TL;DR

The demand for private inference is rapidly increasing to control token costs and protect data privacy.

• Top researchers from New York University, Stanford University, Dartmouth College, and University of Hawaii at Manoa are using B3IQ to support their AI work, including cancer research and dedicated AI model training.

• A survey of 1,800 IT leaders by Broadcom in 2026 revealed that 56% of enterprises are already running or planning to run production-grade inference on private cloud infrastructure, while the usage of public cloud for similar workloads has decreased from 56% to 41% in a year.

• B3IQ currently operates a 27,000-square-foot facility in Oregon and plans to rapidly expand its inventory of US-built NVIDIA GPU systems to meet the growing demand.

Privacy is becoming an increasingly prominent issue in the AI field. Every prompt sent to a centralized AI service provider is like a deposit into someone else's vault: research data, business logic, and creativity flow into a system beyond the sender's control, with terms not set by themselves. The most direct solution is to run AI models on proprietary hardware, but this method was previously out of reach for the vast majority: high-end GPU systems cost tens of thousands of dollars, are in short supply, and require space and professional maintenance to operate.

Today, B3 Labs has released B3IQ to fill this gap: B3IQ is a groundbreaking AI infrastructure service designed to give universities, enterprises, and professional users greater control over the hardware, models, and data on which their AI workloads depend.

Prior to this, institutions needing AI computing power had only two imperfect choices: either rent from cloud vendors, which means prices fluctuate, no ownership, and limited supply when GPUs are scarce; or buy the whole system outright, which means bearing all upfront costs, plus the burden of power, cooling, and maintenance. B3IQ offers a third model that combines both: economically, it is "ownership": assets and returns belong to you; operationally, it is "hosting": data centers and maintenance are handled by the platform.

B3IQ Users Can Purchase Exclusive NVIDIA GPU Systems through Installment Payments: Manufactured in the U.S. by Andromeda, an AI system maker Andromeda, hosted in Oregon. Through the B3IQ dashboard, owners can match idle computing power with power demands to monetize it. The earnings can be used to offset the remaining balance of the purchase or kept as income. After full payment, owners can choose to continue hosting with B3IQ or arrange for physical hardware delivery.

Early users of B3IQ include New York University, Dartmouth College, University of Hawaii at Manoa, and faculty, AI researchers, and student teams at Stanford University. For this group, the budget pressure caused by GPU shortages is particularly real:

“B3IQ’s ‘Owner Control’ model is a viable path between leasing and purchasing, which is why we decided to collaborate with B3IQ. Research funds are fixed and allocated in advance, while cloud computing costs are variable and may silently consume an entire budget line halfway through a project. Turning computing power into a predictable, budgeted cost makes planning easier and accountability to principal investigators or finance departments more straightforward.” Hawaii University AI researcher Pavel Bushuyeu said, “Having our own computing power also shields us from ‘compute scarcity.’ When GPUs are in short supply, centralized service providers often allocate them in limited quantities, with academic users often positioned behind paying enterprise clients. With our own nodes, there is no need to queue for resources when tasks need to be run. And when the system is idle, using a portion of its idle computing power to offset the purchase cost can also spread this investment.”

For researchers and institutions who cannot entrust sensitive data to third-party model services such as Anthropic and OpenAI, this is a substantial market. Pavel Bushuyeu from the University of Hawaii and other B3IQ pilot users are running proprietary models on the platform for cancer research and robot training, data that cannot be leaked at all. In addition, AI service providers may directly block specific keywords and topics, potentially hindering entire areas of research: In Professor Yorke E. Rhodes III's Ethical Tech CoLab lab at New York University's Center for Global Affairs, master’s students are constructing frameworks and simulations based on conflict evacuation data, simulating high-level diplomatic negotiations. Such work triggers content filtering in commercial models and can only run on the team's own infrastructure.

B3 Labs believes that this model is now viable, relying on the rapid advancement of open-source models: Models that are now downloadable and deployable are strong enough, and institutions no longer need to rely on interfaces from centralized vendors like OpenAI, with one caveat: You have to have your own machine to run it. And this is precisely what B3IQ provides.

“Institutions want better control over where AI runs, how data is handled, and how much they pay for compute.” B3 Labs CTO Sean Geng said, “B3IQ consolidates these decisions into a single system: running private workloads on bespoke hardware, then leveraging idle GPU compute through a network that participants can selectively join.”

Whether you're looking to procure GPUs, rent out idle compute power, or directly lease compute power from the network, you can find out more at b3iq.org.

About B3 Labs

B3 Labs builds software and hardware for enterprise AI. Founded in 2024, the team hails from Coinbase and has raised over $21 million from institutions like Pantera Capital and Coinbase Ventures. B3 Labs operates two product lines: B3OS, an enterprise AI execution engine that allows AI agents to perform tasks stably and controllably within enterprise systems, and B3IQ, a customer-owned GPU infrastructure deployed on U.S. soil. For more information, visit B3OS.org and B3IQ.org.

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia