Quantum Supremacy at Seven: What Google and IBM Actually Delivered

Google's claim of quantum supremacy, published in Nature on 23 October 2019, capped a two-year race that began with John Martinis promising a milestone by the end of 2017. The company says its 53-qubit Sycamore processor took 200 seconds on a sampling task that would take a state-of-the-art supercomputer roughly 10,000 years.

That is the headline number. The argument underneath it is narrower, and the people who build competing machines have spent the years since explaining why. What follows is a reconstruction of how the milestone was set, met and disputed, drawn from the reporting that covered it.

The deadline Martinis set himself

In April 2017, MIT Technology Review profiled Martinis, who joined Google in late 2014 from the University of California, Santa Barbara, and led a group of roughly 25 researchers. He told the magazine his team would build a device that achieved quantum supremacy by the end of that year, meaning it could perform a particular calculation beyond the reach of any conventional computer. Proof would come from a drag race between Google's chip and one of the world's largest supercomputers.

At the time, Google had released results from a nine-qubit chip arranged in a line. The newest device had only six qubits, but they sat in a two-by-three grid, which Martinis said showed the technology still worked when qubits were placed side by side. The six-qubit chip also tested a manufacturing method in which qubits and their control wiring are made on separate chips and later bump bonded together, a process Martinis said was working.

He said designs for 30 to 50 qubit devices were already in progress, and that he would need a grid of 49 qubits for the supremacy experiment.

"We think we're ready to do this experiment. It's something we can do now," Martinis told MIT Technology Review.

Not everyone framed the target the same way. Chris Monroe, a University of Maryland professor and cofounder of the startup IonQ, called it an academic milestone. "Afterward you still have to figure out how to make it more scalable and programmable," he said. Simon Gustavsson, a principal investigator in a quantum computing group at MIT, put Google and IBM on comparable footing at that stage.

IBM answers with 50 qubits

IBM moved first on the raw count. On 10 November 2017, the Associated Press reported that IBM scientists had built and measured a processor prototype with 50 qubits. Dario Gil, who led IBM's quantum computing and artificial intelligence research division, said it was the first time any company had built a quantum computer at that scale.

Seth Lloyd, an MIT mechanical engineering professor not involved in the work, said IBM likely still had glitches to work out, but called the announcement a sign of significant progress. The AP did not report a benchmark result attached to the 50-qubit prototype. Scale and demonstrated performance were already separating into two different conversations.

By December 2017, Martinis was making that distinction explicit at a conference. Gizmodo reported from the Q2B event at NASA Ames that he pushed back on qubit counting as the scoreboard. "Press releases always talk about quantum space race in number of qubits," he said. "It's more than just quantity, it's qubit quality."

At the same conference, Martinis said Google was fabricating its 49 or 50 qubit supremacy device that month and would begin testing within two weeks, not counting Christmas delays. He described the experiment: entangle the qubits, let the system evolve, then measure outcomes whose probabilities are so complex that a classical machine must simulate the quantum computer to check them, possibly taking weeks to do what the quantum device does in minutes.

Caltech theorist John Preskill, speaking at the same event, named the period the NISQ era, for noisy intermediate-scale quantum computers, and said progress toward a fault-tolerant machine must continue.

Supremacy, then the pushback

The 2019 result arrived in a paper Fortune obtained after it was briefly posted to NASA.gov, then taken down. Google declined to confirm its authenticity at the time. The company's own blog later described the test computation as taking 200 seconds against thousands of years for the best known algorithms on the most powerful supercomputers. The paper said a Google Cloud server would need 50 trillion hours for the same task, and that on the quantum processor it took 30 seconds.

Sycamore had 53 qubits, scaled back from a 72-qubit design called Bristlecone. The researchers wrote that quantum speedup is achievable in a real-world system and is not precluded by hidden physical laws, and predicted quantum computing power would grow at a double exponential rate.

IBM's Gil rejected the framing rather than the experiment. He told Fortune the test was a highly special case laboratory experiment with no practical applications, and that quantum computers would work in concert with classical machines rather than reign supreme over them. Jim Clarke, Intel Labs' director of quantum hardware, called it a notable mile marker and said a commercially viable quantum computer would require many more advances.

Since then the field has kept moving on the problem that matters more than any single benchmark: error correction. Published research has described factoring with far fewer CPU core-years than earlier estimates, and the recent headlines around IonQ, Infleqtion and a Department of Energy roadmap all point the same direction. None of that erases 2019. It just confirms what Monroe said at the start: the milestone was a beginning, not a finish line.