A new research lab in Kyoto is putting physics at the center of the race to build smarter robots. Nooma Labs, an independent laboratory, states its mission is building "physics foundation action models for robotics" [nooma labs, retrieved 2024]. The tagline is a direct shot across the bow of a sector where billions are being deployed. Competitors like Physical Intelligence and FieldAI have raised hundreds of millions for similar physics-first approaches [The Robot Report, retrieved 2026]. Nooma Labs is entering with a different kind of currency: a specific scientific lineage.
The Cross Labs Continuation
The laboratory is not starting from scratch. Its research program continues and extends the scientific direction previously developed and led by Dr. Olaf Witkowski at Cross Labs [NOOMA Labs, The Space Between Minds, retrieved 2026]. According to an archive post, the program developed at Cross Labs continues at Nooma Labs, which operates with no involvement from the commercial entity Cross Compass Ltd. [About Cross Labs | Historical Research Archive, retrieved 2026]. This positions the new lab as a pure research successor, focused on foundational questions rather than immediate commercial application.
The stated program connects "foundational questions about agency, cognition, and adaptation with new approaches to learning, collective intelligence, and computation" [NOOMA Labs, The Space Between Minds, retrieved 2026]. In a field often dominated by scaling compute and collecting demonstration data, this represents a more theoretical wedge. The bet is that a deeper understanding of physical intelligence and collective adaptation is a prerequisite for robust, general-purpose robot behavior.
The Market Context
The ambition lands in a frothy market. The concept of a "physics foundation model" has become a magnet for venture capital in 2026.
- Physical Intelligence. Raised $400 million, then an additional $600 million, to advance robot foundation models, later open-sourcing its Pi0 model [The Robot Report, retrieved 2026].
- FieldAI. Secured $405 million to scale its 'physics first' foundation models for robots [The Robot Report, retrieved 2026].
These rounds set a high bar for what constitutes a credible player. They also define the competitive landscape Nooma Labs is stepping into,one where capital intensity and rapid scaling are the default modes. Nooma Labs' path, by contrast, appears calibrated for a longer, more research-intensive horizon. The absence of a disclosed funding round or named investors in the public record suggests a deliberately quiet, possibly grant or philanthropy-backed, early phase.
The Open Questions
The laboratory's website is sparse, offering a vision statement but no product details, team bios, or deployment timelines. This opacity is a strategic choice, but it leaves several key questions unanswered. Without a clear commercialization path or partner strategy, it is difficult to assess how the research will translate into tangible robotics advances. The most credible near-term risk is that the lab remains a purely academic endeavor, its insights absorbed by the industry but not directly leveraged by the entity itself.
The counterfactual is that this research-first posture is the point. In a sector chasing near-term demos, a group focused on first principles could uncover the architectural insights that others miss. The continuation of the Cross Labs program provides a thread of continuity and a body of prior work to build upon. The question for observers is whether Nooma Labs can attract the talent and resources to execute on its program before the well-funded incumbents iterate their way to similar insights.
For now, the lab is a statement of intent from Kyoto. It is a bet that the next leap in robotics won't come from more data alone, but from a better model of how physical agents learn and act in a shared world. The next signal to watch will be its first published research or a strategic partnership that brings its theoretical work into contact with real hardware. When Physical Intelligence can point to a $600 million Series B and an open-sourced model, what milestone will define Nooma Labs' first 18 months? [The Robot Report, retrieved 2026].
Sources
- [nooma labs, retrieved 2024] nooma labs homepage | https://www.noomalabs.com/
- [NOOMA Labs, The Space Between Minds, retrieved 2026] NOOMA Labs description | https://noomalabs.org/
- [About Cross Labs | Historical Research Archive, retrieved 2026] Cross Labs archive note | https://www.crosslabs.org/team
- [A New Chapter: Nooma Labs | Cross Labs Archive, 2026-08-31] Nooma Labs announcement | https://www.crosslabs.org/nooma-labs
- [The Robot Report, retrieved 2026] Physical Intelligence raises $400M | https://www.therobotreport.com/physical-intelligence-raises-400m-for-foundation-models-for-robotics/
- [The Robot Report, retrieved 2026] Physical Intelligence raises $600M | https://www.therobotreport.com/physical-intelligence-raises-600m-advance-robot-foundation-models/
- [The Robot Report, retrieved 2026] FieldAI raises $405M | https://www.therobotreport.com/fieldai-raises-405m-scales-physics-first-foundation-models-robots/