Boston-area startup Extropic this week revealed a new kind of AI hardware based on “therdynamic computing.” The company has built its first working chips, which it says could be thousands of times more energy-efficient than current GPUs. The breakthrough comes as giants like Amazon and Apple spend tens of billions on huge, power-hungry data centers. Extropic is betting that the key to scaling AI is not more power, but smarter, radically efficient computation, offering a direct challenge to the massive energy demands of the current AI industry. Artificial intelligence is becoming astronomically expensive, and its primary currency is electricity. As models grow larger, the energy required to train and run them is creating what many call the “AI Energy Wall,” a fundamental barrier to future scaling. U.S. data centers are on track to consume a staggering 426 terawatt-hours by 2030, more than doubling their 2024 usage of 183 TWh, according to independent analysis. This insatiable demand has ignited a frantic, multi-billion-dollar infrastructure arms race among the world’s largest technology companies.

Full report : Startup Extropic has launched a new thermodynamic computing architecture, promising AI chips thousands of times more energy-efficient than current GPUs.

For more see the OODA Company Profile on Extropic.