Honeywell has built the most powerful supercomputer in the world
Crank it up.
Honeywell, a Charlotte, North Carolina-based multinational conglomerate perhaps best known for its home security systems and thermostats, has stunned the world with its announcement of the world’s most powerful quantum computer. The company first announced its quantum capabilities in 2018, but nobody saw a play for global dominance in its future. Well, almost nobody.
“It’s been the goal for four years. We’ve been in this for a decade now with a very thoughtfully laid out strategically laid out plan to get to this point,” says Tony Uttley, the president of Honeywell Quantum Solutions, sitting in the Inverse office.
Uttley speaks with utter confidence about the company’s new announcement. For him, it’s not a surprise at all that Honeywell, which deals in industries ranging from aerospace to factory safety to chemicals, would go into quantum.
“When people think quantum computer, they go, “Honeywell, why would you be doing it?” They don't know what it takes to make a quantum computer. This is not a laptop. This is a room-size apparatus now that takes into account ultra high vacuum chambers and cryogenic systems and magnetic field generating systems and vibrating field ultra high precision control,” he says.
In other words, all things Honeywell already had.
The 114-year old company is measuring its quantum power in a way established by one of its competitors, the equally venerable IBM. It’s not pure qubit power, like the “quantum supremacy” Google claimed last year with 53 qubits.
Computers use bits to transfer information. The more bits, the more data that can be transferred: think of the leap from the Super Nintendo (16 bits) to the N64 (64 bits).
Quantum information is transferred in qubits, which have the same purpose as traditional computer bits but are radically more powerful. These qubits can eventually form quantum gates, which can lead to quantum circuits. That's the measure Google was using.
Honeywell begs to differ. it’s measuring what IBM first called “quantum volume,” which looks at a quantum machine more holistically, taking into account “the number of qubits, connectivity, and gate and measurement errors.”
“The larger the quantum volume, the more complex problems you can solve,” said Dr. Patty Lee, Chief Scientist for Honeywell, speaking in a press statement. “It speaks to the quality of the qubits,” said Paul Smith-Goodson, a veteran quantum analyst with Moor Insights, says in the same statement.
"It really works!”
What does any of this mean? The problem with understanding quantum anything — be it mechanics, entanglement, or computers—is that the quantum world is radically different than our own. Wrapping one’s head around the sheer smallness of atoms is one thing, but then try taking on the fact that at that level the rules of basic physics as we know them are whisked away like spoiled children. Your phone can’t exist in two places, but atoms are a different story.
While bits are limited to being either a 0 or a 1, qubits can be 0 and 1 at the same time, thanks to what’s called a coherent superposition. Seeing superposition would be very cool, but it is impossible to see. Every time it’s viewed, it loses its quantum state. Keeping a coherent superposition is very hard, and the laws of physics dictate that any quantum state that isn’t perfectly isolated will eventually fall into decoherence.
Honeywell can’t completely solve that conundrum, but its new computer offers the next-best thing: being able to pause a quantum calculation midway through. It’s called a mid-circuit calculation and “it’s only possible because of how long we can keep the coherence of the qubits, combined with the precision control we have,” Uttley says. “I mean, real-time engine precision control.”
The perfect trap
Unrecognizable to a modern laptop user — On a physical level, Honeywell’s quantum computer would be unrecognizable to a modern laptop user. All of those high-vacuum chambers and cryogenic chambers and lasers, which Honeywell made custom for this project (Uttley estimates that 90 percent of the computer is bespoke), to support one thing, something that Uttley has brought to Inverse HQ: a ion trap.
He hands me a handheld eyepiece, similar to a jeweler's, and tells me to examine what looks like a computer chip really close. Closer than that. Like, right up to my eye.
Eventually I see a small black rectangle. It looks like the monolith in 2001: A Space Odyssey that trigger evolutionary shifts when discovered. Uttley prefers to compare it to an airplane runway.
An ion trap uses both magnetic and electric fields to capture and manipulate atoms, allowing for Uttley and his team to control atoms and their spin. But it doesn’t work alone.
“If you were to cup your hands together, that’s a trillion trillion atoms in that space,” Uttley says, clasping his hands. “So we have something the size of a basketball and that is an ultra-high vacuum chamber. We pump it down so low that it has five times less particles than outer space. There’s a tiny hole where we insert the atoms. It gets photo-ionized by a laser when it comes through, gives it a charge. And as soon as it gives it a charge, we can manipulate that qubit with an electromagnetic wave form.
So think about it as, in this little runway, there would be qubits floating on top of it. You would create a wave so these would become little qubit surfers, so we have the waves that come together very gently, we can spin them around, we can move them apart and then engage them in other qubits. All because we can control the field within that tiny space.”
During our conversation, nothing gets Uttley quite as excited as these ion traps. When asked if the traps are unique to Honeywell, he offers a definitive “Yes.” Others use traps, he says, but they lacked Honeywell’s institutional and financial resources.
“We started with our own foundry that we use to make integrated circuits,” Uttley says. “That has been a part of our aerospace business for decades. That same technology gets used to basically grow these traps. We started down a path to experiment with that. We found out that not only could we do it, we could do it really well. So we built on that capability. If we can make these, and we can make them well, and make them with enough sophistication to have control over all the situations — you can’t see it, but even in that space there are hundreds of electrodes going down each side, which allows us to change the voltage so minutely that creates these waveforms.
And we got to be experts at it.”
Nothing small about quantum cash
Honeywell is currently ranked 77th on the Fortune 500. The company’s investment in quantum, which Uttley calls “technical debt,” has the potential to help its own engineers on any number of levels. But finding clients for a quantum computer presents challenges of its own. You have to find companies with people who can use quantum computers, in the first place. It’s accessible via cloud—Honeywell has its own API interface—but someone has to know what they’re doing. And then there has to be a clear situation in which they’ll need to use a computer capable of processing information at an exponentially faster rate than any classical computer in existence.
JP Morgan Chase fit the bill. The financial services giant has been beta testing Honeywell’s quantum computing abilities for three months now, and Uttley throws out a number of ways they could be using it: having AI run internal hypotheticals on the stock market, more powerful fraud protection.
It’s here, doing work, where some of Honeywell’s offerings can make themselves felt. The mid-circuit calculation. Measuring a qubit, for example, typically takes it out of the quantum state. As soon as you observe atoms in superposition, they immediately leave.
“But the incredible thing is,” Uttley says, “because of these qubit servers, we can move these far enough away, that when we measure that one, all the others stay in that quantum state. When we take that answer, and we say, oh we’re going to do something else with these, or we’re going to reinitialize that same qubit and turn it back into a zero or a one.”
Uttley says that he thinks about the historic nature of team’s project “all the time,” although he feels like “just a footnote” in the field’s long history, which started around the beginning of the 20th century with giants like Max Planck, Albert Einstein and Niels Bohr. “The geniuses that had to exist for the whole industry to be where we are and the geniuses that we have had on the team to make this happen are incredible. Just incredible.”
Even in the midst of a very corporate project, and amidst the highly competitive nature of titles like “world’s most powerful,” the sheer science of the project still manages to stun him.
“Every time we make little videos of the choreography of these little atoms moving around, when you stop and think that those are individual atoms that are moving to do what we do, just blows the mind that we have the ability to go do that now,” Uttley says.
“Combine that with quantum physics—hey, these two things are in a superposition right now! Hey, these things are entangled! Blows me away. It really works!”