When something in a research lab works that wasn’t supposed to, there’s a certain silence. It’s more like the group holding their breath before anyone dares to say it aloud than a celebration, at least not yet. On October 22, 2025, Google Quantum AI’s 105-qubit Willow processor finished a calculation that, based on the majority of reasonable timelines in the field, shouldn’t have been possible for another ten years. That’s about how I imagine the room felt.
The majority of announcements about quantum computing are theatrical. Strong statistics, well-crafted press releases, and then a protracted period of silence during which the real world doesn’t really change. It feels different with this one. It’s difficult to ignore the fact that since Willow’s findings were released, even the skeptics—of which there are many in the quantum computing community—have been remarkably silent.
| Category | Details |
|---|---|
| Company Name | Google Quantum AI (a division of Alphabet Inc.) |
| Founded | 2012 |
| Headquarters | Mountain View, California, USA |
| Parent Company | Alphabet Inc. (NASDAQ: GOOG / GOOGL) |
| Key Hardware | Willow — 105-qubit quantum processor |
| Landmark Achievement | First verifiable quantum advantage on real hardware |
| Algorithm Used | Quantum Echoes (based on out-of-time-order correlators) |
| Speed Advantage | 13,000× faster than leading classical supercomputers |
| Announcement Date | October 22, 2025 |
| Key Collaborator | University of California, Berkeley |
| Reference Website | quantumai.google |
The chip used an algorithm called Quantum Echoes, which is based on an out-of-time-order correlator in mathematics. To put it simply, the algorithm sends a signal through a quantum system, purposefully disrupts it, flips the signal, and then observes how the disturbance reverberates throughout the system. It’s like listening to quantum memory; it sounds almost poetic. Simulating this process is an almost intolerable computational burden for classical supercomputers. Willow completed the task about 13,000 times more quickly than the world’s greatest classical machine.
But the speed wasn’t the only thing that alarmed physicists. It was the ability to be verified. Previous claims of quantum supremacy, such as Google’s own 2019 Sycamore demonstration, yielded technically impressive but scientifically limited results. The answers were not really cross-checked. The calculation was a kind of performance, impressive but not particularly helpful.
It is truly novel that Willow’s Quantum Echoes results could be independently confirmed on a different quantum device. It indicates that the machine didn’t just run quickly; it also completed the task correctly, as verified by another person.
A problem that seemed almost cruelly ironic has plagued quantum computing for thirty years: the more qubits you added to a processor, the more errors you introduced. It was similar to attempting to construct a taller tower out of blocks that lose stability as you stack them higher. Researchers were aware of the theory and that error correction was crucial, but their hardware attempts consistently failed. Apparently, Willow passed through that wall.
Google showed what is known as “below threshold” operation, which is a situation in which increasing the number of qubits actually lowers errors rather than increases them. The curve finally bent in the proper direction after decades of going in the wrong direction.
The figures associated with Willow’s performance—99.97% fidelity for single-qubit operations and 99.88% for two-qubit entangling gates—are the kind that engineers are hesitant to state out loud. These are not theoretical forecasts or simulated benchmarks. These are measured outcomes from a functional 105-qubit chip functioning in the actual, noisy, imperfect world of a laboratory. As Google moves toward larger systems, it’s still unclear if these fidelity rates will hold, but the direction is correct in a way that has never been the case before.
This becomes viscerally real in the field of drug discovery. It has long been recognized by pharmaceutical researchers that precisely simulating molecular interactions at a quantum level would revolutionize the development of pharmaceuticals by identifying the ideal molecular candidate through simulation rather than years of trial and error.
The issue has always been that these quantum systems could only be approximated by classical computers, and in the medical field, approximations can cost decades or even lives. Google studied two real molecules using the Willow chip and the Quantum Echoes algorithm in partnership with the University of California, Berkeley, and the outcomes matched those obtained using conventional techniques. It was a brief experiment. However, it indicated a massive location.
IBM has been advancing concurrently, recently showcasing quantum speedups in bond trading computations through the combination of quantum processors and traditional machine learning. Better hardware, more intelligent algorithms, and better error handling all seem to be coming together at once in the industry, and the timeline that was previously thought to be measured in generations may actually be measured in years.
Given how this has developed over the past year, it’s possible that the truth is that no one truly knows how close practical quantum computing is. Five years seems hopeful. Twenty years seems like too much caution. The unpleasant reality is that Willow drastically changed the estimate in one direction without providing anyone with a fresh figure to replace the previous ones. Google’s own language is cautious; it “paves a path toward real-world applications,” which is a phrase that covers a lot of ground and has some meaning.
What is evident is that the fundamental challenge that made quantum computing seem so unapproachable—the error correction problem—has been solved in a significant and repeatable manner. There is no commercial version of the Willow processor.
It’s not optimizing logistics for shipping companies or running pharmaceutical simulations for hospitals. However, it illustrated a concept that makes those things feasible in a timeframe that was previously untenable.
There is a sense that something genuine occurred in late 2025, which is cautiously spreading through investment circles and research communities. Not the arrival of quantum computing, but the demonstration that it is no longer a question of whether it will happen. The story is that the question has changed.
