Microsoft’s quantum chip got an upgrade. Critics are still skeptical

Errors are a big problem for quantum computers, befuddling calculations and hindering the machines’ potential to outperform traditional computers. Microsoft’s new quantum chip could bring hope — for those who believe the company’s claims.

Microsoft has unveiled an improved version of its quantum chip, which is designed to resist errors. But the upgrade hasn’t convinced harsh critics of the company’s previous efforts.

Last year, Microsoft reported that it had developed a new kind of a quantum chip that protects information based on principles of topology, the field of mathematics dealing with the geometry of structures with holes. That claim immediately drew skepticism from scientists.

In the new chip, announced June 2 at Microsoft’s Build conference in San Francisco, researchers swapped some of the materials used to create the chip’s quantum bits, or qubits. In particular, the team used lead instead of aluminum as a superconductor, a material that transmits electricity without resistance at low temperature.

Those swaps improved a property called the parity lifetime, a marker of the quality of a topological qubit. These qubits are built from tiny superconducting wires, and the parity indicates whether there is an even or odd number of electrons within the wire. The lifetime indicates how long the parity is maintained before being spoiled by random jitter or outside influences. Previously, the qubit’s parity lifetime was measured in milliseconds; now, it is around 20 seconds, Microsoft researchers report in a paper posted on a Microsoft website and at arXiv.org. The paper is a preprint, so has yet to be peer-reviewed.

“We’re seeing this more than 1000x improvement in this critical metric of the qubit based on this change,” Microsoft physicist Chetan Nayak said in a May 29 news conference.

But topological quantum computing has a history of controversial claims, with multiple high-profile paper retractions in the field. So the bar for convincing physicists is high. And according to some, Microsoft hasn’t cleared it. “Nothing in this preprint resolves the fundamental issues,” says physicist Henry Legg of the University of St Andrews in Scotland.

Microsoft’s device hinges on the existence of groups of electrons that collectively behave like one entity, creating a particle-ish thing called a Majorana. Pairs of Majoranas can store information in a way that is resistant to errors. In the device, a Majorana sits at each end of a thin superconducting wire. Two parallel wires, connected like an H, together form a single qubit with four Majoranas.

The idea is that a pair of Majoranas can store information in a way that is, in quantum parlance, “nonlocal.” That means the information is shared between the two ends of the wire, a setup that protects the information.

Some scientists are positive about Microsoft’s results. The researchers demonstrated a new method that probes the nonlocal properties of the Majoranas, says physicist Kartiek Agarwal of Argonne National Laboratory in Lemont, Ill. That, Agarwal says, supports the case that these are truly Majoranas. “This is fantastic progress.”

But critics such as Legg doubt that Microsoft has a functioning topological qubit at all. Demonstrating a qubit requires two types of measurements, called X and Z measurements. The new paper presents only Z measurements. (Microsoft reported both on their previous chip, but many experts were not convinced by the X measurements.) “Nothing in the presented data proves the existence of a topological qubit or Majoranas in these devices,” Legg says.