To this day, we have yet to see a quantum computer conclusively perform a single useful task. Existing machines are simply too small and error-ridden to solve commercially relevant problems. That hasn’t stopped Donald Trump’s science adviser from promising a “quantum computer powerful enough for scientific discovery by 2028” and Trump from issuing a new executive order to speed up the US quantum computing industry in its competition with China, both on June 22nd.
Companies drive the hype, too. In June, Microsoft announced a new quantum computing chip named Majorana 2. It claimed the chip was a hardware advancement that accelerates its timeline to a “scalable, practical quantum computer” by 2029. But independent experts swiftly criticized the announcement. “This is complete codswallop,” Henry Legg, a physicist from the University of St. Andrews and a longtime Microsoft critic, tells The Verge.
Legg just published a paper in Nature on June 24th criticizing Microsoft’s quantum claims from a year ago — peer review takes a long time — and pointing to what he sees as major discrepancies between Microsoft’s papers and press releases. Nature included Microsoft’s rebuttal. As the arguments continue to roil, the arc of quantum computing’s progress can seem like a mess, alternating between hyped-up announcements from companies, subsequent smackdowns from academic researchers, more fights, and, now, overconfident goals set by heads of state.
Researchers have made genuine progress in quantum computing — it’s just been largely incremental and too esoteric to immediately capture the public’s imagination. Oh, and it’s all very expensive.
Over the last decade, Google, IBM, Amazon, Microsoft, and a slew of national governments and smaller startups have poured billions into quantum computing development. Proponents predict that the technology will lead to discoveries in medicine, as well as advances in materials science and machine learning. Meanwhile, many national security experts frame its development as a new Cold War competition between the US and China.
The promise of quantum computing is that it excels at a fundamentally different type of math than classical computers. Instead of using bits like a classical computer, a quantum computer’s fundamental unit of information is the qubit. Qubits represent information as probabilities rather than ones and zeros. You can think of a qubit as a coin flipping through the air. Before the coin lands definitively as heads or tails, it is a probability of both states. Objects like molecules or processes like photosynthesis inherently involve probabilities, and thus are more “natural” for quantum computers to simulate than classical computers. However, quantum computers are unlikely to be good at classical computing tasks like email or word processing.
Companies make qubits from different materials. Several physicists The Verge spoke to said that the leading qubit types are neutral atoms, ions, and superconducting circuit qubits. Google and IBM both make qubits based on superconducting circuits. Honeywell-affiliated Quantinuum makes qubits out of individual barium ions, whereas Boston-area startup QuEra makes qubits out of individual rubidium atoms. Microsoft’s Majorana particle qubit, which experts dispute exists, is built using a thin wire attached to a superconductor. In pursuing these different approaches, the companies are throwing everything at the wall to develop quantum computing hardware that is both precise and easy to scale.
“This whole Majorana technology, it’s not a technology yet.”
Proponents of the technology say that it could solve problems that today’s supercomputers struggle with. Theoretical research indicates quantum computers should be able to simulate molecules far more easily than supercomputers. These simulations could help to develop new battery materials or medicines.
Some have imagined the quantum computer as a cyberattack tool. In 1994, computer scientist Peter Shor developed a quantum computing algorithm for factoring prime numbers that should be able to break RSA encryption, a ubiquitous family of algorithms used to secure banking and email communications. This promised cryptographic capability has motivated experts to develop more secure protocols known as post-quantum cryptography, still not in widespread use, that quantum computers should not be able to break. Their anticipation of quantum computing’s decryption capability may have rendered this application obsolete. In addition, these cryptographers didn’t actually need a quantum computer to develop a better cryptographic system, so it’s a convoluted argument for quantum computers’ utility. (On June 22nd, Trump issued another executive order aimed to “migrate” government computers to “post-quantum cryptography” by 2030 or 2031.)
Current quantum computers like Google’s Willow are individual chips too primitive to break RSA encryption or implement drug molecule simulations. But the vision is to build scaled-up machines that can. These quantum computers would be specialized data centers of many chips networked together, or perhaps specialized chips within a supercomputer, which a user would log into via the cloud. A quantum computer will not be a consumer gadget that individuals own, nor will it replace classical computers. “It’s a computer with a very specific purpose,” says Dries Sels, a physicist at Boston University.
