SkyPilot has emerged from stealth with $20 million in seed funding to expand its platform for managing artificial intelligence computing infrastructure across multiple clouds, clusters and accelerator types. Lux Capital led the round. Additional investors included Amplify Partners, Coatue Management, Foundation Capital, Race Capital and The House Fund. Technology executives participating in the financing included Databricks CEO Ali Ghodsi, Google Chief Scientist Jeff Dean, Vercel CEO Guillermo Rauch, Replit CEO Amjad Masad, Hugging Face CEO Clem Delangue and dbt Labs CEO Tristan Handy.
SkyPilot helps AI development teams operate computing resources spread across hyperscale cloud providers, specialized neoclouds, Kubernetes clusters and different generations of graphics processing units. Its platform presents these distributed resources through a single control layer.
The company said organizations increasingly need infrastructure from multiple providers because no single cloud or hardware supplier can always provide enough capacity, availability or specialized equipment. Managing these environments separately can slow AI development and make it more difficult to monitor costs, performance and resource availability.
SkyPilot’s technology is designed to make fragmented infrastructure function like one unified AI supercomputer. Teams can use the platform to launch, manage, monitor and scale workloads across providers without rebuilding applications for each environment.
The company was founded by UC Berkeley researchers Zongheng Yang, Zhanghao Wu and Romil Bhardwaj, along with Databricks co-founder Ion Stoica and cloud computing and networking researcher Scott Shenker. The underlying technology was developed through years of AI systems research at Berkeley.
SkyPilot’s open-source project has exceeded 14 million downloads and attracted more than 280 contributors. The technology is used by hundreds of organizations, including Nubank, Abridge, Applied Compute, H Company, Archer and Hippocratic AI.
The newly introduced SkyPilot Platform builds on that open-source foundation with managed capabilities for enterprises operating large GPU fleets. It supports AI development, model training, reinforcement learning, inference, evaluations, agentic workloads and production deployments spanning multiple clusters.
The platform also provides GPU health monitoring, automated remediation, high-availability tools, team management, usage quotas, single sign-on and enterprise governance controls. SkyPilot said it has achieved SOC 2 compliance.
Private preview customers have used the platform to manage more than 10,000 GPUs and support over 200 researchers within individual organizations. Some customers experienced performance improvements of as much as 20 times compared with the open-source version, according to SkyPilot.
The platform is intended to improve GPU utilization and reduce the operational burden of coordinating infrastructure across cloud and hardware providers. It also gives organizations greater flexibility to move workloads without becoming dependent on a single vendor.
SkyPilot refers to the AI systems organizations build around their proprietary information, workflows and industry expertise as custom intelligence. These systems can include specialized agents, applications and models that have been further trained using an organization’s own data.
The company will use the funding to expand product development, engineering and go-to-market operations. It will also continue investing in the SkyPilot open-source ecosystem.
KEY QUOTES:
“Every organization is building custom intelligence around its own data and domains. The challenge is that the AI compute needed to build it is fragmented across clouds.”
“SkyPilot gives frontier AI teams a single platform to manage that infrastructure so they can build custom intelligence faster.”
Zongheng Yang, Co-Founder and CEO of SkyPilot
“Every organization building custom intelligence is running into the same constraint. The compute they need exists, but it’s fragmented across clouds, clusters, and hardware generations.”
“SkyPilot brings that infrastructure together into a single operating environment, helping customers achieve double-digit improvements in GPU utilization while giving frontier AI teams the foundation to build and scale custom intelligence.”
Brandon Reeves, Partner at Lux Capital
