Unlocking the World’s AI Compute: Why We Invested in Hyperbolic

Every breakthrough in AI depends on the infrastructure behind it. Yet for many teams, accessing that infrastructure remains a challenge. Finding available GPUs, comparing providers, navigating contracts, and deploying workloads across different environments all take time and resources that could otherwise go toward building.

For infrastructure providers, the challenge is equally pressing: turning investment in GPUs, power, and data centers into productive capacity and consistent revenue.

We believe connecting these two sides represents a significant opportunity in AI infrastructure. That is why Walden Catalyst is leading Hyperbolic’s oversubscribed $30 million Series A1, alongside Dell Technologies Capital, General Catalyst, Vista Equity Partners, Cambium Capital, Salience Capital, and other strategic investors.

“AI Compute is deeply challenged by supply and demand mismatch, making it incredibly difficult to harness for the AI developer,” said Shankar Chandran, Partner at Walden Catalyst. “Moreover, the myriad of options on compute, interconnects, storage, memory, as well as heterogeneity in the software stack compounds the challenge of accessing available capacity. Hyperbolic’s innovative AI Stack has solved this massive complexity and has aggregated AI compute capacity across compute providers, providing AI developers with simple, transparent, and reliable access to AI compute.”

A founder who saw both sides of the problem

Hyperbolic’s founder and CEO, Jasper Zhang, experienced the compute bottleneck firsthand while conducting research at Berkeley AI Research. Limited GPU access constrained what researchers could attempt, even when they had the ideas and expertise to push further.

His subsequent work as a quantitative researcher at Citadel Securities offered another perspective. In financial markets, connecting buyers and sellers, pricing resources effectively, and providing liquidity are essential to making markets work.

Those experiences shaped Hyperbolic’s approach. GPU capacity is distributed across hundreds of providers, each with different hardware configurations, software environments, and commercial arrangements. AI teams must navigate that complexity to find infrastructure suited to their workloads, while providers need ways to reach customers and keep their GPUs productive.

Jasper recognized an opportunity to bring the engineering of cloud infrastructure together with the principles of efficient markets: make compute easier to access and operate, then build the mechanisms that allow it to move more freely between buyers and suppliers.

Bringing fragmented infrastructure together

At the center of Hyperbolic’s platform is Forge, its unified control plane. Forge transforms heterogeneous bare-metal GPU clusters into standardized infrastructure and provides a common interface for provisioning capacity and managing workloads across providers.

For AI teams, this simplifies the work involved in accessing and operating infrastructure. For providers, it creates a path to bring capacity to market and reach more customers.

Hyperbolic supports more than 250,000 AI builders, including teams at Modal, Higgsfield, Daytona, HeyGen, and RadixArk. Its capabilities span infrastructure provisioning, cloud management, managed AI services, and inference, establishing a foundation for a more connected compute ecosystem.

The new funding will advance Forge and expand Hyperbolic’s multi-cloud orchestration solutions, helping customers manage workloads across providers and optimize placement based on cost, performance, and availability.

Why we invested

At Walden Catalyst, we invest in technologies that address fundamental constraints across the AI stack. As investment in compute expands, the ability to put that infrastructure to productive use becomes increasingly valuable.

Hyperbolic addresses this challenge through a combination of cloud engineering, orchestration, and financial markets expertise. Its team brings experience from Google Cloud, Lambda, Lightning AI, VMware, Supermicro, Citadel, and Point72—backgrounds that reflect both the technical and commercial complexity of the opportunity.

We see the potential for Hyperbolic to become an important platform connecting AI companies with global infrastructure supply. By simplifying access and improving utilization, the company can help customers gain greater flexibility while enabling providers to generate more value from their investments.

“The demand for AI compute will continue to grow, but adding more hardware is only part of the answer. The value of that investment depends on how effectively the capacity is put to work and made available to the teams that need it,” said Lip-Bu Tan, Founding Managing Partner at Walden Catalyst and CEO of Intel.

That conviction is central to our investment. More capacity creates the greatest value when the teams that need it can find it, access it, and use it effectively.

Building a more flexible market for compute

For an AI company, reserving GPUs means making a commitment before knowing exactly how its needs will evolve. For an infrastructure provider, investing in capacity means anticipating demand well ahead of time. When those expectations change, customers can be left paying for resources they no longer need, while providers hold capacity they struggle to monetize.

Hyperbolic’s next chapter aims to give both sides greater flexibility.

Through its planned Global Compute Grid, the company intends to enable customers and suppliers to buy, sell, and resell eligible GPU capacity. A team whose requirements change could recover value from an unused reservation, while another could access capacity that would otherwise remain tied up.

Planned spot instances, dynamic pricing, and market-making strategies would help connect supply and demand more efficiently. Over time, these transactions could also create clearer pricing signals, helping participants make more informed decisions about procurement, investment, and capacity planning.

Hyperbolic is also exploring tools to help buyers and suppliers manage compute price risk. Its longer-term ambition is to build the market infrastructure that could make AI compute easier to value, trade, and finance.

Forge provides the technical foundation for this vision. The planned marketplace would build on it, connecting usable infrastructure with mechanisms that help put more of that capacity to work.

Infrastructure for the next generation of AI

Hyperbolic’s long-term vision extends to how compute will be consumed.

A customer could bring a workload, a budget, and a deadline, and the platform would identify and provision the resources required, continuously optimizing as prices, availability, and performance change.

That capability could also serve AI agents. Within defined budgets and policies, agents could discover available compute, acquire capacity, and run workloads through a programmable interface.

Realizing this vision requires infrastructure that can operate across providers and a market that can respond to changing needs. Hyperbolic is building toward both.

We are excited to partner with Jasper and the Hyperbolic team as they make the world’s compute more accessible, more flexible, and more productive.

Learn more at hyperbolic.ai.