Anthropic reported quarterly revenue of nearly $17 billion, a twelvefold increase from the same period last year. Despite this growth, operating costs reached $17.5 billion, resulting in a net loss of $23.3 billion for the quarter. The company cited $5 billion in short-term compute capacity contracts and $2 billion in employee donation matching as primary drivers of the expense spike. This financial position follows a reported compute crunch since late last year, prompting deals with Amazon, Google, SpaceX, chipmakers, and former crypto miners. Anthropic now expects access to more than 3.5 gigawatts of power this year and plans to double that capacity next year. All told, the company’s updated prospectus shows $895 billion in commitments for compute and data center leases. These costs add pressure on the upcoming IPO, which could raise around $100 billion, with up to $3.5 billion expected from Japanese investors.
The scale of these infrastructure commitments highlights the capital intensity required to train and serve large language models at this stage. Securing gigawatts of power and locking in hardware deals represents a long-term financial liability that influences valuation and investor appetite. The discrepancy between soaring revenue and massive losses underscores the current economic reality for AI infrastructure providers.
* Revenue jumped twelvefold to nearly $17 billion
* Net loss reached $23.3 billion for the quarter
* Total compute and lease commitments total $895 billion
