Quantum computers can’t teleport things—yet

Last November, a group of physicists claimed they’d simulated a wormhole for the first time inside Google’s Sycamore quantum computer. The researchers tossed information into one batch of simulated particles and said they watched that information emerge in a second, separated batch of circuits. 

It was a bold claim. Wormholes—tunnels through space-time—are a very theoretical product of gravity that Albert Einstein helped popularize. It would be a remarkable feat to create even a wormhole facsimile with quantum mechanics, an entirely different branch of physics that has long been at odds with gravity. 

And indeed, three months later, a different group of physicists argued that the results could be explained through alternative, more mundane means. In response, the team behind the Sycamore project doubled down on their results.

Their case highlights a tantalizing dilemma. Successfully simulating a wormhole in a quantum computer could be a boon for solving an old physics conundrum, but so far, quantum hardware hasn’t been powerful or reliable enough to do the complex math. They’re getting there very quickly, though.

[Related: Journey to the center of a quantum computer]

The root of the challenge lies in the difference of mathematical systems. “Classical” computers, such as the device you’re using to read this article, store their data and do their computations with “bits,” typically made from silicon. These bits are binary: They can be either zero or one, nothing else. 

For the vast majority of human tasks, that’s no problem. But binary isn’t ideal for crunching the arcana of quantum mechanics—the bizarre rules that guide the universe at the smallest scales—because the system essentially operates in a completely different form of math.

Enter a quantum computer, which swaps out the silicon bits for “qubits” that adhere to quantum mechanics. A qubit can be zero, one—or, due to quantum trickery, some combination of zero and one. Qubits can make certain calculations far more manageable. In 2019, Google operators used Sycamore’s qubits to complete a task in minutes that they said would have taken a classical computer 10,000 years.

There are several ways of simulating wormholes with equations that a computer can solve. The 2022 paper’s researchers used something called the Sachdev–Ye–Kitaev (SYK) model. A classical computer can crunch the SYK model, but very ineffectively. Not only does the model involve particles interacting at a distance, it also features a good deal of randomness, both of which are tricky for classical computers to process.

Even the wormhole researchers greatly simplified the SYK model for their experiment. “The simulation they did, actually, is very easy to do classically,” says Hrant Gharibyan, a physicist at Caltech, who wasn’t involved in the project. “I can do it in my laptop.”

But simplifying the model opens up new questions. If physicists want to show that they’ve created a wormhole through quantum math, it makes it harder for them to confirm that they’ve actually done it. Furthermore, if physicists want to learn how quantum mechanics interact with gravity, it gives them less information to work with.

Critics have pointed out that the Sycamore experiment didn’t use enough qubits. While the chips in your phone or computer might have billions or trillions of bits, quantum computers are far, far smaller. The wormhole simulation, in particular, used nine.

While the team certainly didn’t need billions of qubits, according to experts, they should have used more than nine. “With a nine-qubit experiment, you’re not going to learn anything whatsoever that you didn’t already know from classically simulating the experiment,” says Scott Aaronson, a computer scientist at the University of Texas at Austin, who wasn’t an author on the paper.

If size is the problem, then current trends give physicists reason to be optimistic that they can simulate a proper wormhole in a quantum computer. Only a decade ago, even getting one qubit to function was an impressive feat. In 2016, the first quantum computer with cloud access had five. Now, quantum computers are in the dozens of qubits. Google Sycamore has a maximum of 53. IBM is planning a line of quantum computers that will surpass 1,000 qubits by the mid-2020s.

Additionally, today’s qubits are extremely fragile. Even small blips of noise or tiny temperature fluctuations—qubits need to be kept at frigid temperatures, just barely above absolute zero—may cause the medium to decohere, snapping the computer out of the quantum world and back into a mundane classical bit. (Newer quantum computers focus on trying to make qubits “cleaner.”)

Some quantum computers use individual particles; others use atomic nuclei. Google’s Sycamore, meanwhile, uses loops of superconducting wire. It all shows that qubits are in their VHS-versus-Betamax era: There are multiple competitors, and it isn’t clear which qubit—if any—will become the equivalent to the ubiquitous classical silicon chip.

“You need to make bigger quantum computers with cleaner qubits,” says Gharibyan, “and that’s when real quantum computing power will come.”

[Related: Scientists eye lab-grown brains to replace silicon-based computer chips]

For many physicists, that’s when great intangible rewards come in. Quantum physics, which guides the universe at its smallest scales, doesn’t have a complete explanation for gravity, which guides the universe at its largest. Showing a quantum wormhole—with qubits effectively teleporting—could bridge that gap.

So, the Google users aren’t the only physicists poring over this problem. Earlier in 2022, a third group of researchers published a paper, listing signs of teleportation they’d detected in quantum computers. They didn’t send a qubit through a simulated wormhole—they only sent a classical bit—but it was still a promising step. Better quantum gravity experiments, such as simulating the full SYK model, are about “purely extending our ability to build processors,” Gharibyan explains.

Aaronson is skeptical that a wormhole will ever be modeled in a meaningful form, even in the event that quantum computers do reach thousands of qubits. “There’s at least a chance of learning something relevant to quantum gravity that we didn’t know how to calculate otherwise,” he says. “Even then, I’ve struggled to get the experts to tell me what that thing is.”