Amazon Sets a 5-to-7-Year Clock for a Useful Quantum Computer

Amazon’s top AI executive, Peter DeSantis, told CNBC on Wednesday that the first commercially useful quantum computers will arrive within five to seven years, framing the next stretch of progress as a Moore’s Law-style climb. The forecast is the first concrete timeline Amazon has offered on useful quantum machines, and it lands in the middle of a sharply divided industry. Google said in March 2025 the technology was roughly five years from running practical applications classical computers cannot calculate. Microsoft has set 2029 as its target for a commercially viable machine, while Nvidia CEO Jensen Huang told investors earlier in 2025 that fifteen years would “probably be on the early side” and later walked the comment back.

DeSantis, a few months into leading a new Amazon organization that bundles AI models, custom chips, and quantum computing, made the comments on June 17, 2026. The new role folds quantum into a single Amazon unit with custom chips and AI models, the first time the company has put all three under one executive.

Amazon’s First Quantum Timeline Points to 5 to 7 Years

Amazon has stayed largely quiet on when its quantum work would translate into a commercial product. DeSantis, who leads a new Amazon organization focused on AI models, custom chips, and quantum computing, ended that silence in his full forecast on useful quantum computing. “I actually do believe, over the next five-to-seven years, we’re going to start to see the first commercially useful small-scale quantum computers,” he told the network.

That makes Amazon the latest big-tech firm to put a public clock on the technology. Once the first system lands, growth should look familiar, DeSantis said: “From there, we’re going to see something that looks a lot like Moore’s Law, where they’re going to get bigger and bigger every year, and they’re going to be able to tackle more and more interesting problems.” The forecast does not specify which qubit technology Amazon expects to win out inside that window, and it does not name a specific year for the first system. Amazon has spread its quantum research between in-house superconducting hardware and cloud platform partnerships that include neutral-atom technologies, so the commitment to a single qubit approach is left open. What the forecast does commit to is the Moore’s Law-style climb and a narrow class of chemistry and materials problems where the first systems are expected to land.

By the numbers, the published timelines across the industry line up in different ways:

• Google: In March 2025, a Google quantum executive told CNBC the technology was about five years from running practical applications classical computers cannot calculate.
• Microsoft: Microsoft has said it will have a commercially viable quantum machine by 2029.
• Nvidia: Jensen Huang told investors in early 2025 that fifteen years would “probably be on the early side” for useful quantum computers, then walked the comment back.
• Amazon: First concrete timeline on useful quantum machines, with Moore’s Law-style scaling after the first useful system.

Quantum Is a Tool for Specific Problems, DeSantis Says

DeSantis used the interview to push back on one of the most common framings of quantum technology. “One of the misnomers is a quantum computer is going to be a faster computer, that’s not it at all,” he said. A quantum machine, in his telling, is built to solve a narrow class of problems that classical computers handle poorly, not to speed up the average workload. “A quantum computer is going to solve a very particular type of problem that isn’t solved well today with a classic computer, and it’s going to solve it much better,” he said.

Molecular and material simulations are likely the first targets of that approach. “The problems that I would think are going to be tackled first are the ones that are quantum-based problems, so things like chemistry, material science,” DeSantis said. “These are the problems where today we cannot run high enough fidelity simulations in a classic computer, and once we have a quantum computer, we’re going to find some real progress,” he added.

The framing narrows the commercial pitch. Quantum machines are tools for problems such as chemistry and material science simulations, not a faster computer, and they are aimed at a smaller, more specialized market than the laptops and data centers most people use today.

How Amazon’s Ocelot Chip Approaches Error Correction

Amazon’s hardware push behind that forecast sits on a single chip called Ocelot, announced by Amazon Web Services in February 2025 and detailed in the journal Nature. The chip was developed at the AWS Center for Quantum Computing at the California Institute of Technology and is built around “cat qubits,” a superconducting architecture named after Schrödinger’s famous thought experiment. Where conventional qubits lose their state at the slightest environmental interference, cat qubits are designed to suppress one of the two main error types at the hardware level.

Ocelot’s design is a deliberate departure from the dominant approach to quantum error correction. AWS researchers built the chip from the ground up with error correction as the top requirement, rather than bolting it onto an existing architecture, the company said in its February 2025 announcement of the chip. Cat qubits, the architecture’s central innovation, suppress bit-flip errors at the physical qubit level, leaving only phase-flip errors for a classical repetition code to handle. That two-step approach is what allows Ocelot to use far fewer physical qubits per logical qubit than conventional surface-code designs, according to the team’s technical write-up of the architecture.

Amazon says the numbers from the announcement are large. Ocelot’s architecture could reduce the resources needed for quantum error correction by up to 90% compared to conventional approaches. A commercial machine built on the architecture could cost as little as one-fifth of current approaches, the company has said. The same team has also said the architecture could accelerate the timeline to a practical quantum computer by up to five years. The team is the first to publish a peer-reviewed result on this kind of cat-qubit chip in Nature, and the work is detailed in a Caltech write-up of the research.

