources and links
- ReportingEuronews: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by Euronews · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
