Extropic, a startup led by Guillaume Verdon and Trevor McCourt, is developing a novel silicon-based chip that uses thermodynamic fluctuations to perform probabilistic computations. Designed for energy-intensive tasks like Monte Carlo simulations, the chip claims to be 1,000 to 10,000 times more energy-efficient than current hardware. With ambitions to disrupt AI datacenter power consumption, Extropic's approach could mark a major shift in computing technology.

Heating and cooling buildings accounts for 18% of global energy use—but AI could help reduce that. From New York to Stockholm, smart HVAC systems are slashing energy costs and carbon emissions. Yet, concerns about AI's own power consumption and environmental toll highlight the need for balanced solutions.