Nebius launches Physical AI Living Lab, pairing robotics startups with NVIDIA AI technologies

Nebius says it has launched the Physical AI Living Lab, a six-month initiative meant to support robotics startups across the United Kingdom and Europe. The program is positioned as a bridge between AI development and physical deployment, pairing selected teams with NVIDIA AI technologies for “physical AI” use cases, according to the announcement carried by Yahoo Finance.

Physical AI is the broader idea of using AI not just to interpret data, but to operate in the physical world, such as robots learning from sensors and adapting to real environments. In practice, that often requires specialized software that can connect perception, decision-making, and control, along with the compute typically associated with accelerated AI development.

Under the Nebius program, the emphasis is on access, with Nebius saying robotics startups will be able to use NVIDIA’s physical AI software tools during the six-month window. The announcement does not outline specific eligibility rules, application requirements, or the number of startups expected to participate.

Nebius also did not provide in the posted report additional detail on the exact NVIDIA components included in the “physical AI software tools” package, such as which frameworks, model families, or robot-development stacks will be used. It also did not specify any hardware partners, on-site test facilities, or whether participating teams will receive dedicated compute credits or engineering support.

For NVIDIA, the news fits into a wider push to translate accelerated AI into robotics and other real-time, sensor-driven applications. NVIDIA’s own newsroom has recently emphasized a co-designed hardware-and-software approach for its next generation AI systems, describing the “Rubin” platform as a tightly integrated stack intended to improve inference efficiency for large-scale AI workloads.

While that Rubin work is not described as part of the Nebius Living Lab in the announcement, it reinforces NVIDIA’s broader strategy: build the computational platform and software ecosystem together, so developers can move from model development to production deployment more quickly. In the robotics context, that matters because latency, throughput, and model behavior in dynamic environments can be as important as raw accuracy.

Investors should also note what is not disclosed. The Yahoo Finance-linked post does not provide financial terms for Nebius, any expected revenue contribution from the initiative, or performance milestones for participating startups. It also does not name specific robotics companies or research labs that are already signed up.

What to watch next is whether Nebius publishes a clearer program outline, including selection criteria and the scope of NVIDIA tooling delivered to cohorts. Participants’ publicly shared results, such as demonstrators, timelines, and technical writeups, could also become a useful indicator of how quickly physical AI can move from development into repeatable robot behavior in real environments.