But development toward these goal applications has not been straightforward, and researchers are still noodling over what that purpose is.
In June, IBM announced it plans to invest more than $10 billion into quantum computing over the next five years. IBM, like Microsoft, aims to build a larger-scale quantum computer by 2029. The company’s investment dovetails with an infusion of public cash into the industry. In May, the Trump administration said it would provide $2 billion in funding to nine quantum computing companies, of which IBM will receive $1 billion.
Similar cycles have played out several times since the technology’s beginnings. Companies announce a breakthrough; independent researchers cry hype, all while investors continue to inject money into the industry. In 2019, Google announced that its quantum computer had performed a task faster than the best supercomputer, a feat now known as quantum advantage. At the time, company spokespeople heralded the achievement as “quantum supremacy,” but today experts widely agree the demonstration, which involved generating random numbers, had no practical application. Regardless, quantum computing investment in 2020 accounted for a third of all investments until that point, according to McKinsey.
Last October, Google claimed it had performed another demonstration of quantum advantage. In the demonstration, Google researchers simulated molecules of 15 and 28 atoms to study their magnetic behavior in a specific scenario. A press release stated that the demonstration showed that “a quantum computer can successfully run a verifiable algorithm on hardware, surpassing even the fastest classical supercomputers.”
“We 100 percent stand behind our results. We stand by our roadmap.”
While the demonstration showed Google’s “high precision” in controlling its machine, it was a contrived experiment designed specifically to show quantum advantage rather than anything useful, says Sels. “It doesn’t simulate anything interesting,” he says. “It would be more interesting if they simulated something that classical methods have been trying for years and cannot do.”
Sels also disputes that Google bested all classical computers. While no one he knows has used a supercomputer to refute Google’s claim, he thinks it’s feasible, as he’s debunked quantum advantage claims before. But he also thinks it’s a waste of time. “Some of these problems, they’re so contrived that we really don’t want to try them,” says Sels. Previously, he has felt obligated to do this debunking as a check on industry hype, but the work isn’t scientifically interesting to him.
“I don’t know if anyone will ever really invest effort into trying to classically simulate Google’s experiment,” says Sels. “I think I won’t unless someone gives me billions of dollars for it.”
Google still thinks the study was significant. “Google’s 2025 result in Nature was the first demonstration of verifiable quantum advantage on hardware. It was not claimed as an immediate practical application, but is relevant for NMR and an indicator of rapid progress towards useful quantum computing,” wrote Google spokesperson Jason Freidenfelds in response to Sels’ criticisms.
The drama can overshadow the real progress in quantum computing. So far, a main technological challenge has been flawed qubits. They cannot execute computing operations perfectly, and the errors compound as algorithms grow longer. This has been the main snag of quantum computing: Any application of interest will require a long algorithm, but the longer the algorithm, the more error-ridden the quantum computer becomes.
Researchers have improved the qubits themselves, so they hold onto information longer. When they hold onto information longer, you can fit in more operations and do more complicated algorithms. Last November, Andrew Houck of Princeton University and his colleagues reported that they’d made a superconducting qubit that can hold onto information three times longer than the previous record holder. The key to their improvement was to make the layered substrate that the qubits sit on out of purer materials than previous chips, with careful attention to the temperature and deposition of the layers that make the chip. “It’s all very subtle tweaks,” says Houck.
And in the last two years, researchers have made substantial strides in what’s known as quantum error correction. “The advances in error correction we’ve been seeing over the past couple years are the most exciting thing going on in the field,” says Sels.
In addition, researchers have developed algorithms to correct errors while the quantum computer operates. The technique involves encoding a single unit of information in multiple qubits, rather than a single qubit, as they did in the past. They refer to the error-corrected collection as a “logical qubit” and its individual constituents as “physical qubits.”
Companies are racing to make logical qubits out of as few physical qubits as possible. In 2024, Google made a logical qubit out of 105 physical qubits. In 2025, IBM and Amazon showed they needed 12 and nine physical qubits to create a logical qubit, respectively. At the end of that year, Quantinuum showed it needed two physical qubits per logical qubit. Fewer physical qubits per logical qubit makes it easier to scale up a quantum computer.
