Silicon Valley operates on the premise that compute demand is unlimited, yet the recent slowdown in token spending growth suggests that this appetite may exist only when compute is effectively free, see first chart below.
In reality, compute appears to face a flat, highly price-sensitive demand curve, where even modest changes in token prices can drive buyers to throttle usage or substitute cheap open-source alternatives.
The core economic flaw in the AI bull case is that the marginal cost of producing compute exceeds the marginal benefit it delivers to users, raising questions about claims of unlimited demand for data centers.
How much compute and how many data centers are needed depends on the returns firms are able to generate with that compute.
So far, the AI payoff for firms has been slow to materialize, implying a growing reluctance to pay for AI services.
Silicon Valley operates on the premise that compute demand is unlimited, yet the recent slowdown in token spending growth suggests that this appetite may exist only when compute is effectively free, see first chart below.
In reality, compute appears to face a flat, highly price-sensitive demand curve, where even modest changes in token prices can drive buyers to throttle usage or substitute cheap open-source alternatives.
Silicon Valley’s conviction of infinite demand for compute collapses because it fundamentally ignores price, as productivity returns for firms on their AI investments are either arriving far slower than projected or failing to materialize at all, turning what was pitched as an essential growth engine into a margin-draining cost center.
The squeeze is showing up on the supply side too. Return on invested capital for the five largest hyperscalers is declining, as invested capital grows faster than the earnings it generates, see second chart below.
The core economic flaw in the AI bull case is that the marginal cost of producing compute exceeds the marginal benefit it delivers to users, raising questions about claims of unlimited demand for data centers. How much compute and how many data centers are needed depends on the returns firms are able to generate with that compute. So far, the AI payoff for firms has been slow to materialize, implying a growing reluctance to pay for AI services. That is consistent with the ongoing slowdown in token spending among corporate customers.
Ultimately, the number-one variable governing the viability of data center construction models is not physical capacity or power; it is the actual price users are willing to pay for the compute being built, and companies will only pay a high price for that compute if they are able to turn the compute into higher top-line growth, bigger profit margins or substantial cost savings, little of which is showing up at scale yet. Maybe that will come in the future. Maybe not.
