What Is Quantum Supremacy, and Has It Actually Been Achieved?

What is Quantum Supremacy? A High-Stakes 'Hello World' Moment

In the cutthroat world of quantum computing, one term carries more weight—and sparks more controversy—than any other: "quantum supremacy." The phrase was coined back in 2012 by Caltech theoretical physicist John Preskill to describe a critical tipping point. The moment a programmable quantum computer pulls off a calculation that no conventional supercomputer could feasibly complete. Ever. It isn't about being better at everything. It's about proving a quantum processor can solve just one problem that's effectively impossible for any classical machine, even if that problem has no practical use. Think of it less as a hostile takeover and more as a 'hello, world' moment for a new kind of computation—a demonstration that these devices, born from the strange laws of quantum mechanics, can access a computational realm forever closed to the binary world of 1s and 0s.

The idea itself isn't new. It traces back to the early 1980s, when physicists Yuri Manin and Richard Feynman argued that to simulate quantum systems, you’d need a computer built on quantum principles. Preskill’s term just captured the drama of finally building one and proving it. This milestone is a scientific benchmark. A proof of principle. It’s a stepping stone, not the final destination, which is why these experiments often run on so-called Noisy Intermediate-Scale Quantum (NISQ) devices, which are powerful but still too error-prone for truly complex applications.

The Shot Heard 'Round the World: Google's 2019 Claim

The debate was abstract. Until it wasn't. In October 2019, reality hit. Researchers from Google AI, working with NASA, published a paper in the journal Nature with a stunning announcement: they'd done it. Their 53-qubit processor, "Sycamore," performed a highly specific task—sampling the output of a random quantum circuit—in about 200 seconds. A simple-sounding feat, but the implications were huge. The team estimated that the same calculation would take the world's mightiest supercomputer at the time, IBM's Summit, a jaw-dropping 10,000 years. Google CEO Sundar Pichai compared it to the first rocket reaching space. A new frontier was open.

The announcement landed with both applause and intense scrutiny. The task was esoteric, sure. But its completion was a landmark achievement. Here, for many, was the first concrete evidence that the exponential power promised by quantum machines wasn't just theory anymore.

The Counterargument: A Supercomputer Strikes Back

The ink was barely dry. Then IBM, Google's chief rival in the quantum race, pushed back. Hard. In a blog post, its researchers argued that Google’s 10,000-year estimate was wildly inflated. With a different classical algorithm—one that cleverly used a supercomputer’s massive hard drives—they claimed Summit could solve the same problem in just 2.5 days. Not millennia. While that's still an eternity compared to Sycamore's 200 seconds, IBM argued it meant the threshold for true infeasibility, the very heart of the term, hadn't been met.

This public spat revealed just how fuzzy the definition is. What does "infeasible" really mean? Is a 2.5-day calculation on a multi-million-dollar supercomputer gulping megawatts of power truly feasible compared to a few minutes on a quantum chip? That debate rages on. The goalposts keep moving, as better classical algorithms constantly raise the bar. Any claim to supremacy, it turns out, might be temporary.

A Crowded Field: Has Quantum Supremacy Been Reached Elsewhere?

Google wasn't alone. Far from it. In December 2020, a team from the University of Science and Technology of China (USTC) led by Pan Jianwei made its own powerful claim. Their machine, a photonic quantum computer named Jiuzhang, tackled a different problem called Gaussian boson sampling. The team said their prototype detected up to 76 photons and did the job in 200 seconds. The time it would take a classical supercomputer? An estimated 2.5 billion years.

And the Chinese teams didn't stop there. In 2021, they announced the 66-qubit "Zuchongzhi 2.1" superconducting processor had achieved a quantum advantage on another task. More recently, in early 2025, they unveiled "Zuchongzhi 3.0," a 105-qubit prototype they claim performs certain tasks a quadrillion times faster than the world's best supercomputers. These successive claims, from different groups using different hardware, suggest that by some metrics, the milestone has been crossed again and again. For a field so young, this rapid-fire progress matters more than any single declaration. It's a sign of the global scale and intense pace of the history of AI and advanced computing.

Why the Term Itself Is Contested: Advantage vs. Supremacy

Beyond the technical fight, there's a linguistic and philosophical one. Is "supremacy" even the right word? Many in the field, IBM included, argue against it for two big reasons. First, it's misleading. Quantum computers will probably never be 'supreme' over classical ones. The future is almost certainly a hybrid model where quantum and classical processors work in concert, like CPUs and GPUs do today. It's about collaboration, not conquest.

Second, the word carries some ugly historical baggage. In a 2019 article for Quanta Magazine, even John Preskill himself acknowledged the term's unfortunate association with the repugnant concept of white supremacy. That's why many researchers now prefer a different term: quantum advantage. This phrasing shifts the goal. It moves from a one-time, absolute victory to the more practical aim of showing a quantum computer can beat a classical one on a useful problem—like discovering new drugs, designing materials, or optimizing financial models. While supremacy is a scientific benchmark, quantum advantage vs supremacy is about real-world application. It’s less about a theoretical stunt and more about delivering tangible value—a challenge explored in articles like how to build an AI strategy that actually delivers business value.

So, has quantum supremacy happened? The answer, unsatisfyingly, is that it depends on who you ask and how you define the terms. Google's 2019 experiment was a watershed moment, proving a quantum processor could, for one specific task, smoke our best classical machines. But the ground beneath this debate is constantly shifting. The conversation has moved on from a simple yes or no. The real question now is where, how, and when these extraordinary machines will finally start solving problems that matter. The ultimate goal isn't supremacy. It's utility.