Quantum supremacy is the point at which a quantum computer performs a task that no classical computer can realistically match. In 2019, Google claimed to have crossed that threshold with its 53-qubit Sycamore processor, which completed a specialized sampling task in about 200 seconds—a job Google estimated would take a state-of-the-art supercomputer roughly 10,000 years. Within days, IBM countered that an optimized classical supercomputer could do the same work in just 2.5 days. The dispute that followed is really a debate about what "supremacy" means and how it should be measured.
Google's 2019 Claim: The Sycamore Experiment
What exactly did Google build and run? Sycamore is a superconducting quantum processor created by Google's Artificial Intelligence division, equipped with 53 working qubits. In 2019, the Google team used it to run what they described as a quantum supremacy experiment, publishing the results in the peer-reviewed journal Nature under the title "Quantum Supremacy Using a Programmable Superconducting Processor."
The task itself was random circuit sampling. Qubits passed through randomly generated quantum circuits—quantum gates drawn from a universal gate set—followed by measurements. According to IBM's technical description of the experiment, the processor implemented a two-qubit gate circuit of depth 20, containing 430 two-qubit and 1,113 single-qubit gates, with a predicted total fidelity of 0.2%.
The headline number came from repeated runs. Google's Nature paper reports that Sycamore took about 200 seconds to sample one instance of a quantum circuit a million times. The team's benchmarks indicated that the equivalent task for a state-of-the-art classical supercomputer would take approximately 10,000 years. Google framed this dramatic speedup over all known classical algorithms as an experimental realization of quantum supremacy for this specific computational task.
IBM's Counter-Argument: A Classical Rebuttal
Why did IBM push back so quickly? Google's announcement was met with a rebuttal from its direct competitor. IBM contended that the calculation Google claimed would take 10,000 years could be performed in just 2.5 days on its own Summit supercomputer, provided its architecture were optimized for the job.
The disagreement was not about the hardware or the raw data. It was about how efficiently a classical machine could simulate the quantum experiment. IBM's researchers—Edwin Pednault, Dmitri Maslov, John Gunnels, and Jay Gambetta, writing on the IBM Quantum blog—argued that the original meaning of "quantum supremacy," as proposed by John Preskill in 2012, described the point where quantum computers can do things classical computers can't. By that definition, they wrote, the threshold had not been met.
IBM's position rested on the specific structure of the task. Because the experiment involved executing a random quantum circuit of a size infeasible for simulation with any available classical computer, the question became whether Google's circuit truly fit that description. IBM's analysis suggested an ideal simulation could be achieved with far greater fidelity than Google's 0.2% figure, and in days rather than millennia.
Google vs. IBM: The Quantum Supremacy Showdown
The table below juxtaposes the key performance metrics from Google's claim with IBM's rebuttal, showing the specific numbers at the heart of the controversy.
| Dimension | Claim |
|---|---|
| Claimed task | Random circuit sampling, performed on Google's 53-qubit Sycamore processor |
| Google's claim (quantum computer) | Sycamore sampled one instance of a quantum circuit a million times in about 200 seconds |
| Google's estimate (classical supercomputer) | Approximately 10,000 years for the equivalent task, based on the computational complexity of the simulation |
| IBM's rebuttal (classical supercomputer) | 2.5 days on IBM's Summit supercomputer, if its architecture were optimized, with greater fidelity |
The gap between 10,000 years and 2.5 days is enormous—roughly a factor of 1.5 million. That chasm explains why the milestone was immediately controversial. Both figures describe the same experiment; they differ only in assumptions about classical simulation efficiency.
The Core of the Debate: Supremacy vs. Advantage
What does the disagreement actually teach us? First, that "supremacy" is a moving target. A claim of quantum supremacy depends on the state of classical algorithms and hardware at a given moment. If a better simulation method appears, yesterday's supremacy demonstration can look less decisive today.
Second, the task matters. Random circuit sampling is a benchmark designed to be hard for classical machines—it is not a commercially useful computation in itself. Even IBM's rebuttal acknowledged, in the broader reporting on the episode, that Sycamore performing a specific calculation significantly faster than the most powerful existing supercomputer was a major scientific achievement, whatever label applied.
Third, the terminology has evolved. The term "quantum supremacy," coined by Preskill in 2012, is now often replaced by "quantum advantage," which describes a quantum computer solving a practical problem more efficiently than classical alternatives. Industry discussion has also introduced "quantum utility," referring to quantum computers delivering value on real-world problems regardless of whether they beat every classical approach. These shifts reflect a move away from a binary supremacy claim toward more nuanced assessments of practical value.
Understanding the Ongoing Significance of Quantum Supremacy
Readers should recognize that "quantum supremacy" represents a contested milestone, with its practical significance debated based on the efficiency of classical simulations. Google's 200-second run was real and peer-reviewed; IBM's 2.5-day estimate was a serious technical argument about simulation, not a dismissal. Neither settles the question definitively, and this article does not adjudicate whether supremacy was "truly" achieved.
The clearest way to track the debate's evolution is linguistic. Watch the continued use and evolution of terms like "quantum advantage" and "quantum utility" in scientific and industry discussions, reflecting a shift from a binary "supremacy" claim to more nuanced assessments of quantum computing's practical value. When researchers stop arguing about beating supercomputers on benchmarks and start reporting results on useful problems, that terminology itself will tell you where the field has moved.
