Cast AI is betting that the next major bottleneck in AI infrastructure isn’t algorithms—it’s access to compute. The Application Performance Automation company today unveiled OMNI Compute, a unified compute control plane designed to let enterprises tap into available cloud capacity across providers and regions as if it were native infrastructure. At the same time, Cast AI announced a strategic investment from Pacific Alliance Ventures (PAV), the U.S.-based corporate venture arm of South Korea’s Shinsegae Group, pushing the company’s valuation beyond $1 billion.

The dual announcement underscores Cast AI’s growing influence as enterprises struggle with GPU shortages, cloud lock-in, and rising infrastructure costs driven by AI workloads.

Why OMNI Compute Matters Now

AI adoption has exposed a structural weakness in cloud computing: capacity is fragmented, region-bound, and often unavailable when demand spikes. Enterprises may have Kubernetes clusters running efficiently—until they need GPUs in a region where supply is constrained or pricing becomes prohibitive.

OMNI Compute is Cast AI’s answer. The platform automatically discovers external compute resources—including GPUs—across cloud providers and regions, and extends existing Kubernetes clusters to consume them transparently. No code changes. No reconfiguration. No operational overhaul.

In practice, that means teams can run workloads where compute is actually available, rather than where their cloud contracts or regions force them to be.

“Enterprises don’t just need cheaper infrastructure—they need infrastructure that adapts automatically as workloads and constraints change,” said Yuri Frayman, Co-Founder and CEO of Cast AI. “That is what our automation agents were built to do.”

Turning GPUs Into a Shared, Global Resource

A central promise of OMNI Compute is fungibility—making GPUs interchangeable at the infrastructure layer. Instead of capacity being trapped inside a single hyperscaler or geography, Cast AI allows workloads to move across clouds while remaining governed and predictable.

According to Laurent Gil, Cast AI President and Co-Founder, the goal is to remove artificial barriers that slow AI deployment. “OMNI Compute makes GPUs fungible so capacity isn’t trapped inside a single cloud or region. Teams can run production workloads wherever compute is actually available.”

This is especially relevant for AI inference, the first workload Cast AI is prioritizing with OMNI Compute. Unlike training, inference must run continuously and close to users, making regional shortages and pricing volatility particularly painful.

Oracle Joins the Multi-Cloud Push

One of the first major providers making GPU capacity available through OMNI Compute is Oracle Cloud Infrastructure (OCI). Through the integration, enterprises running on any hyperscaler can instantly access OCI’s GPU infrastructure across Oracle regions worldwide.

“OMNI Compute removes the barriers that traditionally kept enterprises locked into a single cloud,” said Karan Batta, SVP at Oracle Cloud Infrastructure. For Oracle, the partnership opens access to customers who may not otherwise consider OCI—but need GPU capacity now, not after months of procurement.

This dynamic reflects a broader shift in cloud competition: capacity availability and flexibility are becoming as important as services and pricing.

From Optimization to Infrastructure Orchestration

Cast AI has historically focused on continuous optimization—automating rightsizing, cost control, and performance tuning for Kubernetes workloads. OMNI Compute extends that same logic beyond a single cloud boundary.

External capacity brought in through OMNI Compute is automatically optimized using Cast AI’s existing tooling, including GPU sharing, monitoring, and rightsizing. The result is consistent behavior across environments, even as workloads span multiple clouds and regions.

For enterprises, this means scaling AI services without pinning workloads to a single provider, while still meeting compliance, regulatory, and data residency requirements.

Early Enterprise Validation

Customers already running AI in production see OMNI Compute as a practical solution to real-world constraints.

Uniphore, which operates real-time AI workloads globally, says the ability to provision GPUs across clouds without changing application code fundamentally alters how it deploys inference. “Access to reliable, affordable GPU capacity exactly where and when you need it is mission-critical,” said Erik Johnson, VP of Product Management at Uniphore.

Samsung Electronics also sees broader implications. “OMNI Compute’s unified control plane has the potential to change how enterprises like Samsung run AI infrastructure globally,” said Kyotack Tylor Kim, Head of Next Gen Cloud Group at Samsung Electronics.

The $1B Valuation Signal

The strategic investment from Pacific Alliance Ventures, backed by Shinsegae Group, adds more than capital. Shinsegae operates across retail, consumer, and digital platforms—industries increasingly dependent on AI-driven applications at scale.

PAV’s backing follows Cast AI’s recent Series C round led by G2 Venture Partners and SoftBank Vision Fund 2, with participation from Aglaé Ventures and others. Together, the funding validates Cast AI’s thesis that automation—not manual cloud management—will define the next phase of infrastructure operations.

“We see strong global demand for Cast AI’s platform,” said Hyuk Jin Chung, Managing Partner at PAV, pointing to expansion opportunities across Asia.

Global Growth and Market Momentum

Cast AI’s customer roster already includes Akamai, BMW, Cisco, FICO, HuggingFace, NielsenIQ, Swisscom, and more—spanning industries from telecom to automotive to AI-native companies.

Following its Series C, the company has expanded aggressively, opening offices in Bangalore, London, New York, and Tel Aviv, and establishing subsidiaries across Europe, Asia, and North America. That footprint reflects the global nature of the problem Cast AI is addressing: infrastructure scarcity doesn’t respect regional boundaries.

The Bigger Picture

As AI workloads proliferate, enterprises are discovering that cloud-native doesn’t automatically mean cloud-flexible. GPU shortages, regional constraints, and vendor lock-in are becoming strategic risks—not just operational headaches.

OMNI Compute positions Cast AI at the intersection of AI infrastructure, Kubernetes automation, and multi-cloud strategy. By abstracting compute availability from provider boundaries, the company is effectively arguing that the future of AI infrastructure is adaptive, automated, and provider-agnostic.

For marketing and digital leaders watching the AI stack evolve, the message is clear: performance, cost, and scale will increasingly depend on how intelligently infrastructure adapts behind the scenes.

Get in touch with our MarTech Experts.