Ray Song
Ray Song is one of Capricorn's main contributors and the founder of Computable, a company that builds pricing tools and cash-settled derivatives for graphics-processing-unit (GPU) compute. Before founding Computable he worked as a trader and in crypto infrastructure at firms including Susquehanna International Group, Flow Traders, and Jump Trading, and contributed to the Pyth Network market-data oracle and to the Capricorn/aPriori project. He is based in the San Francisco Bay Area.[1][2]
Early Life and Education
Song attended The Harker School, a private school in San Jose, California, before enrolling at Claremont McKenna College in Claremont, California, where he earned a bachelor's degree in Applied Mathematics.[1]
During and around his undergraduate years he took a series of finance and research internships. In the summer of 2019 he was a summer analyst at Oceanic Partners, where, by his own account, he "looked at the private markets, wrote memos, built spreadsheets," and handled miscellaneous office tasks. In January 2020 he completed a quantitative trading internship at Susquehanna International Group (SIG), which he summarized as covering "options theory, game theory, research, & poker." From August to December 2020 he served as a project manager on a Mathematics Clinic collaboration with the Claremont Graduate University, working on a Sandia National Laboratories project applying machine learning to predict atomic-scale glass failure; he described himself as "the least qualified person on a team of 3 PhDs and 2 Master's students."[1]
Trading Career
Flow Traders
Song began at Flow Traders, a global technology-enabled liquidity provider that specializes in exchange-traded funds (ETFs), with a summer business course in January 2020, after which he worked as a trader in New York from January 2020 to January 2022. He summarized the trading role, with characteristic brevity, as "buy low, sell high," and logged a brief "garden leave" in January 2022 that he annotated "decaffeinating myself."[1]
Jump Trading and Pyth Network
From June 2022 to August 2023 Song worked in the crypto operations of Jump Trading LLC, a research-driven quantitative trading firm founded in 1999 and headquartered in Chicago. His work spanned the trading arm and Jump Crypto, the group's Web3 infrastructure and investment division, where his role was listed as "Special Projects & Partnerships." Over the same period he was a core contributor to the Pyth Network, described as a next-generation oracle that brings high-frequency, high-fidelity market data on chain from more than 120 financial institutions and traders.[1]
Reflecting on the Pyth work in 2022, Song wrote that he was "proud to have been a part of this journey, bringing high-fidelity, low-latency data on chain," and noted how much the network had accomplished "in just one year."[1]
Capricorn and aPriori
Since September 2023 Song has been a core contributor to Capricorn, the entity behind aPriori, a nonprofit that describes itself as "an intelligent order flow coordination layer for high-performance blockchains." aPriori is a small organization headquartered in San Francisco with a workforce distributed across Singapore and the United States, and it reported $28 million in total funding across one prior round. Song's role label at Capricorn is styled as "janitor," an operations designation consistent with his self-presentation.[1][2]
In July 2024, the project announced the completion of a fundraising round and named Binance Labs (later YZi Labs) as an investor; Song shared both announcements.[1]
Computable
Song is the founder of Computable, which he began operating in the San Francisco Bay Area in August 2026. The company builds financial infrastructure around the price of GPU computing power, aimed at artificial-intelligence firms that rely on large fleets of GPUs. Song has said Computable grew out of internal tools the team built after fielding requests from AI companies on "how to hedge their gpu price risks through cash settlement derivatives."[1][2]
GPU Index and Compute Exchange
Computable positions itself, in its own words, as "the only exchange for physical compute where companies can buy, sell, and redeem GPU hours for any week with instant liquidity."[3] On September 1, 2026, the company introduced the Computable GPU Index (CGI), which it described as "the first open-source price index for GPU compute, with a mathematically robust methodology anyone can verify and reproduce," built on open data, open method, and open code. Song marked the launch with the line, "The price of compute. Finally computable."[2]
Rate Floor Pricer
On August 24, 2026, Computable launched the Rate Floor Pricer, a tool it likens to "insurance on GPU prices." The company framed the problem it addresses in terms of long-term hardware commitments: GPU leases are "massive price bets," with companies signing three-to-five-year deals whose value is exposed to depreciation, obsolescence, and oversupply. Computable stated that companies were "already using it to hedge 8-figure exposure" and that the tool, previously internal, was being opened to all users.[2][1]
According to biographical material for the Energy Investors Forum, Computable is backed by Y Combinator and OpenAI, and Song is identified there as a co-founder. The same source states that Song previously worked at Jump Trading and co-founded a trading infrastructure company that was valued at $250 million.[3]
Public Appearances
Song took part in the Energy Investors Forum on July 23, where he was listed under the affiliation BlockSpaceForce. He moderated a session titled "Securing Bitcoin in the Post-Quantum Future" and was a panelist on "Private Markets: Financing the Compute Economy," the latter aligning with his work at the intersection of capital markets and GPU compute.[3]
Song is active on X, where he comments on markets, technology, and sports culture, and has used the platform to preview Computable's product launches.[2]
