The Underdog Qubit: How Spin Quietly Caught Up
The Setup: A Simple, Devastating Question
In the race to build a useful quantum computer, everyone knows the favorites: superconducting circuits, running full-scale systems for Google and other major players, and neutral atoms, encoding information in the electrons of cold atoms held together by light. Spin qubits — built the way ordinary computer chips are, by carving patterns into semiconductor wafers — have spent years as the quiet also-ran, mentioned mostly as a promising idea that hadn't caught up yet. What happens when the underdog stops being underestimated all at once, in four different labs, in the same few months?
The Breakthrough: Four Teams, One Sudden Leap
Here's what actually shifted, with the technical noise cleared away.
Three years ago, the state of the art for a spin-qubit quantum computer was a system with roughly 4% error on single-qubit measurements — workable, but far behind what's needed for reliable computation. This year, that number collapsed. Two papers report complex measurements on single silicon chips with error rates around 0.02% and 0.3%: one from HRL Laboratories in Malibu, California, which built a device with 18 spin qubits, and one from QuTech in Delft, the Netherlands. Other groups are racing forward on the same technology: Groove Quantum, a Berkeley-based startup, reported a comparable 18-qubit germanium-based device with a 0.2% error rate on a preprint server in April. A month earlier, RIKEN, a Japanese research institute, reported a five-qubit system with an error rate below 0.01% in another preprint.
Spin, in this context, describes the angular momentum of a particle — able to be oriented in any spatial direction, and controllable enough to serve as the basic unit of a quantum computation. In the late 1990s, physicists Daniel Loss and David DiVincenzo, then at IBM, realized that electron spins made ideal qubits precisely because they could be manipulated the way ordinary transistors are, using electric pulses — a similarity that made the approach immediately attractive, even if it took decades of engineering to catch up with that promise.
Why It's Bigger Than It Looks
The obvious framing is "spin qubits get better." The bigger framing is what "better" now unlocks. Error rate is the single most decisive number in quantum computing — it determines how many operations a system can run before noise overwhelms the calculation, and by extension whether a given qubit technology is a laboratory curiosity or a viable path to something useful. Dropping from roughly 4% to 0.02% in a few years isn't incremental progress; it's the kind of jump that changes which architecture serious money and serious research attention get pointed at next.
It also matters because spin qubits inherit something the leading approaches don't have as naturally: manufacturing lineage. They're built using techniques adapted from decades of semiconductor fabrication, the same industrial base that makes ordinary computer chips at enormous scale. A qubit technology that error-corrects like the front-runners but scales more like existing chip fabs is a genuinely different proposition for anyone thinking about how quantum computing eventually gets built at volume.
The Part Nobody Talks About: The Numbers Still Trail the Leaders
It's worth keeping this leap in proportion. Superconducting circuits and neutral-atom systems are still far ahead in scale — superconducting quantum computers already run with more than 100 qubits, and neutral-atom systems with thousands, while the latest spin-qubit results are working with single digits to the high teens. Low error rates on a handful of qubits are a necessary foundation, not a finished computer; the technology still faces the much harder problem of scaling those low error rates up to the qubit counts the leading approaches already have. The breakthrough is real, but it's a breakthrough in quality at small scale, not yet in scale itself.
The Meta-Twist: The Biggest Vote of Confidence Came From a Rival
The clearest sign of how seriously this shift is being taken didn't come from inside the spin-qubit community — it came from IBM, a company that has bet its entire quantum roadmap on superconducting circuits. IBM announced it would acquire HRL Laboratories, the same lab behind one of the record-low error rate results, a move that surprised much of the quantum-computing community given IBM's public focus on a rival architecture. IBM's own director of research, Jay Gambetta, framed the deal as pushing the company further toward the frontiers of quantum innovation — which reads less like abandoning superconducting circuits and more like hedging: if spin qubits really are catching up, the safest move is to own a piece of that progress rather than compete against it from outside.
Conclusion: Underestimated Isn't the Same as Behind
Spin qubits didn't win the quantum computing race this year — nobody has. But they proved something narrower and, in its way, more important: that being the quiet option for a decade doesn't mean a technology has stalled, it can mean it's been accumulating incremental gains nobody was tracking closely, right up until four labs cross a threshold in the same season. The lesson for anyone watching any technology race is the same one IBM just acted on — the also-ran you stopped watching is exactly the one worth checking on again.
References:
https://www.nature.com/articles/d41586-026-02357-z
https://newsroom.ibm.com/2026-07-23-ibm-to-acquire-hrl-laboratories-to-power-the-future-of-quantum
https://quantumzeitgeist.com/ibm-acquisition-qubit-hrl-labs/