Zenithon raises $10M Seed

Backed VC, Lunar Ventures, Seraphim Space, MMC Ventures and SOSV led the round, with angel investors also taking part. Valuation was not disclosed. Spending is split evenly between headcount and compute, with a goal of releasing a new model every three months. Staff is expected to grow from 11 to 17 within 3 to 6 months, with a push into San Francisco. CEO Alex Higginbottom said the seed combines two funding rounds from the past year into one, and follow-on extensions remain possible. The product is a set of neural models trained on both simulation output and real experimental results. Zenithon says they can explore a million design points in the time a conventional solver simulates one. That claim is unverified. TechFundingNews names London-based PhysicsX as its main competitor. Co-founder Abetharan Antony is a plasma physicist. Advisers include Charlie Songhurst, Dan Brunner of Commonwealth Fusion Systems and NYU's Zongyi Li. For founders building AI for physical engineering, the round shows investors will fund a domain specialist at seed if compute is a large share of the budget. It also shows what buyers will ask for: proof that a surrogate model matches real experiments, since speed alone will not persuade engineers working on fusion or rocket hardware.