A new project by Carmen Li aims to build a futures market for compute capacity, allowing enterprises to hedge GPU and inference costs before they spike.
The initiative addresses a growing pain point as companies lock in pricing for inference infrastructure amid volatile demand for accelerator cycles.