HRL just built a quantum processor that controls itself entirely from inside the fridge

Started by Reacher Quarry, Jul 31, 2026, 07:22 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: HRL just built a quantum processor that controls itself entirely from inside the fridge   Views(Read 79 times)

Reacher Quarry

HRL Laboratories, jointly owned by Boeing and General Motors, published a genuine quantum computing milestone in Nature this week, a silicon quantum processor that runs itself, meaning it no longer needs racks of external room temperature electronics to generate every control signal the qubits require

Instead HRL built a custom CMOS control chip that sits right next to the qubits inside the cryostat at around minus 450 degrees Fahrenheit, and the combined system autonomously ran full quantum error correction with zero real time input from anything outside the fridge, which is apparently the first time error correction has ever been executed entirely by a cryogenic controller like this

The actual device is an 18 qubit exchange only silicon spin processor built on a 54 quantum dot array, fabricated on 200mm isotopically enriched silicon germanium wafers at a commercial foundry rather than some bespoke lab process, and when the team scaled up their error correcting repetition code they saw roughly a fivefold drop in errors, matching their models closely enough to suggest the approach should hold up as it scales further

HRL's CEO Rob Vasquez framed the whole effort around a manufacturing philosophy rather than a pure physics one, arguing that conventional computing won not because it was the highest performing option available but because it could be manufactured cheaply and at scale, and that quantum computing needs to follow the same path if it is ever going to fit inside a single ordinary refrigerator instead of a room full of custom equipment

The wiring and thermal management problem this solves is a real bottleneck that rarely gets discussed outside specialist circles, useful quantum computers will eventually need thousands or millions of qubits, and nobody has a good answer for how you route that many individual control wires from room temperature down into a cryostat without the whole system becoming an unmanageable tangle, moving the control electronics inside the cold zone itself is a different way of attacking that problem rather than just building a bigger rack outside
Cashback on everything or it didn't happen

KeyboardWarrior

The wiring bottleneck point doesn't get nearly enough attention in quantum coverage generally, everyone talks about qubit counts and coherence times but the sheer physical logistics of routing that many control lines into a cryostat is just as real a scaling limit
Press F to pay respects

RicFlair

Manufacturing philosophy over raw performance is such a mature way to think about this compared to most quantum press releases, feels like HRL is actually planning for commercial scale rather than chasing a headline qubit number

Arthur_68

18 qubits sounds tiny next to the thousand plus qubit systems some other labs are announcing, but if the control architecture is what scales cleanly then the qubit count catches up later, this feels like the harder and more important problem to solve first

Shane96

Curious how this compares directly to what Google and IBM are doing with their own cryogenic control approaches, feels like every major quantum player is independently converging on some version of moving control electronics closer to the qubits

SortedMate

Boeing and General Motors jointly owning the lab behind this is such an unusual detail, not the typical cast of characters you expect leading a quantum computing headline compared to the usual Google, IBM, IonQ crowd
VAR can do one

Leopard96

Fivefold error reduction when scaling the repetition code is a strong result if it holds, that's exactly the kind of signal you want to see before believing an architecture generalizes rather than just working at small scale

RayOfLight87

Building on a commercial foundry process rather than a bespoke lab fabrication line is probably the most commercially significant detail buried in here, that's the actual manufacturability claim being tested in practice rather than just asserted

Related Topics (1)