Quantum computers turned out to be more useful than expected in 2025

Over the past year, I kept telling the same story to my editor: Quantum computers are on the verge of becoming useful for scientific discovery.

Of course, that was always the goal. The idea of ​​using quantum computers to better understand our universe is part of their origin story, and it even featured in a speech by Richard Feynman in 1981. When thinking about how best to simulate nature, he wrote: “We can abandon our rule about what the computer was, we can say: let the computer itself be built from quantum mechanical elements that obey the laws of quantum mechanics.” »

Today, Feynman’s vision has been brought to life by Google, IBM, and dozens of other companies and academic teams. Their devices are now being used to simulate reality at the quantum level – and here are some highlights.

For me, this year’s quantum developments began with two studies that landed on my desk in June that dealt with high-energy particle physics. Two separate research teams used two very different quantum computers to simulate the behavior of pairs of particles in quantum fields. One used Google’s Sycamore chip, made of tiny superconducting circuits controlled by microwaves, and the other used a chip produced by quantum computing company QuEra, based on extremely cold atoms controlled by lasers and electromagnetic forces.

Quantum fields encode how a force, such as the electromagnetic force, would act on a particle at any position in the universe. They also have a local structure that dictates the behaviors you should observe if you zoom in on a particle. Such fields are difficult to simulate in the case of particle dynamics – when the particle does something over time and you want to make something that looks like a movie out of it. For two very simplified versions of the quantum fields that appear in the standard model of particle physics, the two quantum computers tackled this precise task.

Jad Halimeh of the University of Munich, who works in the field but was not involved in either experiment, even told me that a more muscular version of these experiments, simulating more complex fields on larger quantum computers, could eventually help us understand what particles are doing inside particle colliders.

Three months later, I was on the phone with two other teams of researchers, again discussing these same two types of quantum computers, now used in condensed matter physics. Condensed matter physics is close to my heart because I studied it in graduate school, but its impact extends far beyond this columnist’s inclinations. This has been particularly crucial for the development of semiconductor technologies that underpin everyday devices such as smartphones.

In September, researchers from Harvard University and the Technical University of Munich in Germany used quantum computers to simulate two exotic phases of matter that had been predicted in theory but had eluded more traditional experiments. Quantum computers have proven capable of predicting the properties of these strange materials, something that growing and analyzing crystals in the laboratory has so far failed to accomplish.

October brought the prospect of practical use of a new superconducting quantum computer from Google, called Willow. The company’s researchers and their colleagues used Willow to run an algorithm that could be used to interpret data from nuclear magnetic resonance (NMR) spectroscopy, a technique commonly used to study molecules in biochemical research.

Although the team’s demonstration with real NMR data failed to accomplish anything a traditional computer couldn’t, the mathematics of the algorithm promises to one day surpass the capabilities of classical machines, allowing researchers to learn unprecedented details about molecules. How quickly this comes to fruition depends on the rate at which quantum computing hardware improves.

A month later, a third type of quantum computer entered the conversation. A company called Quantinuum has shown that its Helios-1 quantum computer, made from trapped ions, can run simulations of a mathematical model for perfect electrical conductivity, or superconductivity. Because they conduct electricity without any losses, superconductors can open the door to extremely efficient electronics or even make the power grid more sustainable. However, all known superconductors operate only under high pressure or extremely low temperatures, making them impractical. A mathematical model that reveals exactly why certain superconducting materials would provide a crucial stepping stone toward building useful superconductors.

Helios-1 simulated what Henrik Dryer, a researcher at Quantinuum, told me was probably the most important model of its kind; one that has captured the attention of physicists since the 1960s. And while this specific simulation didn’t offer radical new insights into superconductivity, it nevertheless announced that quantum computers would be valuable players in physicists’ long-standing quest to better understand them.

Just a week later, I found myself on the phone with Sabrina Maniscalco of quantum algorithm company Algorithmiq, to ​​discuss metamaterials. These are materials whose microscopic details can be engineered to have special properties that natural materials do not have. They can also be custom-made for specific purposes, from rudimentary invisibility cloaks to chemical ingredients capable of speeding up reactions.

Metamaterials are also something I had dabbled in as a graduate student, and Maniscalco’s team figured out how to simulate one using an IBM quantum computer made from superconducting circuits. Specifically, they could track how a metamaterial scrambles information, including in regimes where a more conventional computer might struggle. While this might seem like a rather abstract setup, Maniscalco told me it could advance research into chemical catalysts as well as solid-state batteries and some devices that convert light into electricity.

As if particle physics, new phases of matter, molecular research, superconductors and metamaterials weren’t enough, while I was writing this article I received a tip about a study in which a team of researchers from the University of Maryland in the United States and the University of Waterloo, Canada, used a trapped ion quantum computer to determine how particles bound by the strong nuclear force behave at different temperatures and densities. Some of this behavior is thought to occur inside neutron stars, which are poorly understood cosmic objects, and also to have occurred in the early universe.

Even though the team’s quantum calculation involved approximations that don’t quite fit the most realistic models of the strong force, the study makes the case for another area of ​​physics in which quantum computers are thriving as discovery machines.

Of course, with this abundance of examples also comes an abundance of caveats and question marks. Most mathematical models simulated on quantum hardware require a number of simplifications and approximations compared to the most realistic models, most quantum computers are still so error-prone that they require post-processing of the results of their calculations to mitigate or remove these errors, and the question of benchmarking the results of quantum computers against what the best conventional computers can do remains thorny.

Simply put, traditional methods of calculation and simulation are another area where progress has been rapid and encouraging, placing classical and quantum computing researchers in a dynamic back-and-forth where yesterday’s most complex or fastest calculation inevitably becomes tomorrow’s runner-up. Over the past month, IBM has even teamed up with several other companies to launch a publicly available “quantum advantage tracker,” which will eventually become a ranking showing where quantum computers are or are not ahead of their conventional counterparts.

But even though quantum computers won’t be at the top of my list anytime soon, last year’s reporting still shifted my priorities toward excitement and anticipation. In effect, these experiments move quantum computers from being a subject of scientific study to being a tool for doing science, which was impossible just a few years ago.

At the start of this year, I expected to write primarily about benchmarking experiments, in which quantum computers run protocols that emphasize their quantumness rather than solving useful problems. Such calculations often serve to highlight how quantum computers are different from conventional computers and can highlight their potential to do radically new things. But the path to a useful calculation for a working physicist seemed long and not at all obvious. Now, albeit cautiously, I think this route might be shorter than I expected. I am sure that more quantum surprises will await me in 2026.