Extropic Prepares First Shipment of Thermodynamic Computing Prototype Device

Extropic, the startup co-founded by Guillaume Verdon, known online as "Beff – e/acc," is set to ship its inaugural thermodynamic computing prototype device. The announcement, made by Verdon on social media, signals a significant milestone for the company pioneering a novel approach to computation. "Which startup should get the first thermo prototype device shipment? 👀📦" Verdon posted, hinting at an imminent dispatch.

The device represents Extropic's thermodynamic computing chip, a radical departure from traditional and quantum computing paradigms. Unlike quantum computing, which seeks to minimize thermodynamic fluctuations, Extropic's technology aims to harness these fluctuations to generate "programmable randomness" for complex computations. This approach promises a more energy-efficient method for modeling uncertainty.

Extropic, which secured $14.1 million in seed funding, is positioning its chips as accelerators designed to enhance classical computers. The company's technology is expected to find initial applications in fields requiring probabilistic simulations, such as high-tech trading and medical research. Verdon envisions a future where entire workloads, including large language models, could run on thermodynamic hardware.

Guillaume Verdon, a former Google quantum computing researcher, founded Extropic after becoming disillusioned with the pace and challenges of quantum computing development. He is also a prominent figure in the "effective accelerationism" (e/acc) movement, an ideology advocating for unrestrained technological progress. This philosophy underpins Extropic's rapid development and ambitious goals for its computing hardware.

The company plans to make its first operational chips available to the market later in 2025, with the prototype shipment marking a crucial step towards broader commercialization. Extropic's innovative technology aims to address the escalating demand for processing power in the age of artificial intelligence by embracing, rather than fighting, the fundamental physics of computation.