A famous result lands in a top journal. Press releases call it a major step toward useful quantum computing. Five years later, another team carefully repeats the experiment — and finds the original signals can be explained without the breakthrough. They submit. The journals tell them replication “lacks novelty.” Two years pass. The paper finally lands in Science in January 2026, and it should make us reread half a decade of quantum computing news.
That’s the situation Sergey Frolov and collaborators from Pittsburgh, Minnesota, and Grenoble walked into when they decided to test the most exciting claims in topological quantum computing.
The breakthroughs they checked
For nearly a decade, one of the most thrilling stories in the field has been topological superconductors — exotic materials that would, in principle, host particles called Majorana modes. A quantum computer built from those would be radically more stable than the noisy qubits in today’s machines, and orders of magnitude cheaper to error-correct than the kind of devices we covered in Physicists Hit Rewind on Time. High-profile papers presented experimental signatures that looked like Majoranas were finally appearing in nanoscale superconductor–semiconductor devices. Industry roadmaps were redrawn around the idea.
Frolov’s team set out to do something deeply unglamorous: reproduce the actual measurements.
What the replications found
The signals were real. The interpretation was not. With access to fuller datasets, the team consistently identified simpler alternative explanations — ordinary disorder, conventional Andreev physics, plain measurement artefacts — that fit the data at least as well as the breakthrough story. None of that proves Majoranas aren’t there. It proves the original experiments don’t establish that they are.
That distinction is the whole game in physics. A signal that’s consistent with an exotic effect is not a signal that requires one. Several of the most celebrated quantum results of the last ten years sit in exactly that uncomfortable middle zone.
The two-year fight to publish
This is the worst part. Journals rejected the replication study. The reasons were almost cartoonish: “replication work lacks novelty”; “the field has moved on.” Frolov’s group spent roughly two years in review before Science finally ran the paper — a publication delay long enough for more lab money to be spent chasing the original story.
That’s not unique to quantum computing. It’s the quiet structural reason a quantum computing replication crisis can build up without anyone really noticing: there’s almost no incentive to confirm or refute a result once everyone has moved on to the next exciting one. The careers are made on the first paper, almost never the second.
How to read the next quantum breakthrough headline
Three things worth keeping in mind when you see the next “scientists achieve” quantum result:
- The important question is whether the effect requires the exotic explanation, not whether it’s consistent with it. That’s a higher bar than press releases use.
- Single-lab demonstrations are not yet results. Watch for the second team that gets the same numbers under different conditions.
- The quiet replication papers are the ones that decide whether a field is real. They almost never make headlines. Go find them anyway.
None of this means topological quantum computing is over. It does mean the past decade of Majorana headlines should probably be filed under “open question” rather than “confirmed” — and that the way mainstream quantum journalism handles these claims needs to change.
For a different model of how a contested scientific question can be settled honestly, see Two Big Theories of Consciousness Fought in Public — where two rival camps agreed in advance which experiment would prove one of them wrong. That’s the version of science where replications happen up front, not two years late.
Honest physics is slow. Honest journalism about it should probably be slower too.