Normal Computing Secures $50M From Samsung Catalyst to Tackle AI Energy Crisis
Normal Computing raises $50 million to solve rising AI chip design costs and develop energy-efficient processors. The startup's software already assists over half of the top 10 global semiconductor companies.
Normal Computing raises $50 million in a funding round led by Samsung Catalyst to address the skyrocketing costs and power demands of artificial intelligence hardware. The startup operates on two fronts by providing AI software that helps semiconductor firms design chips more efficiently and by building its own experimental processors. New investors like Galvanize and ArcTern Ventures join existing backers such as Eric Schmidt’s First Spark Ventures and Micron Ventures to support this dual approach.
The company's software platform is already in use by more than half of the top 10 semiconductor companies by revenue. This tool targets the intense complexity of modern AI chip design, where advanced models pack tens of billions of transistors and cost over $500 million to develop before manufacturing even begins. By reducing errors before the critical "tape-out" phase, Normal Computing helps these massive companies avoid incredibly expensive delays and rework.
Beyond its software, Normal Computing is developing its own "thermodynamic" AI chips that use the natural randomness of physical systems to compute more efficiently than traditional GPUs. Founder Faris Sbahi states that this hardware approach aims to solve the impending AI energy crisis as data centers expect to hit an energy wall around 2030. The former Google Brain and Palantir team has already taped out a prototype chip as part of a growing movement of startups seeking alternatives to conventional, power-hungry AI hardware.