CoreWeave Unveils CoreWeave Physical AI Launchables, Taking the Setup Work Out of Robotics and AV Development

BERLIN — October 21, 2026 — CoreWeave, Inc. (Nasdaq: CRWV), The Essential Cloud for AI®, today announced CoreWeave Physical AI Launchables, a catalog of physical AI workloads packaged as Templates, built and tested by CoreWeave engineers, that let robotics, autonomous vehicle (AV), and simulation teams start a run with one command instead of rebuilding their environment first. Unveiled at NVIDIA GTC Berlin, CoreWeave Physical AI Launchables extends CoreWeave’s platform into the engineering layer where physical AI work happens.

The hard part of robotics and AV engineering is rarely perception, motion planning, or control systems. It is the week before the run, when someone has to stand up the environment, configure storage, and wait on the scheduler before the first job can start. CoreWeave engineers built and tested the configurations for the workloads these teams run most often, so teams start from something that already works instead of rebuilding the same environment for every new workload.

CoreWeave Physical AI Launchables builds on a stack CoreWeave already operates: the infrastructure these workloads run on, observability that shows how a run is performing while it runs, and CoreWeave Forge for developing and improving models and agents. CoreWeave Physical AI Launchables extends that stack into the engineering layer, the software that engineers and developers run to build robots, AV, and digital twins. That software comes from NVIDIA and the other companies leading this field. The configurations that make it run here are our own engineering work. Each Template is a tested starting point: teams point it at their own scenes and data, and run it against reserved capacity for a sustained program or as a burst job to test a single idea.

At launch, Templates span five tools: NVIDIA Isaac Sim, CARLA, MuJoCo, NVIDIA AlpaSim, and NVIDIA Cosmos 3, covering data generation, evaluation and grading, and policy training across the physical AI loop. Every tool ships a blank Template, so teams are not limited to the workloads CoreWeave packaged. CoreWeave Launchables also includes the new Cosmos Data Factory agentic skills for generative AI data augmentation across video, image and text, as well as auto-labeling, data curation and retrieval. More Cosmos Data Factory workflows are coming soon as Launchable templates. 

“Robotics and autonomous vehicle teams are doing some of the most demanding engineering in this industry, and they should be spending their time on it," said Richard Ahlfeld, senior vice president of physical AI at CoreWeave. "Physical AI Launchables gives them a tested starting point for the tools they already use, so more of their time goes to the work only they can do."

CoreWeave Physical AI Launchables ships its own agent skills, tuned for high performance on CoreWeave’s cloud and arriving with every Template, for Claude Code, Codex, and other coding agents. An agent knows which settings are worth changing, issues the same command an engineer would, and reads what the run produced separately from whether the job finished. That is the difference between an agent that can start a run and one that can finish a useful one.

“NVIDIA’s reference agent skills for physical AI are built to run on any cloud,” said Nakul Arora, director of product management, NVIDIA. “CoreWeave has built their own version optimized for their AI cloud and ships it with Physical AI Launchables, so these workflows run natively on their infrastructure and developers can focus on the physical AI problem rather than the orchestration underneath.”

For engineering teams, Templates change the level the work happens at:

• Work at the domain level. Teams build with the models, datasets, environments, and evaluations they already understand, while CoreWeave handles the packaging, deployment, and execution beneath.
• Reuse and adapt working templates. Settings or application code can change without rebuilding what runs them. Shared Templates mean the next experiment starts from something that already ran.
• Keep GPUs busy on GPU work. Where the workload permits, Templates finish CPU-only preparation and post-processing before GPU execution begins, and release capacity as soon as the GPU stage ends.
• Run in parallel without collisions. Each Template declares its own resource and concurrency limits, so parallel work does not compete with itself.
• Automate through scripts and agents. Every operation is documented and every response machine-readable. Scripts and coding agents can launch runs, monitor them, and retrieve results.
• Get every result back in one place. Reports, metrics, media, and diagnostics all return through the same interface. Outcomes are easy to assess and failures easy to find.

CoreWeave is also working with software providers in physical AI to run their products on CoreWeave infrastructure, so teams can stay in the tool they already use. Partners include Cadence, Convergent Science, Encord, nTop, PhysicsX, and Voxel51, spanning vision and data tooling through engineering simulation.

“Physical AI demands a new level of performance and scale across the entire AI development lifecycle, from data enrichment through to model inference,” said Eric Landau, Co-founder and CEO, Encord. “By bringing together CoreWeave’s accelerated compute infrastructure and Encord’s AI data platform, we’re making it easier for teams to work with multimodal data and advance their AI model development.” 

"Physical AI puts extraordinary demands on infrastructure, from processing dense multimodal recordings to training and evaluating compute-intensive models,” said Brian Moore, Co-founder and CEO, Voxel51. “By combining Voxel51’s data workflows with CoreWeave’s high-performance compute, we’re helping developers accelerate the data flywheel, turning raw sensor data into better models with every iteration."

To learn more and request a briefing on CoreWeave Physical AI Launchables, visit the website here.

Built on Proven AI Infrastructure

CoreWeave consistently delivers industry-leading performance, demonstrated by record-breaking MLPerf benchmark results in inference and training, and its position as the only AI cloud to earn SemiAnalysis ClusterMAX™ Platinum ranking three times in a row.