Tech & Appscomparison
Anthropic's S-1: $6.6 million in revenue per employee and 94 pages of risk
The ANT prospectus shows a company growing revenue 13x in a year, spending up to $120.5 billion on compute, and warning at length about its own technology.
A filing with big numbers on both sides
Anthropic's revenue grew 13x in a year while its headcount nearly tripled, and the figures come from the S-1 investor prospectus the company filed publicly on a Monday. [1] The filing comes ahead of a planned IPO later this fall under the ticker symbol ANT, so it matters to anyone who follows AI stocks or builds on Claude. [1]
Goldman Sachs and Morgan Stanley are leading the offering. [1] The same document also devotes 94 pages to risk factors, the largest section of the filing. [2]
Anthropic next to OpenAI and Nvidia
Anthropic released Claude in 2023, a few months after OpenAI's ChatGPT. [1] As of May, Anthropic was valued at $965 billion, while OpenAI had an $852 billion valuation in March. [1]
Anthropic's popularity has grown due to its Claude Code AI coding tool and its Claude Cowork enterprise offering. [1] As of Sept. 30, it had about 5,300 employees. [1]
Revenue per employee gives a second comparison. Anthropic's figure over the past year was more than $6.6 million, up from $1.4 million a year earlier. [1] Nvidia brings in about $7.2 million per employee, by Yahoo Finance calculations. [1]
How Anthropic buys compute
Under the SpaceXAI arrangements, Anthropic agreed to spend an aggregate of up to $120.5 billion over terms extending into 2029. [1] Anthropic expects the capacity to be fully delivered in December 2026, at approximately $3.1 billion in aggregate per month once all of it has been delivered, with different fees during ramping periods. [1]
Anthropic calls its data center leasing initiative Anthropic Direct. [1] It says the initiative addresses its compute needs beyond leasing capacity from cloud providers such as Amazon and Google. [1]
The company says it has moved to a hybrid model. [1] In it, Anthropic buys specialized AI accelerators directly and operates that hardware in dedicated data center facilities leased and operated on its behalf. [1]
What the market did
Anthropic's S-1 showed its revenue growth outpacing rising costs, and chip stocks climbed after the filing. [1] Nvidia, Broadcom, Micron, and Intel shares popped in midday trading, helping lift the Nasdaq Composite 0.8%. [1]
The risk factors
The prospectus states that sufficiently advanced AI systems could, if not developed, deployed, and operated with appropriate safeguards, pose catastrophic or existential risks to humanity. [2] It also warns that AI systems could take harmful actions independently and at scale without human intervention, due to misalignment, errors in objective specification, or self-preserving behaviors. [2]
For AI models that direct physical equipment, the prospectus warns that errors or misuse could cause bodily injury, property damage, or the release of hazardous substances. [2] AI researcher Dave Banerjee noted that Anthropic believes the insurance industry as a whole may not be sufficient to cover the liability it could face in a catastrophic AI-related incident. [1]
Recursive self-improvement
Recursive self-improvement (RSI) describes a state in which humans hand over the research and development of AI models to other AI models. [2] The prospectus warns that risks will increase as AI systems substantially automate AI research and development if they accelerate AI progress beyond the company's ability to evaluate and safeguard new models. [2]
Anthropic calls the recursive dynamic 'closing the loop' and calls it one of the most significant structural advantages in frontier AI. [2] It says failing to 'close the loop' could result in falling progressively further behind competitors who continuously use AI to enhance their own development capabilities. [2]
Former Anthropic researcher Jacob Coxon resigned loudly and went viral, and put curbing RSI at the top of his wish list for the industry. [2] Multiple bills in Congress, including one cosponsored by Bernie Sanders, propose a ban on RSI. [2]
Other items in the filing
Anthropic's S-1 says AI's full economic impact by 2031 could be even greater than the $22.5 trillion figure forecast by the International Data Corporation. [1] The lab has also faced a run-in with the Trump administration over how the government can use the company's AI models. [1]
Anthropic listed a total of 39 terms in the early pages of its prospectus. [1]
What this means for you
If you follow the ANT listing, the filing says the IPO is planned for later this fall under that ticker, with Goldman Sachs and Morgan Stanley leading it. [1] The 94 pages of risk factors are the largest section of the document. [2]
If you use Claude Code or Claude Cowork, those are the products behind the growth in Anthropic's popularity. [1] If you hold chip stocks, Nvidia, Broadcom, Micron, and Intel shares popped after the filing. [1]
If you track AI policy, multiple bills in Congress propose a ban on RSI, which Anthropic itself calls a significant structural advantage in frontier AI. [2]
Sources
Written from the sources above and checked twice before publishing. How we work · Report an error
