The race to build a useful quantum computer has a new frontrunner — and it's made of silicon.

Four independent research teams published major breakthroughs this week in silicon spin qubits, a quantum computing technology that had long been overshadowed by more headline-grabbing approaches. The results, including two papers published Wednesday in the journal Nature, show that spin qubits can now achieve error rates on par with rival technologies — and they do it using manufacturing techniques borrowed directly from the semiconductor industry.

What are spin qubits?

All quantum computers rely on qubits — quantum bits that can exist in multiple states simultaneously. But there's no single best way to make a qubit. Google, IBM, and others have focused on superconducting circuits. Other groups work with ions trapped by lasers. Spin qubits take a different approach: they use single electrons — and specifically the quantum property called "spin" — as qubits, confined within tiny wells carved into silicon chips.

The appeal is obvious. Silicon is the same material that powers every classical computer and smartphone. In theory, spin qubits should be manufacturable using the same equipment and expertise that already powers the modern semiconductor industry. In practice, spin qubits have progressed slowly, and until this week, the best systems contained only a handful of qubits operating at error rates too high to be useful.

The breakthrough

Three years ago, state-of-the-art spin qubit systems had error rates around 4% for critical measurements. This week, two groups reported a roughly 20-fold improvement.

HRL Laboratories in Malibu, California, described an 18-qubit silicon processor achieving error rates of approximately 0.2%. QuTech, the quantum research center at Delft University of Technology in the Netherlands, published results from a 5-qubit silicon device hitting the same benchmark. Both results appeared in Nature on July 29, 2026.

They were not alone. Groove Quantum, a start-up also based in Delft, reported in April an 18-qubit germanium-based device with a similar error rate. Most impressively, RIKEN, Japan's national research institute, recently posted preprint results showing a 5-qubit system with an error rate below 0.01% — the lowest achieved by any qubit platform on record.

"It's just great for the community that such advancement has been made," said Daniel Loss, a theoretical physicist and a pioneer of the spin qubit concept. An independent peer reviewer of the HRL paper called it "a significant milestone in the technological maturity of semiconductor spin qubits."

Why it matters

Quantum computers are still far from replacing classical machines. Superconducting systems from Google and IBM operate with more than 100 qubits, and neutral-atom platforms boast thousands. But scale alone does not determine usefulness — error rates matter enormously. Low-error qubits perform reliable calculations; high-error ones produce noise and garbage.

Spin qubits' use of standard silicon manufacturing means they could, in principle, be scaled up using existing chip fabrication facilities — potentially giving them a long-term advantage over technologies that require exotic materials or highly specialized equipment.

The commercial world is already paying attention. IBM, which has focused on superconducting qubits, announced last week that it would acquire HRL Laboratories. "The HRL team will help IBM push even farther forward toward the frontiers of quantum innovation," said Jay Gambetta, IBM's director of research.

For a technology that many had quietly written off, the spin qubit comeback could not be better timed.