Daryl Xu, chief executive of the New York‑based startup B3IQ, has introduced a rent‑to‑own programme that lets university researchers acquire high‑end GPU machines without the uncertainty of hourly cloud fees.
How the model works for academics
PhD candidate Pavel Bushuyeu at the University of Hawaii, who studies skin‑cancer impacts on Pacific Island populations, struggled to secure on‑campus compute and faced privacy hurdles when considering commercial cloud services. Through B3IQ's scheme he can place a 30 % down‑payment on a machine such as an Nvidia H200 rig priced at $54,516 and spread the balance over five years.
Why privacy and predictable costs matter
University grants often require fixed budgets, a condition hard to satisfy with fluctuating cloud rates. Owning a dedicated rig also helps researchers comply with health‑data regulations like HIPAA and avoids the scrutiny that can accompany overseas cloud providers.
Customers may choose to have the hardware assembled on‑site or, more conveniently, hosted at B3IQ's 27,000‑square‑foot data centre in Oregon.
Growing academic adoption
Beyond Bushuyeu, researchers at New York University are using B3IQ machines for war‑zone evacuation simulations and diplomatic‑negotiation models. The startup also lists academics from Stanford and Dartmouth among its users and is negotiating broader contracts with university procurement offices.
From gaming to AI hardware
The company began in 2024 as a video‑game venture founded by Xu, Viktoriya Hying and Sean Geng, all former colleagues at crypto platform Coinbase. Their initial focus shifted after they observed a surge in demand for GPU chips across gaming and AI workloads.
"AI compute was exponentially increasing. I think we had a lot of gaming customers just kind of yell at us because GPU prices were so expensive, and we thought there was something there. And so we did what we've always done: build vertical, meaning we built both software and hardware in the AI space," says Xu.
Running the Oregon data centre gave the founders hands‑on experience assembling and upgrading GPU rigs, enabling B3IQ to offer on‑demand upgrades and additional compute capacity.
Position in the market
While major cloud providers such as Amazon Web Services continue to dominate hourly compute, and firms like Vast.AI provide long‑term financing, B3IQ distinguishes itself by catering to the privacy and budgeting needs of academia. The company also allows owners to sell spare capacity through a managed dashboard.
Financial outlook and next steps
With a staff of roughly 30, B3IQ projects $50 million in revenue for the current year. It has secured $21 million in funding from investors including Pantera and Coinbase Ventures, and plans a larger Series A round in the coming months.
Looking ahead, the firm aims to expand its university client base across Europe and the United States, and to refine its financing terms to further ease the financial planning of research projects.