HRL Laboratories Demonstrates Autonomous 18-Qubit Silicon Spin Quantum Processor as IBM Moves to Acquire the Lab
HRL published a Nature paper on an 18-qubit silicon spin processor run by an onboard cryogenic chip, days after IBM agreed to acquire the lab.
Overview
HRL Laboratories has published a paper in Nature describing a silicon spin quantum processing unit that runs its own control sequences from an onboard cryogenic chip rather than relying on racks of room-temperature electronics, according to arXiv, where the underlying preprint is posted. The demonstration, dated July 29, 2026 according to Quantum Computing Report, lands less than a week after IBM announced a definitive agreement to acquire HRL Laboratories, according to IBM’s newsroom.
What We Know
The chip integrates a three-rail array of 54 exchange-coupled quantum dots that can be configured to host up to 18 exchange-only qubits, according to the arXiv preprint. Control of those qubits is handled by a custom-designed cryogenic CMOS controller built on a 130-nanometer RF-CMOS process with roughly 70 million transistors, which operates inside the cryostat at 4 kelvin while drawing approximately 3.5 watts, according to the paper’s full text and corroborated by Quantum Computing Report. That controller routes 150 time-varying control signals plus 35 static device biases to the qubits through a superconducting interconnect that adds less than 10 microwatts of thermal load to the system, per the arXiv text.
The paper reports mean single-qubit gate errors of 2×10⁻⁴ and mean CNOT (two-qubit) gate errors of 3×10⁻³, with the lowest reproducible CNOT error reaching 9×10⁻⁴, according to the arXiv preprint. The authors describe these results as advancing the state of the art for exchange-only qubits “by an order of magnitude,” per the paper’s abstract. The Quantum Insider similarly describes the demonstration as achieving control errors roughly ten times lower than any prior showing of this qubit type.
The team also validated the chip by running error-correction experiments: a distance-5 repetition code produced a logical error rate of 5.0×10⁻³ across 200 syndrome-extraction rounds, and a distance-3 version of the same code produced a logical error rate of 3.2×10⁻³, according to the arXiv text. A separate [[4,2,2]] quantum error-detecting code reached a logical fidelity of 0.95 after three rounds of syndrome extraction when post-selected on the error-detection outcome, compared with an average fidelity of 0.59 without that post-selection, per the same source.
The Nature publication arrived days after IBM said it had signed a definitive agreement to acquire HRL Laboratories, a research institution jointly owned by Boeing and General Motors since 1997, according to NAND Research. Both Boeing and GM will continue partnering with IBM on quantum applications and other technology development after the transaction closes, according to IBM’s announcement, which did not disclose financial terms. “The HRL team will help IBM push even farther forward toward the frontiers of quantum innovation. This talented group of researchers brings a broad portfolio of technologies that will strengthen IBM’s long-term plans to deliver useful quantum computing to the world,” said Jay Gambetta, IBM’s Director of Research and IBM Fellow, in the announcement. HRL President and CEO Rob Vasquez called the deal “the natural next chapter for what we have built at HRL, where our team has dedicated years to exploring paths to how future quantum computers could be built at scales that today seem impossible,” according to the same IBM release.
IBM’s existing quantum roadmap, including the planned 2029 delivery of its Quantum Starling system, has been built around superconducting qubits, according to NAND Research, with a follow-on system called Blue Jay planned for the mid-2030s, according to Techzine. HRL’s silicon spin qubits operate at roughly 1 kelvin, a substantially warmer and easier-to-engineer regime than the 15 millikelvin required for superconducting qubits, according to NAND Research. IBM said in the same announcement that it anticipates HRL’s expertise in silicon spin qubits, cryogenics, control electronics and packaging will “fuel its quantum program for decades to come,” and pointed to Anderon, a quantum wafer foundry IBM said in May 2026 it would establish and described as “the world’s first pure-play quantum wafer foundry,” as a potential partner for future spin-qubit manufacturing, according to IBM’s newsroom.
What We Don’t Know
Neither IBM nor HRL disclosed the financial terms of the acquisition. It is not yet clear when the deal is expected to close, how HRL’s roughly 250-author research team will be organized within IBM, or on what timeline — if any — silicon spin qubits might be folded into a shipping IBM quantum system alongside the superconducting Starling and Blue Jay machines.
Analysis
The timing links two otherwise separate developments into a single narrative about how IBM is hedging its quantum bet. The company has spent years committing publicly to a superconducting roadmap, but the HRL acquisition — arriving the same week HRL demonstrated an autonomously controlled, error-corrected silicon processor in a peer-reviewed journal — gives IBM a second, CMOS-compatible qubit technology as a fallback path if superconducting qubits hit a scaling wall before fault tolerance is reached.