Reka and Moonvalley join forces to accelerate physical AI infrastructure

Reka positioning itself as a “foundational intelligence for the physical world” has always hinted at a more embodied vision of AI than pure text-and-image labs. Folding Moonvalley’s team into that effort signals a deliberate bet that the next competitive frontier won’t just be bigger LLMs, but models and infrastructure that reason about space, dynamics, and action. Physical AI is compute-hungry, data-hungry, and deeply tied to hardware; combining research and engineering talent around that stack is how you turn a lab into a platform.([en.prnasia.com](https://en.prnasia.com/all_releases/industry_ITE.shtml?utm_source=openai))

In the broader race to AGI, this move matters because embodied capabilities are likely to be a key differentiator among frontier labs over the next five years. Systems that can plan in 3D, manipulate objects and coordinate fleets of robots are closer to general problem-solving in the real world than chatbots, and they generate rich feedback loops that accelerate learning. If Reka can turn this partnership into a coherent “physical AI OS” — models, simulation, data generation, and deployment infrastructure — it becomes a much more credible rival in a space currently dominated by US hyperscalers and a handful of robotics-first labs.

For incumbents in industrial automation and robotics, the message is that the software stack is consolidating: generic motion planning and perception are being subsumed into foundation models, and the value is shifting to data, deployment, and vertical integration around those models.