Amazon Exec Sees Commercial Quantum Computing in 5-7 Years

A commercially useful quantum computer will arrive within five to seven years, Amazon's (AMZN) top artificial intelligence executive said this week. He compared the technology's scaling path to the early years of Moore's Law.

The comparison is provocative but imprecise. Moore's Law described a decades-long trend in transistor density that held with surprising regularity. Quantum computing has yet to produce a single machine that outperforms a classical supercomputer on any commercially relevant task.

Several companies have claimed quantum advantage on narrow academic problems. Google's Sycamore processor solved a calculation in seconds that would take a classical machine thousands of years. The problem was designed specifically to show quantum speed, not to solve a real business problem. Classical algorithms later caught up to Sycamore's result. No quantum computer has cracked a problem a company would pay to solve faster or cheaper than existing hardware.

Amazon's own quantum effort takes a different technical path than most competitors. The AWS Quantum Computing Center uses neutral atoms trapped by lasers rather than the superconducting circuits that Google and IBM rely on. The company also sells access to multiple quantum architectures through Amazon Braket, letting customers experiment without betting on any single design. That hedge suggests Amazon sees quantum as real but is not committing the business to one approach.

The five-to-seven year timeline falls within what most quantum companies project. The industry has a history of missing its own predictions. IBM said in 2017 that a 50-qubit system would outperform any classical computer. It did not. Google declared quantum supremacy on Sycamore in 2019; classical algorithms later closed the gap. Quantum progress is real. Classical computing keeps advancing too, raising the bar for what counts as useful.

For Amazon, the quantum question is whether the technology creates an advantage for its cloud business. AWS already dominates public cloud infrastructure. If quantum stays experimental for another decade, AWS loses nothing. If it materializes sooner, the early work on Braket and neutral-atom hardware gives Amazon a foothold. The risk is not missing the technology but overpaying to be first.

The same factor that makes the timeline plausible – incremental hardware gains – also makes it uncertain. A five-to-seven year forecast in an industry that has never delivered a useful machine on schedule is not a prediction. It is a directional bet. Amazon is placing it with options, not an all-in position.