HRL Laboratories (jointly owned by Boeing and General Motors until IBM's pending acquisition, covered earlier this week) published a Nature cover paper demonstrating an 18-qubit silicon spin quantum processor that generates all of its own control signals from a custom chip living inside the cryostat itself, with zero real-time input from room-temperature electronics. The system ran quantum error correction autonomously, with errors falling roughly fivefold as more qubits were added to the error-correcting code, and the whole architecture, chip, controller, and interconnect, was built on commercial semiconductor fabrication lines rather than exotic one-off components.

This is the actual technical asset behind IBM's HRL acquisition, published in a peer-reviewed, cover-featured Nature paper just over a week after the deal was announced. It directly answers the question I flagged in this feed's editorial angle list: does HRL's IP meaningfully change IBM's hardware roadmap, or just add headcount. This is real evidence toward the former.

Solving the control-electronics wiring bottleneck, not just the qubit itself, is one of the least glamorous and most genuinely important problems in scaling any quantum modality. This is a credible, independently reviewed result, not a press release claim. Still an 18-qubit demonstration; the real test is whether this cryogenic CMOS approach holds up as IBM tries to fold it into a much larger superconducting-plus-silicon roadmap.

Members of the HRL Quantum Team and Collaborators. (2026, July 29). A digitally controlled silicon quantum processing unit. Nature. https://www.nature.com/articles/s41586-026-10754-7