Engineering Superintelligence.
Modern machines demand more than any one discipline can explain. A launch vehicle, a fusion reactor, a race car at the edge of performance — each is a system of systems. A change in one part can reshape the requirements of another. Engineers must reason across physics, materials, manufacturing, and mission requirements, while their tools often separate them. The iteration loop is slow. The cost of being wrong can be measured in years.
Apex Ultra is a frontier AI lab building Engineering Superintelligence (ESI). Our aim is AI that can design, simulate, test, and build complex machines and physical systems — reasoning across disciplines, testing its assumptions, and learning from the consequences of its decisions.
We work from the SRI International campus in Menlo Park, with teams across San Francisco, Boston, London, Zurich, and Paris.
We believe intelligence must be stewarded, not unleashed. That it must serve all life. And that what we build will ultimately mirror who we are and what we stand for.
Built by people who have built things.
Founded by two Stanford professors together with championship-winning Formula 1 engineers and Silicon Valley entrepreneurs, and built by researchers and engineers from the world's leading AI labs, aerospace programs, and research departments.
The team we have assembled brings together AI researchers, engineers, and computer scientists whose experience spans fundamental research and the demands of building complex machines in mission critical environments.
Where we trained
- Stanford AI Lab
- Stanford Aerospace
- MIT
- Brown
- UC Berkeley
- Imperial College London
- University of Oxford
- University of Cambridge
Where we built
- Google DeepMind
- OpenAI
- Anthropic
- NASA
- U.S. Space Force
- SpaceX
- Mercedes-AMG Formula 1
- McLaren Formula 1
We keep the team small and senior, and we work closely together across six cities.
Trained where engineering is extreme.
Foundation models for engineering cannot be trained on the open internet. An intelligence grounded in physics that can reason across disciplines has to be tested against the physical world. Wind tunnels, test stands, simulation clusters, and manufacturing lines provide more than data: they connect design decisions to outcomes. So we train and evaluate our models where engineering actually happens, and the results are measurable.
We work with a small number of strategic sandbox partners who operate at the extremes of their fields — Formula 1, hypersonics, aerospace and defense, fusion energy, automotive, robotics — and we train and evaluate our engineering foundation models against their real problems, under their real constraints. Every deployment is isolated. No partner's data or intent is pooled with another's.
Partners receive early access to the systems trained in their environment, and help shape what we build and how we test it — with a direct line into how ESI develops.
Learn moreWe are hiring.
AI researchers. Aerospace, mechanical, and materials engineers. Scientific machine learning experts. Computer scientists. We bring these disciplines together to teach learning systems how the physical world works — and how to build within its constraints.
We believe the intelligence we build will reflect the people who build it — so we pay as much attention to the person as to the résumé.
Menlo Park · San Francisco · Boston · London · Zurich · Paris
Join us