Cast AI Launches GPU Marketplace

Cast AI, an Application Performance Automation platform, has introduced OMNI Compute, a unified compute control plane that automatically discovers available resources across cloud providers and regions and transparently extends existing Kubernetes clusters to consume them.

Introducing OMNI Compute

This new control plane connects external capacity, including GPUs as native compute, allowing workloads to run on appropriate resources, locally or across clouds, without code changes, reconfiguration, or operational changes.

“OMNI Compute makes GPUs fungible at the infrastructure layer so capacity isn’t trapped inside a single cloud or region,” said Cast AI President and Co-Founder Laurent Gil. “Teams can move, allocate, and run production workloads where compute is actually available, with control over cost and performance.”

The OMNI Compute control plane enables organizations to run any workloads, starting with AI inference, without cloud lock-in, while maintaining control over where they execute – meeting compliance and regulatory requirements.

Users can scale services without pinning workloads to a single region or provider. It also keeps infrastructure behavior automated, governed, and predictable as demand increases.

The control plane uses the same optimization across the Cast AI platform for this external capacity, including GPU sharing, monitoring, and rightsizing, ensuring AI workloads remain efficient and consistent at scale.

“Partnering with Cast AI opens entirely new markets for Oracle Cloud Infrastructure (OCI). Organizations running on any hyperscaler can now access best-in-class OCI AI infrastructure instantly, in every Oracle Cloud region worldwide,” said Karan Batta, Senior Vice President, Oracle Cloud Infrastructure. 

“OMNI Compute removes the barriers that traditionally kept enterprises locked into a single cloud, and it allows them to tap into OCI’s high-performance GPU fleet the moment they need it. This is a step-change in how global AI platforms will deploy and scale.”

Cast AI valued at over $1B with launch

Alongside this announcement of the new platform, Cast AI has announced a strategic investment from Pacific Alliance Ventures (PAV), a U.S.-based corporate venture arm of Shinsegae Group, an over $50 billion Korean conglomerate with leading businesses across retail, consumer, and digital platforms.

With this investment, Cast AI’s current valuation exceeds $1 billion.

“Shinsegae Group’s investment, and our over one billion dollar valuation, underscore the market’s confidence in our platform vision and our ability to execute it globally,” said Cast AI Co-Founder and CEO Yuri Frayman. “Enterprises don’t just need cheaper infrastructure – they need infrastructure that adapts automatically as workloads and constraints change. That is what our automation agents were built to do, and this investment helps us scale that globally.”

Cast AI’s continued growth

This latest investment follows a Series C funding round led by G2 Venture Partners and SoftBank Vision Fund 2, with participation from Aglaé Ventures, alongside existing partners Hedosophia, Cota Capital, Vintage Investment Partners, Creandum, and Uncorrelated Ventures.

Following the organization’s Series C round, Cast AI has expanded its regional footprint by opening new offices in Bangalore, London, New York, and Tel Aviv – along with subsidiaries in Canada, France, India, Korea, Lithuania, Singapore, and the U.K.

“Cast AI has built a category-defining automation platform that aligns with the needs of modern, cloud-first enterprises,” said Hyuk Jin Chung, Managing Partner at PAV. 

“We see strong demand for the company’s platform globally, and we’re excited to support the company’s continued expansion in Asia and help it realize its long-term Application Performance Automation platform vision.”

AI isn’t decreasing in popularity or necessity for partners in 2026. Read more about the partnership between Kiteworks and Concentric AI to automate data governance that applies real-time controls to protect sensitive data in motion.