Auctionomics, a market design firm co-founded by Nobel laureate Paul Milgrom and Dr. Silvia Console Battilana, and OCX Group Inc., the parent company of OneChronos, have announced a strategic partnership to develop the first financial market dedicated to GPU and compute capacity.
This collaboration aims to establish the financial infrastructure necessary to help various organizations, including investors in data centers, data center operators, power generators, chip manufacturers, and enterprise buyers, manage the unprecedented demand for AI compute resources anticipated over the next decade.
Auctionomics specializes in applying game theory and mathematical models to design complex auctions. The firm’s work has generated billions in economic value and addressed critical challenges in telecommunications, natural resources, and digital markets, including the Emmy-winning design of the FCC auction that facilitated television streaming. Dr. Silvia Console Battilana, CEO of Auctionomics, noted that the compute market has reached a scale and complexity requiring sophisticated market designs akin to those for mature financial asset classes. She emphasized that Auctionomics’ expertise in designing multi-billion dollar auctions, combined with OneChronos’ Smart Market platform, will enable the creation of a market mechanism to bring transparency, efficiency, and risk mitigation to compute resource markets.
The new off-exchange marketplace will leverage a novel application of OneChronos’ Smart Market technology to deliver real-time price discovery and liquidity through bi-lateral forwards. This will allow users to lock in future capacity and pricing in advance, providing price certainty, capacity guarantees, and planning visibility essential for enterprise AI initiatives and large-scale computing projects. OneChronos, a technology company innovating at the intersection of capital markets, mechanism design, and operations research, currently operates an alternative trading system for U.S. equities, regulated by the SEC and FINRA, through its subsidiary OneChronos Markets LLC.
The auction-based approach seeks to more efficiently allocate data center and energy resources, including GPU capacity. It directly addresses a fundamental challenge faced by organizations providing access to compute or building AI applications: the inability to reliably hedge against fluctuations in price, availability, and efficiency of compute resources and upstream supply chains within a rapidly evolving market. Existing on-demand and reserved compute models expose organizations to price volatility and capacity bottlenecks during peak demand, as well as uncertainties regarding standards and performance. This new market structure will also introduce a layer of predictability to the energy aspect of compute, assisting data center operators and power providers in hedging against volatile power prices and long-term consumption risks.
The market is anticipated to serve a diverse group of participants, including AI research labs, universities, firms investing in data center infrastructure, power generators engaging in long-term power purchase agreements, data center operators seeking to hedge against market volatility, and organizations developing AI applications.
OneChronos’ Smart Market technology, which underpins this new market, utilizes combinatorial auctions and mathematical optimization to match counterparties based on complex, multi-asset preferences. This technology enables participants to express nuanced trading goals across portfolios, unlocking efficiencies that traditional one-to-one markets cannot achieve. Applied to compute, the technology can optimize across interdependencies between different types of resources, capturing synergies that conventional auctions might miss. Eventually, participants are expected to be able to bid on bundles of compute capacity, power capacity, energy storage, and other compute-upstream resources, facilitating efficient allocation across diverse and dynamic computing needs.
Kelly Littlepage, CEO and co-founder of OneChronos, highlighted the success of Smart Markets in equities trading, which currently surpasses $6.5 billion in daily volume. He stated that the technology, made possible by advances in machine learning, can electronify more of the financial and real economy beyond capital markets. Littlepage noted that the market for computing resources, already larger than some major commodity markets like copper, is ill-suited for conventional auctions, and that this partnership will bring improved price discovery and risk mitigation to this critical resource.
The parties foresee the financial market operating through a user-friendly interface that simplifies its underlying mathematical complexity, allowing institutional participants with varied technical backgrounds to engage in sophisticated trading without requiring deep knowledge of auction theory or optimization algorithms. The system will incorporate monitoring tools designed to detect potential market manipulation and ensure fair access for all participants.