And error correction is central to Microsoft’s controversial announcement. Microsoft claimed, which experts dispute, that it made an object made of electrons known as a Majorana particle, predicted to exist in a tiny wire made of the semiconductor indium arsenide stuck to a superconductor. Theory predicts that under specific experimental conditions, the electrons in this wire, thinner than a human hair, would perform a collective “dance” in which they start to behave in a strange unit known as a Majorana particle. In particular, researchers hypothesized that the Majorana particle should make fewer errors than other physical qubits, and thus, would be easier to scale up.
Legg says Microsoft has not successfully created a Majorana particle, the basic building block of its machine’s design. Microsoft’s approach has “fundamental issues,” he says, which were already a problem in the chip’s predecessor, the Majorana 1, released last year. “You know the phrase, ‘years, not decades’?” says Legg. “I think it’s more like centuries, not decades.”
“We 100 percent stand behind our results. We stand by our roadmap,” Microsoft’s quantum lead, Chetan Nayak, responded in an interview with The Verge. In an email statement, he added that Microsoft’s “papers do show that we are creating and controlling Majorana [particles]. He also wrote that Legg has not “proposed an alternative model that fits all of our data.”
Microsoft’s supporting evidence is unconvincing, according to Legg. What it claimed as evidence of a Majorana particle, he says, could actually be due to quantum dots forming in its device. Quantum dots are electron-containing objects that are not useful for Microsoft’s quantum computer. It also bases its claim on data from a single device, says Legg. He wants to see Microsoft replicate the results in multiple chips. “If you repeatedly try and find Jesus in your toast, eventually you’ll find Jesus in your toast,” he says. “But that one piece of toast doesn’t mean you had some kind of epiphany.”
“While we appreciate the religious fervor, our data maintains the strength and consistency of our roadmap, as we have for the past several years across previous milestones. We look forward to delivering the world’s first quantum machine and sharing the energy of our achievements with the world,” wrote Nayak in response.
Past spurious work from Microsoft-affiliated researchers adds to the doubt. In 2021, the journal Nature retracted an article from Microsoft-affiliated researchers in which they’d claimed strong experimental evidence that they’d created a Majorana particle.
“This whole Majorana technology, it’s not a technology yet,” says Rajibul Islam of the University of Waterloo.
“We are computing with these systems, and look forward to delivering a quantum computer that utilizes them to full advantage in the future,” wrote Nayak in response.
For the qubit types that experts can agree exist, companies are now promising bigger machines. By 2029, IBM plans to build a data center-sized quantum computer with 200 logical, or error-corrected, qubits. Quantinuum has set a similar goal, a machine with hundreds of logical qubits, for 2030.
“There’s no evidence of the scalability of any platform to the level that you would need to do useful quantum computations within a decade, or probably a couple of decades.”
While certainly larger, it’s unclear whether these machines will be able to do anything useful. “I’ve been saying, half-jokingly, that if someone gives me an [error-corrected] computer right now with a few hundred qubits, it’s not clear to me what we will do with it,” says Sels.
Even hopeful experts have varying opinions about when a quantum computer will demonstrate something useful. Eleanor Crane of King’s College London recently was awarded time on Google’s quantum computer to simulate a simple model of photons interacting with electrons, which occurs in both solar cells and photosynthesis. “If we were to understand this process, not only would we understand what’s happening in nature, but we would also understand how to build better solar cells,” says Crane.
She thinks that researchers will have demonstrated a useful scientific simulation on a quantum computer by 2028. Houck thinks it’s likely to happen before 2035. Crane thinks a quantum computer could break RSA encryption by 2030, while Islam thinks it will be at least a decade.
Legg is more skeptical and thinks some have underestimated the fundamental challenges of scaling. “There’s no evidence of the scalability of any platform to the level that you would need to do useful quantum computations within a decade, or probably a couple of decades,” he says.
While researchers have made progress toward building a useful quantum computer, it’s not clear what that use should be. “It’s such a nascent technology,” says Islam. “If you ask, what is a quantum computer good for, I do not know of an application which is a sure shot.”
