Business conglomerate and defence contractor Honeywell announced today that it will bring what it claims is the most powerful quantum computer yet to market in the next three months.
Though quantum computers are still closer to science experiments than useful computing devices, most of the major computing companies are in the midst of developing machines of their own or offering access to these devices on the cloud. Honeywell’s newly announced machine will rely on a different underlying architecture than IBM’s or Google’s, and it may have some advantages when it comes to quantum processing power or the ability to scale up in the future. But a device’s power doesn’t really matter at a time when there aren’t actual uses for quantum computers. Instead, this announcement is significant mainly in that yet another company has introduced another architecture to compete with the biggest players in the quantum computing game today.
“We’re absolutely delighted to be in a position where we can bring this to market,” Tony Uttley, president of Honeywell Quantum Solutions, told Gizmodo. He explained that the device combined many of Honeywell’s strengths, including high-precision controllers, magnetic systems, and integrated circuits.
Quantum computers are computers based on a different architecture than classical computers, where the underlying unit used to perform computations, called a quantum bit (a qubit), obeys the same mathematical laws as electrons around an atom. Today’s quantum computers have little more capability than random-number generators and quickly lose their quantum behaviours due to interaction with the outside world. But companies and researchers developing these computers are working to overcome those limitations by creating better qubits that can stay quantum for longer.
The artificial atoms at the core of quantum offerings from Google and IBM are superconducting loops controlled by radio-frequency pulses stored in a super-cold dilution refrigerator. Honeywell, instead, has developed what it calls a trapped-ion quantum charge coupled device. A 2D system traps ions of the element ytterbium, which serve as the qubits, while a helper atom assists in cooling the ytterbium atoms. Optical pulses put the qubits into their initial states (the quantum version of setting a computer bit to 0 or 1) and then measure the qubits’ final values (either 0 or 1), while further pulses act as the gates, or quantum logic operations between two qubits.
One of the largest concerns with quantum computers today is whether or not engineers will be able to scale them up to the number of qubits required to run powerful quantum algorithms. For the short term, Honeywell’s architecture has scaling built in—they can add more qubit-containing units and can transport individual qubits between these units.
Honeywell performed a series of tests on its device in order to measure quantum volume, a benchmarking metric developed by IBM based on the number of qubits as well as how effective the device is at performing quantum logic operations. The researchers tested their device with four ions—so four qubits—and calculated it to have a quantum volume of 16. But Uttley told Gizmodo that Honeywell plans to release a commercial device in three months with a quantum volume of 64, the highest yet. They are using this test to claim that their upcoming quantum computer is the most powerful in the world.
Chris Monroe, founder of competing trapped-ion effort IonQ, was excited by Honeywell’s announcement. “The reason this is a big deal is that there must be competing architectures for progress,” he told Gizmodo in an email. Monroe explained that competing architectures like the ones employed by IBM and Google will require breakthroughs in physics to scale further, but in trapped-ion computers, scaling is more of an engineering problem. Honeywell has shown a path forward to larger quantum computers with cleaner qubits, he said.
As for the “most powerful” claim, some, several industry analysts have endorsed quantum volume as a useful method to measure quantum computers’ performance in a way that isn’t reliant solely on the number of qubits, according to the Honeywell press release. Jerry Chow, senior manager in quantum system technology at IBM Research, told Gizmodo in an emailed statement that he was excited to see IBM’s metric in use as “the best hard agnostic benchmark to measure progress of quantum computers.”
But Monroe wasn’t convinced, saying he didn’t think that quantum volume was the best metric to measure quantum computers at this stage—because it doesn’t take use cases into account. He thought it better to base power on the computer’s ability to run well-known algorithms. Uttley told Gizmodo that Honeywell had successfully run some well-known algorithms on the device.
The computer is exciting, but not because of its power. Like every other quantum computer around today, this is still just a very interesting science experiment; it’s like announcing you have the best volcano at the science fair. Basically, Honeywell has successfully demonstrated a new quantum computing architecture, an architecture that has been in the works for almost 20 years and which Honeywell has worked on for the past decade, with very clean qubits that have extremely low error rates. Honeywell also has a plan for how to scale the computer up in the near future.
In the meantime, Honeywell is partnering with JPMorgan Chase to develop algorithms for the devices and future devices based on this architecture, as well as with startups to build software. Honeywell will also offer its device for use by developers over the cloud via Microsoft’s Azure Quantum Services.
An important note is that the applications companies are pursuing when they talk about commercialising quantum computers aren’t the kind that make the world better. Uttley mentioned chemical development, the oil and gas industry, and the aerospace industry as places where Honeywell thinks its device will bring impacts in the future. These are applications that will improve weaponry and exacerbate the effects of climate change. Quantum computers will be used to make companies even wealthier and will likely have little benefit to regular people.
But perhaps it’s too early to worry about that. Those use cases are still beyond the capability of today’ “NISQ-era” quantum devices, which algorithm developers are still trying to devise useful applications for. Meanwhile, companies like Honeywell, Google, and IBM will continue racing to scale their quantum computers up.