Google's Supremacy Claim, IBM's Rebuttal and the Classical Catch-Up: A Case Study
Why start with this story
If you only learn one case study about quantum hype, make it this one. It has everything: a bold headline, a big number, a rival lab pushing back, follow-up papers over several years, and a final lesson that is more nuanced than "the claim was false". It also shows how science is supposed to work, with claims tested in public. For the wider background see Quantum Supremacy vs Quantum Advantage Explained and Quantum History Part 4.
What Google reported in 2019
Google's research blog, dated October 23, 2019, described its Sycamore processor, a 54-qubit device whose hardest circuits used 53 qubits. Google reported that the machine performed the target computation in 200 seconds, and said the same output would take the world's fastest supercomputer 10,000 years. The figure caption notes that this estimate assumed 1 million CPU cores and a method called Schrodinger-Feynman simulation.
Three details matter. The 10,000-year number is an estimate by the claimant, not a measured run. The task, random circuit sampling, is a benchmark designed to be hard for classical machines, and the blog does not present it as useful work. And the team checked smaller circuits against classical simulation, then pushed to sizes where checking became infeasible.
IBM's reply, October 22, 2019
IBM, a rival, published a blog and a paper (arXiv:1910.09534) arguing that an ideal simulation of the same task could run on a classical system in about 2.5 days. The key idea was storage: Google's estimate leaned on a method that trades memory for time, while IBM's approach also used hard drive space on the Summit supercomputer to hold the quantum state. The IBM paper itself says such circuits can be simulated at high fidelity "in a matter of days".
Notice what IBM did not say. It did not claim to have run the full calculation in 2.5 days. It argued the task could be done, and it argued the word "supremacy" was being misunderstood. Both sides were making projections, which is normal, and also a reason to stay humble about either number.
The classical catch-up continues
- 2021 preprint: a paper (arXiv:2111.03011) reported generating one million uncorrelated bitstrings for the 53-qubit, 20-cycle Sycamore circuit in about 15 hours on a cluster of 512 GPUs, and projected that an exascale machine could take a few dozen seconds. The fidelity achieved was low, about 0.0037, and the fast time is a projection for an ideal implementation.
- 2022 theory result: Aharonov, Gao, Landau, Liu and Vazirani posted a polynomial-time classical algorithm for noisy random circuit sampling. The authors say it is not practical as it stands and does not settle finite-size experiments, but it is evidence that noisy random circuit sampling may not scale as a lasting quantum win.
What came next: Willow in 2024
Google kept using the same benchmark. In its Willow announcement the company reported that the 105-qubit chip finished a random circuit sampling computation in under five minutes that it estimates would take the Frontier supercomputer about 10 septillion years. Google itself says the estimate uses generous assumptions for Frontier, expects classical methods to keep improving, and states that the benchmark has no known real-world application. Read Google Quantum AI for the full picture. That candour is worth noticing: the company's own page contains the caveat that headlines often drop.
Lessons for reading any headline
- Ask who made the classical estimate. If the claimant also estimated the competition, treat the number as a starting point.
- Ask what the task is. A task built to be hard for classical computers tells you about the boundary, not about usefulness.
- Expect rebuttals. Rival labs and academics will try to simulate the result. A claim that survives years of attempts is stronger.
- Check the fidelity. A classical shortcut that reaches low fidelity is not the same as matching the experiment.
- Separate the machine from the milestone. Sycamore and Willow were real engineering achievements even though the benchmark debate continues.
What it means for a memecoin holder
Quantum headlines can move attention toward anything quantum-themed, including community tokens. Attention is not value. The QNT token is an independent memecoin, not connected to Google, IBM or Quantinuum Ltd. Nothing here is financial advice. See QUANTUM (QNT) Risk Factors and QUANTUM (QNT) Memecoin vs Quantinuum Ltd.
Sources and further reading
- Google Research blog: quantum supremacy using a programmable superconducting processor (Oct 23, 2019)
- IBM: on quantum supremacy (Oct 22, 2019)
- arXiv:1910.09534, Leveraging secondary storage to simulate deep 54-qubit Sycamore circuits
- arXiv:2111.03011, solving the sampling problem of the Sycamore circuits
- arXiv:2211.03999, a polynomial-time classical algorithm for noisy random circuit sampling
- Google: Willow quantum chip announcement
Reported as of 2026-10-09. Scientific debates move, so check the newest papers before relying on any claim. This page is education, not financial advice, and it makes no price predictions. The QNT memecoin is independent of Quantinuum Ltd, the real company, and of every lab, company and regulator named here.
Frequently asked questions
Did IBM prove Google's supremacy claim wrong?
Not by running the full task. IBM argued in October 2019 that a better classical method could do it in about 2.5 days, which is a projection. Later classical papers narrowed the gap further, but the debate is about the benchmark and the word supremacy, not about whether Sycamore worked.
Is random circuit sampling useful?
Google itself says it has no known real-world applications. It is a test of whether a quantum chip can beat classical machines on a task designed to be hard.
Why do headline comparisons say 'years' or 'septillion years'?
They are estimates made by the claimant for a specific classical method. A new classical algorithm can shrink them a lot.
Does this story affect the QNT memecoin?
No. QNT is an independent community memecoin with no link to Google, IBM or Quantinuum Ltd. Nothing here is financial advice.
Keep reading
- Quantum Supremacy vs Quantum Advantage Explained
What the terms quantum supremacy and quantum advantage mean, and why headline claims are often debated. - Google Quantum AI: Willow, Error Correction and Quantum Echoes
What Google's Willow chip and the Quantum Echoes result showed, why below-threshold error correction matters, and what is still unproven. - Quantum Benchmarks Explained: Why Qubit Count Misleads
How do you measure a quantum computer? Learn what error rates, fidelity, coherence time and quantum volume mean, and why qubit count alone is not enough. - Quantum History Part 4: NISQ and the Supremacy Claims (2018 to 2022)
Preskill names the noisy era, Google claims a milestone in 2019, IBM pushes back, USTC joins in, and the qubit counts climb.
All Quantum computing guides | Back to top | Search the site
Main pages: Quantum computing explained | Quantum and crypto | Companies | Quantum news | Glossary