With the recent advancements in quantum research, it is no longer a matter of if, but when practical, fault-tolerant quantum computers will be available for real-world applications. Ocelot is an important step on that journey.

Oskar Painter, AWS’s director of quantum hardware and head of the AWS Center for Quantum Computing at Caltech, made the comments in the February 2025 announcement of the Ocelot chip. The Nature paper itself describes Ocelot as a “first realization” of a scalable architecture, and the team has framed future work as exponentially driving down logical error rates through larger code distances and improved components.

Big Tech’s Quantum Race in Numbers

Amazon is not the only major player betting on a quantum future. Google’s most advanced quantum processor, Willow, is a 105-qubit superconducting chip with significantly improved coherence times compared to its earlier Sycamore chip, the company has said. Microsoft has staked its quantum effort on topological qubits and is targeting a commercially viable machine by 2029, while IBM’s modular Quantum System Two platform hosts the world’s most powerful fleet of utility-scale quantum computers, the company has said.

Superconducting, topological, ion-trap, and neutral-atom approaches are all still in play, with no clear consensus on which will reach commercial scale first. Amazon, for its part, has spread its research dollars between in-house superconducting hardware at Caltech and cloud platform partnerships that include neutral-atom technologies. The result is a race where timelines vary as much as the physics underneath them. Each forecast is built on its own qubit technology, its own definition of “useful,” and its own funding pipeline.

The race is already showing up in the technical numbers, and the published targets and the Ocelot chip’s measured performance are the clearest read on where each side thinks the technology stands:

• Cat qubit bit-flip protection: more than a thousand times longer than the lifetime of conventional superconducting qubits
• Cat qubit bit-flip time: approaching one second
• Cat qubit phase-flip time: 20 microseconds
• Ocelot logical error rate, distance-3 code: 1.72% per cycle
• Ocelot logical error rate, distance-5 code: 1.65% per cycle

What Could Move the Five-to-Seven-Year Window

DeSantis’s forecast carries a number of embedded assumptions. The Moore’s Law-style climb only kicks in after the first useful small-scale system lands, and the Ocelot team’s claim of a timeline cut of up to five years rests on a chip that is still a prototype.

The other wild card is the rest of the field. Google’s earlier five-year forecast, Microsoft’s 2029 target, and Nvidia’s later (and walked-back) fifteen-year comment show the industry is still far from consensus on the timeline. Each prediction comes with its own definition of “useful,” its own qubit technology, and its own funding pipeline. Amazon’s new quantum timeline is the most recent executive-level prediction on the record, putting it between Microsoft’s near-term push and Google’s more aggressive five-year horizon. The same scaling curve is what eventually puts public-key cryptography in the line of fire, and the digital assets built on today’s encryption are already on a separate migration clock for digital assets.

Frequently Asked Questions

When will Amazon’s first commercially useful quantum computer launch?

Peter DeSantis told CNBC on June 17, 2026, that the first useful small-scale quantum machines will arrive in five to seven years, with growth after the first system following a Moore’s Law-style curve. DeSantis took on the new role leading Amazon’s bundled AI, chip, and quantum effort only a few months earlier, and the forecast is the first concrete quantum timeline Amazon has put on the record.

What will the first useful quantum computers be used for?

DeSantis said chemistry and materials science are the most likely early applications, calling them problems where ‘today we cannot run high enough fidelity simulations in a classic computer.’ The Ocelot team has separately named drug discovery, new materials, and financial risk modeling as areas the technology could eventually reach.

How does Amazon’s Ocelot chip differ from a conventional quantum chip?

Ocelot uses ‘cat qubits,’ a superconducting architecture designed to suppress bit-flip errors at the hardware level. The chip has 14 core components, including five cat data qubits and four ancillary qubits, on two stacked silicon microchips. The team’s published Nature paper reports cat qubit bit-flip times approaching one second, more than a thousand times longer than the lifetime of conventional superconducting qubits.

How does Amazon’s 5-to-7-year forecast compare to Google, Microsoft, and IBM?

Google said in March 2025 the technology was about five years from running practical applications that classical computers cannot calculate. Microsoft is targeting a commercially viable machine by 2029. Nvidia CEO Jensen Huang said fifteen years would ‘probably be on the early side’ in early 2025 and later walked the comment back, leaving Amazon’s five-to-seven-year forecast as the most recent executive-level prediction on the record.

Will quantum computers replace classical computers?

DeSantis specifically said quantum computers are not ‘a faster computer.’ Quantum machines are designed to solve a narrow class of problems classical computers handle poorly, and they will work alongside classical hardware rather than replace it. The first commercial systems, both Amazon and its peers have said, will be small and specialized, not general-purpose.