Featured coverage image. Source: Editorial desk.
Nebius and NVIDIA have unveiled a new Physical AI Living Lab designed to help robotics startups bridge the gap between simulation and real-world deployment. The six-month program provides emerging companies with access to high-performance AI cloud infrastructure and advanced NVIDIA software tools.
Nebius and NVIDIA have unveiled a new Physical AI Living Lab designed to help robotics startups bridge the gap between simulation and real-world deployment. The six-month program provides emerging companies with access to high-performance AI cloud infrastructure and advanced N...
What is the Physical AI Living Lab? The Physical AI Living Lab is a pioneering six-month incubation initiative launched by AI cloud specialist Nebius in collaboration with NVIDIA. Engineered to empower European and British robotics startups, this program bridges the gap between simulated training environments and physical deployment. It provides early-stage companies with bare-metal high-performance computing, end-to-end synthetic data pipelines, advanced foundational models, and direct access to world-class machine learning engineers.
In an era where "physical AI" is rapidly becoming the next frontier in intelligent automation [2], the convergence of cloud-native orchestration and mechanical execution is creating a paradigm shift. Startups historically struggled with high computational costs and the operational overhead required to test physical machinery. By establishing this state-of-the-art living lab, Nebius and NVIDIA aim to dismantle these financial and technical barriers, enabling developers to build, test, and deploy intelligent agents with unprecedented speed.
What is Physical AI? Physical AI refers to intelligent systems embedded in physical environments, enabling machines like robots, autonomous vehicles, and drones to perceive, reason, and act in the real world [2]. Rather than processing static datasets or text queries behind a computer screen, these systems interact directly with the dynamic variables of human-centric environments.
It's about connecting digital intelligence with physical execution, allowing AI to move beyond data analysis and take tangible actions. By binding cognitive capability with spatial intelligence, physical AI equips machinery with real-time feedback loops. This lets robots handle everything from micro-manipulation in cleanrooms to navigating heavy vehicles across complex industrial terrain.
This technology is transforming industries by giving machines greater awareness of and interaction capabilities with their surroundings. As deep reinforcement learning and spatial computing mature, the demands for advanced computing power have spiked exponentially. The development of physical AI is crucial for creating robots that can operate autonomously and adapt to dynamic environments, a vision championed by NVIDIA CEO Jensen Huang, who has predicted a future with "a billion robots".
What is the Sim-to-Real Gap? The Sim-to-Real (Simulation-to-Reality) gap is the operational challenge where an AI model trained flawlessly in a pristine virtual simulation environment behaves unpredictably or fails when deployed on physical hardware in the messy, variable conditions of the real world.
Overcoming this gap requires a delicate mix of hyper-realistic physics engines, deep domain randomization, and massive-scale computing. Developing sophisticated physical AI requires massive-scale simulation, extensive synthetic data generation, and substantial accelerated computing power – resources that are often beyond the reach of nascent companies [6]. Early-stage startups often exhaust their funding on compute resources or cloud architectural setups, diverting focus from their primary product innovations.
By democratizing access to high-fidelity virtual simulations and deep-learning training frameworks, the Living Lab changes this dynamic. This hands-on experience, coupled with Nebius's AI cloud infrastructure, ensures that startups have the computational power and tooling necessary for effective AI product development [5]. Instead of troubleshooting virtual physics in isolation, founders can validate models using NVIDIA’s end-to-end stack.
Evan Helda, Head of Physical AI at Nebius, underscores the direct value proposition of the program:
"Most robotics teams can build a strong model — the bottleneck is getting the simulation, synthetic data, and compute in place to take it further. The Living Lab is built around that problem: founders get the full NVIDIA physical AI stack on Nebius AI Cloud and direct time with our engineers, so they spend time building robots, not assembling infrastructure."
What is the NVIDIA Isaac Platform? The NVIDIA Isaac platform is an end-to-end development suite tailored for intelligent machines. It combines hardware-accelerated simulation, modular libraries, and world foundation models to streamline spatial training and synthetic data pipelines.
Through this partnership, startups gain access to NVIDIA’s state-of-the-art developer ecosystems. NVIDIA Isaac Sim, built on NVIDIA Omniverse, is a key component, offering an open-source framework for robotics simulation, testing, and synthetic data generation in physically based virtual environments. Furthermore, Omniverse itself is a collection of libraries and microservices designed for developing physical AI applications, including industrial digital twins and robotics simulations, serving as the collaborative engine for unified spatial design.
These tools, combined with NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs, provide the foundational compute power and sophisticated software necessary for advanced AI development [5]. This infrastructure allows startups to run billions of parallel virtual trials, refining motor control policies and perception architectures before deploying them to physical machinery.
