Liquid Compute has raised $15 million in seed funding to build a marketplace and financial infrastructure platform for buying, selling, pricing and managing AI computing capacity. FirstMark and Chemistry co-led the round, with participation from K8 Capital, Night Capital, TrueBridge, Brainchild Holdings, UFO Holdings and investor Dmitry Balyasny.
The New York-based company, formerly known as Pluto, is developing what it describes as a physical grid for compute alongside a planned regulated financial market that could allow companies and investors to trade contracts tied to the price of AI infrastructure.
Liquid Compute has applications pending with the U.S. Commodity Futures Trading Commission for Designated Contract Market and Derivatives Clearing Organization status.
If approved, those registrations would allow the company to develop regulated cash-settled markets tied to prices established through its underlying compute marketplace.
The company was founded by Ronit Jain and Aarav Patel, who met while studying engineering at the University of California, Berkeley.
Liquid Compute participated in Y Combinator’s Winter 2024 program.
The founders developed the business around the idea that computing capacity increasingly resembles electricity infrastructure more than a conventional commodity such as oil.
AI compute is not completely interchangeable.
Different GPU configurations, data center locations, power environments, network connections and deployment schedules can make one unit of computing capacity substantially different from another.
Compute is also time-sensitive because unused GPU capacity during a particular period cannot necessarily be stored and sold later in the same way as a physical commodity.
Liquid Compute is therefore developing a marketplace intended to match supply and demand across hardware, geography, infrastructure and time.
The company wants to connect otherwise fragmented pools of computing capacity into a more transparent market where AI companies, enterprises and other buyers can locate infrastructure while suppliers can monetize unused capacity.
The $15 million financing will support the buildout of Liquid Compute’s matching and clearing infrastructure while helping the company expand the amount of compute available through its network.
Liquid Compute is also developing pricing and market data around its physical order book.
That data could give buyers, sellers, traders and lenders greater visibility into how computing capacity is priced and how AI infrastructure is financed.
The company sees that information becoming increasingly important as AI infrastructure spending expands and computing capacity becomes a larger component of corporate capital expenditure.
AI model developers and enterprises can face significant uncertainty around future compute requirements and pricing.
GPU infrastructure providers, meanwhile, need to determine how much capacity to build, where to deploy it and how to monetize periods when infrastructure is not fully utilized.
Liquid Compute believes a more standardized marketplace could help both sides manage those challenges.
Above the physical marketplace, the company plans to establish a financial layer through which participants could eventually manage exposure to changes in compute prices.
Liquid Compute compares the opportunity with the development of established financial markets around other strategically important resources.
Its long-term ambition is to become financial infrastructure for computing capacity in a similar way that established exchanges helped create transparent pricing, hedging and risk management markets around energy and other commodities.
The company’s pending CFTC applications are central to that strategy.
Liquid Compute is seeking Designated Contract Market status, which could allow it to operate a regulated derivatives exchange, and Derivatives Clearing Organization status, which could support clearing of transactions conducted through the market.
Regulatory approval has not yet been granted.
If Liquid Compute receives the necessary approvals, the company plans to create cash-settled products linked to pricing generated across its underlying physical compute market.
That could eventually allow companies dependent on GPU infrastructure to hedge future computing costs.
Infrastructure providers could potentially use the same markets to manage revenue exposure, while financial participants could trade contracts linked to changes in compute pricing.
Liquid Compute believes that type of financial infrastructure will become more important as access to GPUs increasingly influences the speed and cost at which companies can develop and deploy AI systems.
The company has already established trading and data licensing partnerships with Susquehanna Predictions, BGC Group and Wintermute.
BGC is working with Liquid Compute as it develops an over-the-counter market for compute.
Those relationships could help Liquid Compute establish institutional participation and market data around an asset class that remains comparatively fragmented.
The company is also hiring across compliance and market operations to support its CFTC applications and future regulated-market ambitions.
FirstMark sees compute infrastructure becoming strategically important to the U.S. economy as AI adoption increases.
The firm believes more transparent pricing and allocation mechanisms could become important not only for technology companies but also for governments seeking visibility into domestic computing capacity.
Chemistry is backing Liquid Compute around the company’s decision to structure compute more like a power market than a traditional fungible commodity market.
Electricity markets have to continuously balance location-specific and time-sensitive supply and demand because electricity production in one place and period is not always directly interchangeable with production somewhere else.
Liquid Compute believes computing capacity increasingly has similar characteristics.
A particular GPU cluster may be useful for one AI workload but unsuitable for another because of chip type, networking, location, availability or other infrastructure requirements.
That makes matching specific supply with specific demand an important part of the market.
Liquid Compute’s physical marketplace is intended to create that matching layer first.
The company then plans to use transaction information generated through the marketplace to establish more transparent pricing benchmarks and financial products.
That sequence could help create a market where compute capacity is not only purchased when needed but increasingly financed, traded and hedged.
The opportunity is growing alongside enormous investment in AI data centers and GPU infrastructure.
As computing capacity becomes a strategic input for AI development, Liquid Compute believes companies will increasingly need market mechanisms similar to those used for other critical infrastructure resources.
The company ultimately wants to create an environment where AI developers can efficiently acquire capacity, infrastructure operators can monetize unused GPUs and financial institutions can price and manage risks associated with compute.
The new seed financing gives Liquid Compute additional capital to build that market structure while pursuing the regulatory approvals needed for its planned financial products.
KEY QUOTES:
“Our goal is to be the regulated financial infrastructure for a new asset class, similar to what NYMEX was to oil, CBOE was to volatility, and Kalshi was to event contracts. Access to compute increasingly determines access to intelligence. We are creating the essential marketplace through which both natural users and financial participants can access, price, and manage that resource.”
Ronit Jain, Co-Founder and CEO of Liquid Compute
“Compute is rapidly becoming strategic infrastructure for the United States. That creates a need for market infrastructure that gives industry and government a transparent view into how capacity is priced, allocated, and financed. Liquid Compute is building a regulated physical and financial layer for the American compute economy.”
Adam Nelson, Partner at FirstMark
“Most people building in this space are treating compute as a fungible commodity like oil, and most of the infrastructure being built reflects that. Liquid Compute started from a different premise. They are building the physical market first, then the regulated financial layer on top of it, closer to how power markets actually work.”
Mark Goldberg, Managing Partner at Chemistry
“BGC is excited to have signed a data partnership with Liquid Compute to support our build out of the OTC market for Compute and congratulate Ronit and team for their successful capital raise.”
Arran Rowsell, Head of Strategy at BGC Group
