Earlier this week, Honeywell announced “a breakthrough in quantum computing that accelerates the capability of quantum computers and will enable the company to release the world’s most powerful quantum computer within the next three months.” There’s a lot to unpack here, but the bottom line is this: Honeywell is making a big claim that is difficult to back up. The validity of this claim depends on two things: quantum volume being an accepted measure and time working in Honeywell’s favor.
The basic units of information in classical computing are binary digits (bits) that are always in a state of 0 or 1. Quantum bits (qubits) meanwhile can be in a state of 0, 1, or a superposition of the two. Quantum computing leverages qubits to perform computations that would be much more difficult, or simply not feasible, for a classical computer. The resulting applications include everything from cryptography and optimization to machine learning and materials science. In other words, any industry impacted by computing today will be impacted by quantum computing tomorrow.
But tomorrow is a long way away. Quantum computing is still so new that companies haven’t even settled on the equivalent of classical computing’s transistor. Major players are jostling to lead the space, or at least set the terms for what leadership might look like — including establishing a framework for comparing quantum computers. Because no clear leader has emerged, Honeywell is able to claim that title.
What is quantum volume?
IBM invented the phrase “quantum volume” three years ago. Here’s how the company described it in January, when it achieved a quantum volume of 32:
Quantum Volume (quantum volume) is a hardware-agnostic metric that we defined to measure the performance of a real quantum computer. Each system we develop brings us along a path where complex problems will be more efficiently addressed by quantum computing; therefore, the need for system benchmarks is crucial, and simply counting qubits is not enough. As we have discussed in the past, quantum volume takes into account the number of qubits, connectivity, and gate and measurement errors. Material improvements to underlying physical hardware, such as increases in coherence times, reduction of device cross talk, and software circuit compiler efficiency, can point to measurable progress in quantum volume, as long as all improvements happen at a similar pace.
Unsurprisingly, Honeywell backs up IBM’s description.
“Quantum volume is currently the fairest metric for demonstrating a quantum computer’s full capabilities,” Honeywell Quantum president Tony Uttley told VentureBeat. “It goes beyond simple qubit number to identify the size of the problem that a quantum computer is capable of running. It accounts for error, cross talk, and qubit connectivity, in addition to the number of qubits.”
Honeywell projects that its quantum computer will have a quantum volume of 64, or “twice that of the next alternative in the industry” — hence its claim of “world’s most powerful quantum computer.”
The problem with quantum volume
The “next alternative in the industry” is IBM, which also happens to be the sole alternative. IBM and Honeywell are the only companies that use quantum volume as a measuring stick. Minutes after Honeywell’s announcement dropped, IBM emailed us to say it is “excited to see the wider quantum computing community embrace the quantum volume metric, which IBM first developed in 2017.”
The metric determines the performance of a quantum computer for any underlying qubit type, accounting for gate and measurement errors, device cross talk, as well as device connectivity and circuit compiler efficiency. This result confirms quantum volume is the best hard agnostic benchmark to measure progress of quantum computers. The higher the quantum volume, the more of the quantum computational space can potentially be used for solving challenging problems on real near-term hardware. IBM has reached a quantum volume of 32 on a 28-qubit superconducting system, which was announced in January of this year — and the company expects to continue to double quantum volume every year. The scaling and doubling of quantum volume each year we refer to as Gambetta’s Law. Honeywell’s paper shows exciting new progress in programmable trapped-ion quantum systems.
We reached out to a handful of quantum computing companies about quantum volume. The response was a collective shrug. D-Wave said quantum volume doesn’t fully capture the nuances of different approaches to quantum computing and applications. IonQ pointed to the fact that quantum volume doesn’t take use cases into account. And Rigetti said it doesn’t have a strong stance on quantum volume because it only cares about quantum advantage, the inflection point when quantum computers begin to solve real, useful problems. Of course, the first step is quantum supremacy, when a quantum computer can solve a problem that classical computers practically cannot.
But let’s assume IBM really has found a good yardstick for measuring quantum computers. The other issue with Honeywell’s claim is that the company hasn’t launched its quantum computer. A lot can happen in three months.
Just take a look at the previous quarter. In November, Microsoft announced Azure Quantum, a cloud service that will let you tap into quantum hardware providers Honeywell, IonQ, and QCI. In December, AWS announced Amazon Braket, a cloud service that will let you tap into quantum hardware providers D-Wave, IonQ, and Rigetti. In the span of two months, the two biggest cloud providers decided they were going to shake up the quantum computing space.
Within the next three months, IBM could release a quantum computer with a quantum volume of 64. Or a third contender that uses quantum volume could enter the race and happen to hit that number. Most importantly, even if quantum volume was the accepted metric, Honeywell can’t claim “the world’s most powerful quantum computer” before its release.
So why is Honeywell counting its chickens before they hatch? In short, because quantum volume is an unused metric. IBM expects to double its quantum volume every year — and it already hit 32 this year. Honeywell is thus betting that IBM won’t have a breakthrough in the next three months and nobody else will bother with quantum volume. That’s certainly possible.
Ultimately, businesses only care whether quantum computing can bring results. IBM will tell you its customers are pleased, and its competitors will say the same. We can’t ask Honeywell, of course, because it doesn’t have any customers yet (JPMorgan Chase will be the first once the quantum computer actually debuts).