What is the Nebius AI Cloud? Nebius AI Cloud is a high-density, bare-metal GPU cloud tailored for complex machine learning projects. Built from the hardware level up, the platform optimizes data orchestration, system performance, and network fabrics to accelerate AI training and inference.
Headquartered in Amsterdam and listed on Nasdaq (NBIS), Nebius is dedicated to delivering a unified platform that spans the entire AI journey, from data and model training to production runtime and deployment [13]. The company has been actively building full-stack infrastructure to support the global AI industry, including GPU clusters and cloud platforms. By bypassing the overhead of standard, non-specialized hyper-scalers, Nebius offers fine-tuned control over raw computation.
Nebius's commitment extends to offering services such as cloud platforms for AI workloads, development team services for autonomous vehicles, and generative AI development. Their collaboration with NVIDIA on the Physical AI Living Lab underscores their strategic focus on empowering AI developers and enterprises worldwide, particularly in sectors like robotics and physical AI. Nebius has also made significant investments in AI infrastructure, including data centers and GPU clusters in various global locations. This worldwide footprint lets builders compute where their data resides, minimizing latency and maximizing training throughput.
The Startup Advantage By integrating specialized cloud-native GPU resources with advanced simulation APIs, the Physical AI Living Lab cuts the developmental runway for robotics startups, shifting focus from hardware configuration to cognitive asset generation.
The first cohort of the Physical AI Living Lab is set to begin in September 2026 [6]. This timeline serves as a milestone for robotics in Europe and the UK. Looking ahead, Nebius and NVIDIA have plans to extend the program to other regions as it grows, aiming to foster global innovation in robotics and physical AI, establishing a standardized ecosystem for physical autonomy worldwide.
Macro Market Dynamics: Physical AI serves as the cognitive layer of modern heavy industry, enabling raw, unstructured physical environments to be actively interpreted and navigated by autonomous systems.
From manufacturing and logistics to healthcare and construction, the ability of machines to perceive, reason, and act in the real world is driving unprecedented efficiency and innovation [1]. This growth is reshaping the competitive landscape. Industry analysts predict substantial growth in the AI robots sector, with projections reaching $124.26 billion by 2034.
This growth is driven by technological milestones across multiple dimensions:
NVIDIA's CEO, Jensen Huang, has emphasized that "NVIDIA's full-stack platform — spanning computing, open models and software frameworks — is the foundation for the robotics industry, uniting a worldwide ecosystem to build the intelligent machines that will power the next generation of factories, logistics, transportation and infrastructure". By opening the doors to this technical framework, Nebius's Living Lab serves as a launchpad for the startups leading this next wave of innovation.
The principal objective of the Nebius Physical AI Living Lab is to accelerate how robotics startups design, train, and deploy autonomous systems. By offering cloud-native computational power alongside NVIDIA’s specialized developer tools, the program helps early-stage companies bypass major financial and engineering obstacles. Founders can focus their energy on refining physical behavior and software design, rather than wasting capital building backend infrastructure.
The Living Lab integrates a comprehensive suite of NVIDIA tools. Key technologies include NVIDIA Isaac Sim for robot simulation and testing, NVIDIA Isaac Lab for training control models, and NVIDIA Omniverse for developing digital twins. Startups also leverage NVIDIA Cosmos world models, NVIDIA OSMO for secure workload orchestration, and the NVIDIA Physical AI Data Factory Blueprint for synthetic data generation. This software runs on Nebius's bare-metal GPU clusters, backed by NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs.
The Sim-to-Real gap is one of the steepest challenges in robotics. The program tackles this by giving startups access to high-fidelity virtual simulations and deep domain randomization libraries. By running millions of simulated physical interactions in parallel, startups can train robust neural networks that adapt smoothly to real-world anomalies—such as shifting lighting, friction, and dynamic human movement—minimizing the need for slow, manual recalibration on physical hardware.
The initial cohort is tailored for early-stage robotics and physical AI startups based in the UK and Europe, starting in September 2026. Because Nebius operates high-performance data centers globally, the compute and support elements are delivered primarily via a secure cloud environment, paired with direct access to physical engineering facilities. As the lab matures, Nebius and NVIDIA plan to expand the program globally to support other regions.
Featured image by Morthy Jameson on Unsplash
AI BlogX is committed to high editorial standards. For time-sensitive or critical topics, please verify claims against original primary sources.
Authoritative and trend-focused coverage across business, sports, entertainment, health, lifestyle, politics, science, and technology.
More Desks
© 2026 AI BlogX. All rights reserved.
Trend-focused editorial workflow
Stories are monitored from trending signals, then processed for accurate summaries, fact-checking, and desk oversight.
Editorial policy
