Why Google’s Top Quantum Scientist Walked Away: An Inside Interview

In April, John Martinis, the leading quantum‑hardware scientist at Google who helped achieve the landmark quantum‑supremacy result published in Nature , announced his resignation, surprising the community.

Martinis, formerly a professor at UCSB, joined Google in 2014 with a team of about a dozen researchers to build a functional superconducting quantum computer. He and his team published roughly 200 papers during their time at Google.

PSG : Why did you leave Google? What prompted the decision?

Martinis : I was concerned about losing my leadership role. Hartmut Neven, who ran the overall quantum‑AI project, hired me to lead the hardware team, but over time my authority was reduced, and I felt my optimistic, single‑goal focus conflicted with the broader management.

PSG : Did the team’s focus on quantum supremacy cause tension?

Martinis : Yes. I pushed the team to concentrate on the supremacy experiment, which created friction, especially when others wanted to pursue multiple projects simultaneously.

PSG : How did the leadership structure change?

Martinis : After a year, three other members were made co‑leaders while I became a consultant. I still contributed technically, inventing methods to scale qubit numbers and solve wiring challenges, but I no longer directed the hardware effort.

PSG : What are your thoughts on the future of quantum computing at Google?

Martinis : Google plans to build a million‑qubit system within a decade with error rates low enough for quantum error correction. Even a few hundred high‑fidelity qubits could enable useful applications. The team will continue producing strong results based on the Sycamore chip.

PSG : Would you consider returning to Google?

Martinis : Possibly, if a mutually satisfactory arrangement can be found. I still respect the team and wish them success.

The interview provides a rare, candid look at the scientific, technical, and managerial challenges behind one of the most celebrated achievements in quantum computing.