Harvard physicists crack quantum error correction, clearing path to supercomputers

For thirty years, quantum computing has lived in a frustrating gap: the machines could theoretically solve problems no classical computer could touch, but they were too error-prone to actually do it. A team at Harvard just changed that.

Physicists led by Mikhail Lukin have demonstrated a "fault tolerant" quantum system using 448 qubits—the quantum equivalent of classical computing's bits—that can detect and correct its own errors as it runs. They published the findings in Nature, and the achievement marks the first time researchers have combined all the essential pieces needed to scale quantum computing from lab curiosity to practical machine.

"For the first time, we combined all essential elements for a scalable, error-corrected quantum computation in an integrated architecture," Lukin said. The work involved physicists from Harvard, MIT, the University of Maryland, and the National Institute of Standards and Technology, alongside QuEra Computing, a startup spun out from Harvard-MIT research.

Why This Matters

Quantum computers don't work like your laptop. Instead of storing information as zeros and ones, they use quantum bits that exploit the weird rules of subatomic physics to hold multiple states simultaneously. In theory, this lets them process certain types of problems exponentially faster than any classical machine. In practice, qubits are fragile. They degrade. They make mistakes. For decades, that fragility made scaling them up nearly impossible—adding more qubits meant adding more errors, not solving bigger problems.

The Harvard team's breakthrough is architectural. They built layers of error correction into the system itself, using techniques like "quantum teleportation" (transferring a quantum state from one particle to another without physical contact) and logical entanglement to catch and fix mistakes before they cascade. More importantly, they proved you can do this at scale. When they added more qubits to their system, errors actually went down—crossing what physicists call the critical threshold.

"This is the first time we have an architecture that is conceptually scalable," said Dolev Bluvstein, the paper's lead author. Significant technical challenges remain before anyone builds a practical quantum computer with millions of qubits, but the path is now visible.

The team works with neutral rubidium atoms, using lasers to encode them as qubits. In September, the same group published another Nature paper showing they could operate a system of over 3,000 qubits continuously for more than two hours—solving another major engineering headache: keeping atoms from escaping the system.

Lukin, co-director of Harvard's Quantum Science and Engineering Initiative, has spent decades chasing this moment. "This big dream that many of us had for several decades, for the first time, is really in direct sight," he said. The core pieces for building quantum computers that actually work are finally falling into place.