According to Woofun AI, Anthropic PBC has disclosed in its draft IPO prospectus that despite having secured long-term compute resources valued at over $600 billion, the company still faces the stark reality of being unable to obtain all the compute capacity it desires. This contradiction highlights the core pain point in the current AI infrastructure landscape: even with massive capital commitments, compute supply remains structurally short.
The prospectus details payment schedules showing that even if Anthropic does not fully utilize the compute capacity covered by its agreements, it must still pay $111 billion to Alphabet (GOOGL.US, GOOG.US), $110 billion to Amazon.com (AMZN.US), and $31 billion to Microsoft (MSFT.US). These major cloud computing giants represent only a portion of the relevant parties. Under agreements running through 2033 and 2034 respectively, Anthropic may also pay up to $86 billion to two suppliers, Nscale and Lambda, to lease compute resources powered by NVIDIA (NVDA.US) chips. Due to a chip leasing agreement between Alphabet (GOOGL.US) and Broadcom (AVGO.US), Anthropic must bear over $160 billion in Google chip leasing costs, pushing its total long-term infrastructure spending potentially beyond the $600 billion mark.
Woofun AI's compiled data shows that the driving force behind this massive investment comes from explosive revenue growth and the uncertainty of compute acquisition. In the six months ended June 30, Anthropic's revenue reached $16.3 billion, compared to just $1.1 billion in the same period last year. Meanwhile, its spending on compute and infrastructure rose to nearly $9 billion, up from $2.3 billion in the first half of 2025. This six-month expenditure total has already exceeded Anthropic's full-year 2025 compute and infrastructure spending of $7.3 billion.
However, the risk disclosure section of the prospectus notes that the acquisition of compute resources remains uncertain, as an increasing number of cloud service giants view Anthropic as a competitor and are reluctant to provide excessive compute resources. Alphabet, Amazon.com, and Microsoft are all selling competing AI models or supporting competing labs, and the companies receiving the most payments from Anthropic are also restricting their compute supply to it.
Additionally, Anthropic is exploring further agreements that could add another $150 billion to its total commitments, bringing the grand total to approximately $750 billion. On paper, the company's losses are quite severe—of its $80.7 billion loss in the first half, $72 billion stemmed from accounting charges related to value changes in convertible notes and other financial instruments issued to raise capital.
The timing of this draft S-1 prospectus is rather unusual, as on the same day Space Exploration Technologies (SPCX.US) released a quarterly report that captured most of the attention, with its superintelligence business segment generating $7.6 billion in revenue through cloud services and compute infrastructure deployment. Nevertheless, for the broader AI industry, Anthropic's disclosure may be more significant—a customer that has already committed $600 billion and is still seeking more resources implies enormous demand for chips, data centers, and cloud compute over the coming years, with Alphabet, Amazon.com, Microsoft, NVIDIA, and Broadcom all positioned to benefit.
Investors will soon gain deeper insight, as Anthropic is preparing for a potentially blockbuster IPO that could value the company at as much as $2 trillion. However, the latest reports indicate that after the company disclosed its third-quarter results, some existing and potential investors have lowered their valuation expectations to the $1.5 trillion range.
Strong AI demand is the root of this urgency, and the compute arms race has evolved from a pure technological contest into a deep entanglement of capital and supply chains. This series of data points indicates that the competitive landscape of the AI industry is undergoing a fundamental shift. Anthropic's predicament is not an isolated phenomenon but a microcosm of the entire industry struggling under compute bottlenecks. When cloud service giants are both suppliers and competitors, the political economy of compute allocation becomes increasingly complex. Anthropic is attempting to hedge risk by locking in long-term contracts, but shifting market dynamics make this strategy full of uncertainties. As the IPO approaches, the market will reassess the company's true value and potential risks. This concerns not only Anthropic's own fate but will also influence capital flows and the technological evolution path of the entire AI ecosystem. In the years ahead, the scarcity of compute resources will continue to drive up industry costs while potentially spawning new business models and technological breakthroughs.
Comments