Odaily reported that Alex Svanevik, chief executive officer of the on-chain analytics platform Nansen, has commented on pricing conditions in artificial intelligence infrastructure. Svanevik said the AI industry’s bubble may burst when companies begin to effectively use Chinese large models. He also noted that the U.S. regulatory environment may restrict that process, while the broader trend remains that Chinese models continue to become more efficient and are able to run on hardware that is not at the cutting edge.
Svanevik also pointed to changes in the compute market. According to his comments, global GPU supply is increasing, including supply from chips that are not made by Nvidia. At the same time, recent declines in H100 and H200 GPU rental prices reflect a change in the supply-demand structure for computing power. He raised the question of how to interpret falling GPU rental prices as a market signal, placing that signal alongside improving model efficiency and expanding compute supply.
In Svanevik’s view, the combination of more efficient Chinese large models, the ability to operate on less advanced hardware, and broader GPU availability is putting AI infrastructure pricing under review. With rental prices for H100 and H200 GPUs moving lower, the AI infrastructure market may be entering a repricing phase. His remarks link enterprise adoption paths, model efficiency, chip supply, and the current state of GPU rental pricing into one discussion about how AI infrastructure is being valued.