Fermilab's QICK, the Quantum Instrumentation Control Kit, the open source hardware used to steer qubits in real time, is moving to commercial production under Qblox through a federal Cooperative Research and Development Agreement (CRADA) licensing deal.
The U.S. Department of Energy did not just license a piece of open-source quantum hardware to a Dutch company. It identified a bottleneck that has nothing to do with qubits, and quietly consolidated it.
Fermilab's Quantum Instrumentation Control Kit, or QICK, began life as an in-house architecture for reading out and pulsing the superconducting and other qubits that physicists use to do their experiments. The kit runs on field-programmable gate arrays (chips whose logic can be rewired after manufacturing) and on RF system-on-chip devices, which put radio-frequency signal generation and digitisation onto a single piece of silicon. Built on FPGA and RF-SoC development boards, QICK gave research labs a programmable way to control qubits in real time without buying the expensive, vertically integrated commercial alternatives that dominated the field. The DOE frames it as a piece of shared infrastructure that university groups can actually afford.
For most of its life, QICK was simply a piece of shared lab infrastructure, useful but unglamorous. Then the U.S. government decided it mattered enough to formalise.
Under a new Cooperative Research and Development Agreement with Qblox, the company gains exclusive U.S. manufacturing, domestic supply-chain coordination, and commercial distribution rights for QICK-based control hardware. Fermilab retains the open-source code, and the licensing arrangement gives the federal lab a structured seat at the table for ongoing research priorities. Qblox, for its part, gets a federally validated product line and a path to convert an open-source standard into commercial revenue.
Workforce development is part of the same envelope. Qblox will design hands-on QICK training curricula for engineering students, academic researchers, and facility technicians, according to the Department of Energy. The stated goal is to make sure U.S. laboratories, not just commercial customers, can actually operate the hardware the company will ship.
The CRADA is the headline, but a parallel announcement around the same time was nearly as telling. Qblox also said it would integrate its quantum control topologies into Hewlett Packard Enterprise's classical high-performance computing and AI data-center stacks, per the company's separate release. The two partnerships line up: if Qblox becomes the default U.S. manufacturer of the control layer for quantum research, getting that same control layer to plug cleanly into classical HPC infrastructure is the next commercial problem to solve.
That double play reframes what the "quantum race" actually is. The press cycle still treats it as a contest between qubit modalities and chip counts. The control plane, the analog and digital hardware that turns human instructions into precisely timed microwave pulses and reads the answer back out, gets far less attention. It is also the layer every quantum program needs, the layer where open-source architectures are relatively scarce, and the layer where supply-chain control is most easily regulated by a national government.
The arrangement is not exclusive in the legal sense. Qblox gets preferred commercial manufacturing rights, not a closed monopoly on the QICK design itself. Anyone can still use the open-source code. What changes is who builds the integrated hardware at scale, who handles the U.S. supply chain for radio-frequency components and FPGA boards, and whose name appears on the box that ends up in a quantum lab. SDxCentral and ExecutiveGov both characterised the deal as a manufacturing-scaling arrangement rather than a technology transfer; the open-source lineage remains intact.
What to watch next is whether the federal bet pays off across three places that have nothing to do with qubit counts: whether Qblox can stand up a domestic supply chain for the specialised analog and RF components QICK depends on without cost overruns that price university labs out; whether the workforce-development curricula produce a pipeline of engineers who can operate control hardware that is now a private-company product rather than a freely licensed lab resource; and whether the HPE integration ever yields a hybrid system in production, rather than in a press release.
If any of those three numbers moves in the next eighteen months, the DOE's announcement will read less like a routine research partnership and more like the moment the United States decided the control plane, not the qubit, was the layer worth controlling.