Startup B3IQ has found a new AI niche: Supplying rent-to-own machines to privacy-conscious university researchers
Fortune Jeff John Roberts
A startup is renting researchers AI machines they can eventually own. It helps them keep costs steady and data private, which campus and cloud options often don’t.
Based on reporting by Fortune, Jeff John Roberts — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
A PhD student at the University of Hawaii wanted GPU power for cancer research, but campus computers were limited and outside cloud services were pricey and awkward for private data. He found a different route: B3IQ, the AI hardware business from B3 Labs, which lets customers put 30% down on dedicated machines and pay the rest over five years.
That setup is aimed squarely at people who hate surprise bills. A frontline Nvidia H200 rig starts at $54,516, and B3IQ says its buyers can avoid the problem of variable hourly pricing on rented compute. For university work, that predictability matters. Grants often come with fixed budgets, and AI training can chew through cash faster than a lab can write the next progress update.
Privacy is the other selling point. Researchers handling health-related data can keep a dedicated machine under their control, which makes compliance with rules like HIPAA easier. It also sidesteps the headaches that can come from using overseas compute providers, where regulators and national security officials may start asking questions. B3IQ can place the hardware on a customer’s own premises, or host it in its 27,000-square-foot facility in Oregon.
The company only began offering this rent-to-own model this year, but it is already landing academic customers. New York University researchers are using B3IQ compute for simulations based on war-zone evacuation data and for modeling nation-level diplomatic negotiations. The startup says Stanford and Dartmouth are customers too, and it is talking with university procurement offices about broader deals.
That’s a sharp turn for a company that started in 2024 with video games. The co-founders, who met at Coinbase, ran into the same GPU shortage everyone else did in early 2025 and shifted hard into AI hardware and software. Now B3IQ says it can assemble rigs, help customers upgrade to newer chips, and even resell spare capacity to approved third parties. Xu says the company has around 30 employees, expects $50 million in revenue this year, and has raised $21 million from backers including Pantera and Coinbase Ventures.
My take — AI-written commentary, not fact-checked reporting
This is the kind of unglamorous AI business that actually makes sense: sell the metal, finance it, and stop pretending every researcher wants a surprise cloud bill. B3IQ is basically turning GPU scarcity into a lease agreement, which is less sexy than a model launch and probably more useful. The real tell is that universities care about budgets and privacy first, hype second.
Read more about this at: Fortune