Zenithon raises $10M to build world models for extreme physics

UK-based AI startup Zenithon has secured $10 million in funding to develop groundbreaking world models for extreme physics, with the aim of accelerating the creation of transformative technologies that will shape the 21st century. Backed by notable investors including BACKED, Lunar, Seraphim, MMC, and SOSV, as well as founders and directors from hyperscalers, Zenithon is set to revolutionize industries such as fusion reactors, rocket technology, and semiconductor manufacturing.

The company's primary focus is on building AI models that can accurately predict the behavior of intricate physical systems and rapidly evaluate the outcomes of various designs. Zenithon argues that as these technologies become increasingly critical to shaping the future, engineers must possess the knowledge to control complex physical phenomena.

However, traditional tools often fall short when dealing with the time-consuming nature of simulations and experiments, limiting the scope of design exploration for even the most skilled teams. By leveraging its advanced world models, Zenithon claims its technology can evaluate a million design points in the time it takes to simulate a single scenario, while simultaneously learning from real-world experiments.

This capability will empower engineers to venture beyond conventional designs, uncovering new levels of performance and continuously refining the operation of their machines. Co-founded by Alex Higginbottom and Abetharan Antony, Zenithon boasts a team of researchers with notable accomplishments in machine learning for physics, alongside experts from leading laboratories across the globe.

The advisory board includes prominent figures such as Charlie Songhurst, a Board Director at Meta and former Head of Strategy at Microsoft; Dan Brunner, co-founder and former CTO of Commonwealth Fusion Systems, which has raised approximately $3 billion; and Zongyi Li, an Assistant Professor at NYU. The recently acquired funds will be allocated between talent acquisition and computational resources, enabling Zenithon to launch new models every three months and expand its workforce in San Francisco and across the United States.