AI continues to get a lot more attention from the hyperscaler community than quantum computing, and the reason is clear: AI is the here and now, while quantum computing lies somewhere over the rainbow.
So, while Amazon Web Services (AWS) got a lot of attention last week for launching its own chip focused on accelerating AI training applications, it got less notice for an announcement that could have a huge impact on the future of quantum computing
In an evening presentation at last week’s AWS re:Invent conference, AWS EC2 General Manager Peter Desantis unveiled what he described as a prototype, custom developed “quantum error correction chip.” The announcement was the culmination of years of work since Amazon opened its AWS Center for Quantum Computing (CQC) at the California Institute of Technology (Caltech back in 2019.
The greatest commercial relevance for the quantum computing industry could still be some years away, and it is not yet clear how involved hyperscalers like AWS, Google, and Microsoft ultimately will be in the sector. AWS and Microsoft both offer quantum computing-as-a-service capabilities through their cloud services, offerings that allow researchers and others to access quantum processing units made by quantum computing start-ups like IonQ. All three cloud giants also have dabbled in quantum computing R&D in a number of ways, with AWS also hosting a Center for Quantum Networking and a Quantum Solutions Lab.
But Desantis noted that one of the barriers keeping quantum computing from achieving greater commercial relevance is its “error-prone” nature. Desantis said, “It's a reasonable question to ask: Why haven't quantum computers started to change the world yet? And like many things, you have to read the fine print. The fine print with quantum computers says they're noisy and prone to error. All the computers we use today use error correction… Quantum objects are far more sensitive to noise from their environments.”
Many companies are working on quantum error correction, which generally speaking involves encoding blocks of physical qubits into higher-quality logical qubits. However, it takes many physical qubits to make one logical qubit.
“Today, the state of the art is probably about one error per 1000 quantum operations,” Desantis said, further describing that stat as a 100x improvement from five years ago. “The problem is, qubits are still far too noisy to be useful. The quantum algorithms that we get excited about require billions of operations without an error.”
AWS thus focused on better ways to improve error correction and more efficiently resolve the bit flip and phase flip errors that bog down quantum computing performance. It’s answer is the new prototype chip, manufactured in-house by AWS
“The unique thing about this chip is how it approaches error correction by separating the bit flips errors from the phase flips. With this prototype device, we've been able to suppress bit flip errors by 100x by using a passive error correction approach. This allows us to focus our active error correction efforts on just those phase flips.”
Through this approach, Desantis said AWS has demonstrated that it can “theoretically achieve quantum error correction six times more efficiently than with standard error correction approaches.” That achievement could help accelerate the whole science of quantum computing, and bring the industry much closer much more quickly to commercially-relevant quantum computers that can solve problems and carry out computations that classical computers are not able to address.
Desantis called it an important step toward “developing the hardware-efficient and scalable quantum error correction that we will need to solve interesting problems on a quantum computer.”
This development does not necessarily mean that AWS will start manufacturing its own quantum error correction chips. Chips of this type most likely would cwork with other kinds of quantum processing units within a full-scal quantum computer. Desantis demurred from discussing what will come next in AWS’ pursuit of higher-quality quantum computing, but it could signal that AWS is getting more serious and more deeply involved in the technology discipline that could have the massive impact on the computing sector tomorrow that AI is having today.