Harvard Crosses Quantum Threshold

Harvard Just Crossed the 30-Year Quantum Threshold

For three decades, quantum computing has had the same problem: errors pile up faster than you can fix them.

To put the problem in perspective, imagine filling a bucket with holes in the bottom. Pour water in the bucket and it leaks out. Not only does it leak, it leaks faster than you can fill it. So you pour more water — but more holes develop. That’s the problem. Quantum computing has been bailing water for 30 years.

Harvard just built a bucket that seals itself faster than it springs leaks.

A team led by Harvard, with MIT, QuEra, and the University of Maryland, just demonstrated the first complete fault-tolerant quantum architecture. Not a piece of one. Not a promising step toward one. The whole thing, working together.

This is important because quantum computers are fragile. Every calculation introduces errors. The dream has always been to correct those errors faster than they accumulate — cross that threshold, and you can scale. Stay below it, and you’re just building expensive noise machines.

This team hit 2.14x below threshold. Errors are getting fixed faster than they’re created. That’s a sealed bucket.

What they actually built

448 neutral atoms running as logical qubits with surface codes. They ran dozens of logical qubits and hundreds of logical teleportations in a single deep circuit.

The sleeper hit is mid-circuit qubit reuse. Instead of burning through atoms and discarding them, they recycle. They ran 150-cycle operations on the same atoms with measurement and reset between each one. The system expels errors continuously instead of accumulating them.

That’s not a hack. That’s architecture.

The competition is paying attention

Hartmut Neven, VP of Engineering at Google Quantum AI — a direct competitor using superconducting qubits:

“This work represents a significant advance toward our shared goal of building a large-scale, useful quantum computer,”

When Google’s quantum team tips their hat to a neutral atom approach, the field is shifting.

What’s still missing

Scale. This architecture works at 448 qubits. Cryptographically relevant quantum computing needs millions. But the physics now works. What remains is engineering.

Mikhail Lukin, one of the lead researchers, put it plainly:

“This big dream that many of us had for several decades, for the first time, is really in direct sight.”

Steady progress from Harvard

Back in September of this year, Harvard physicists overcame the persistent problem of “atom loss” (qubits escaping and losing their coded information) by developing a continuous resupply system using optical lattice conveyor belts and optical tweezers, enabling the first quantum computer that can run indefinitely without needing to stop and restart.

In another article from The Harvard Gazette, Kermit Pattison writes:

“A machine with 300 quantum bits could simultaneously store more information than the number of particles in the known universe. Now process this: Harvard scientists just unveiled a system that was 10 times bigger.”

Go read the article (https://news.harvard.edu/gazette/story/2025/09/clearing-significant-hurdle-to-quantum-computing/). It is truly mind-blowing stuff they’re building.

Sources:

• Nature — A fault-tolerant neutral-atom architecture for universal quantum computation (https://www.nature.com/articles/s41586-025-09848-5)
• Harvard Gazette — “A potential quantum leap” (https://news.harvard.edu/gazette/story/2025/11/a-potential-quantum-leap/) (November 12, 2025)