Nvidia splashes $10 billion to bet big on Anthropic: valuation gambles behind the compute power “closed-loop”
Recently, there have been reports that Nvidia is in talks to invest up to $10 billion in AI unicorn Anthropic, with the intention of serving as a cornerstone investor in its IPO. This is not a simple financial investment, but rather Nvidia’s big move in the next round of “compute binding.”
First, this $10 billion is meant to lock in major customers. Anthropic’s Claude models rely heavily on Nvidia chips. Previously, the two sides had an agreement in which Anthropic committed to purchasing $30 billion worth of cloud compute powered by Nvidia chips. Now Nvidia is pouring in more capital—effectively using a capital linkage to deeply bind the customer. Even if Anthropic successfully goes public, it would still have to keep purchasing Nvidia chips, forming an ecosystem “closed loop” that is hard to break.
Second, Anthropic’s valuation of as high as $2 trillion has sparked heated debate. From a fundamentals perspective, its annualized revenue had already surged to $65 billion by the end of July—more than 7 times higher than at the end of last year—and in the second quarter it first reported adjusted operating profitability. If we base this on the company’s internal forecast for 2028 revenue of $190–200 billion, a $2 trillion valuation implies a forward price-to-sales multiple of about 10x. For a company whose year-over-year growth is still doubling, this number is not necessarily outlandish, but the key condition is that extremely stringent revenue expectations must be perfectly met, leaving almost no room for error.
Overall, this deal is Nvidia’s defensive battle to consolidate its compute hegemony with capital, and it is also Anthropic’s gamble on sustaining high growth in the future. For investors, closely tracking the revenue guidance in its IPO prospectus is far more practical than obsessing over the $2 trillion valuation figure.
Recently, there have been reports that Nvidia is in talks to invest up to $10 billion in AI unicorn Anthropic, with the intention of serving as a cornerstone investor in its IPO. This is not a simple financial investment, but rather Nvidia’s big move in the next round of “compute binding.”
First, this $10 billion is meant to lock in major customers. Anthropic’s Claude models rely heavily on Nvidia chips. Previously, the two sides had an agreement in which Anthropic committed to purchasing $30 billion worth of cloud compute powered by Nvidia chips. Now Nvidia is pouring in more capital—effectively using a capital linkage to deeply bind the customer. Even if Anthropic successfully goes public, it would still have to keep purchasing Nvidia chips, forming an ecosystem “closed loop” that is hard to break.
Second, Anthropic’s valuation of as high as $2 trillion has sparked heated debate. From a fundamentals perspective, its annualized revenue had already surged to $65 billion by the end of July—more than 7 times higher than at the end of last year—and in the second quarter it first reported adjusted operating profitability. If we base this on the company’s internal forecast for 2028 revenue of $190–200 billion, a $2 trillion valuation implies a forward price-to-sales multiple of about 10x. For a company whose year-over-year growth is still doubling, this number is not necessarily outlandish, but the key condition is that extremely stringent revenue expectations must be perfectly met, leaving almost no room for error.
Overall, this deal is Nvidia’s defensive battle to consolidate its compute hegemony with capital, and it is also Anthropic’s gamble on sustaining high growth in the future. For investors, closely tracking the revenue guidance in its IPO prospectus is far more practical than obsessing over the $2 trillion valuation figure.