Hyperbolic: Building the Liquidity Layer for Compute

Today, Hyperbolic is announcing its latest round including DTC’s participation. This brings the company’s total funding announced to date to $59M. We’re excited to partner with Jasper Zhang and team to help build the Global Compute Grid.

The company today already supports more than 250,000 AI builders, including teams at Modal, Higgsfield, Daytona, HeyGen, Voio, RadixArk, Pika, and many other AI-native companies. The business is scaling quickly, aiming to hit a revenue run-rate of $100M by the end of the year.

The platform they’ve built

Hyperbolic today provides an AI infrastructure platform that unifies compute from hundreds of providers – their platform turns heterogeneous bare-metal compute into a unified AI cloud. It gives AI companies a single platform to buy, manage, and orchestrate compute at scale while helping GPU providers bring capacity to market faster, increase utilization, and maximize yields on their infrastructure.

In June, Hyperbolic launched Forge, the infrastructure layer that manages the full lifecycle of each machine, from provisioning and configuration through security hardening, monitoring, and sanitization after each run. Customers get a consistent experience regardless of which data center supplies the hardware. It also lets Hyperbolic work with providers that lack mature provisioning APIs, adding about 2% virtualization overhead in the company’s internal testing.

Where this goes: a financial layer for compute

Hyperbolic today is useful and differentiated on its own terms, but we think it is also well positioned for what may come next – a spot marketplace to both buy and sell compute, and physical settlement infrastructure for compute contracts.

Turning compute into useful AI work is both a technical challenge and a capital allocation problem. AI infrastructure requires enormous upfront investment, often before demand is secured. Also, compute is now becoming a key line item for every AI company’s financial statement. New financial instruments are helping providers fund the buildout, and marketplaces that aggregate supply to improve price discovery are emerging alongside them.

Over time, greater standardization and price transparency could support forwards, futures, and hedging products. This layer is still early, and compute is not yet truly fungible. But if capacity becomes easier to finance, price, and reallocate, more infrastructure can be built and more of what already exists can be put to productive use.

As that market matures, the companies best positioned to anchor it will be those that can standardize and verify the underlying capacity. That is what Forge does. Hyperbolic is building toward a Global Compute Grid where standardized capacity where suppliers and customers can buy, sell, and resell eligible GPU capacity through a unified platform. Those primitives are a foundation for a secondary market and, eventually, tools to hedge compute prices and manage compute risk through actual physical settlement and delivery of compute.

After earning his PhD at UC Berkeley, Jasper joined Citadel, where he saw how liquid markets form around scarce, valuable resources, and how much they depend on trust. That perspective informs how the company is approaching the broader opportunity here. This month, they filed a detailed public comment with the CFTC on listing compute derivatives, supporting well-regulated markets built on verifiable, transaction-based pricing.

None of this will be built overnight, but the team has already shipped the hard physical layer, and we’re excited to back them as they build more valuable primitives around it.

Welcome to the DTC family, Jasper and team.

‍

‍