Amazon May Boost Quantum Power for AWS
Though quantum computing is far from an everyday reality, the possibilities it offers are powerful. Amazon may want in on the ground floor.
Sign up for smart news, insights, and analysis on the biggest financial stories of the day.
Amazon could be looking at a new way to keep AWS on top.
The company filed a patent application for “quantum computing task translation,” which is meant to support multiple quantum computing technologies. Amazon’s patent details what it calls a “quantum algorithm development kit,” hinting at the possibility of beefing up quantum capabilities to AWS customers.
Amazon’s tech essentially acts as a middleman between a quantum computer and the user interface. First, a user makes a request with this service through an “edge computing device” — their own device that isn’t connected to the quantum computer itself. Then the system will “automatically translate the quantum task, quantum algorithm, or quantum circuit” into a representation that a quantum computer can understand. This system will then pick the right quantum computer for a certain job, and work with it on the back end to complete the request.
Amazon noted that different types of quantum computers are better at computing and generating different things. Plus, things like “costs, run-times, error rates, availability, etc. may vary across quantum computing technologies.” Therefore, picking the technology one can be important to getting the desired outcome.
The company notes that its service aims to bring the power of quantum technology to the layperson, “enabling customers to seamlessly use the different quantum computing technologies without requiring the customers to have specific knowledge of the underlying quantum computing technologies.”
For decades, quantum computing — applying the properties of quantum mechanics to perform complex calculations — existed only in theory and experiments. But in the last decade, a number of tech companies have invested heavily in bringing those theories into reality, said Todd Brun, professor at the USC Viterbi School of Engineering.
While Amazon hasn’t publicly released news of quantum hardware, it offers Amazon Braket, a quantum computing service, as part of AWS, which is mostly used for scientific research. This patent seems to be a more robust version of its current quantum services, enabling non-experts to take advantage of the emerging tech.
“What (Amazon) is trying to do, I think, is build a software environment that will allow you to develop applications that could be run on many different quantum computers from different providers,” said Brun.
Brun noted that a small number of quantum processors exist, with two of the major ones being IBM’s and Google’s, which are “building blocks towards larger systems that may exist in the future.” But even the ones that have been built only have around a hundred quantum bits of processing power, which, compared to the billions of bytes of processing that a classical computer offers, makes them “very limited machines.”
Several things stand in the way of making quantum computing an everyday reality. For one, quantum computers need to be kept at very low temperatures, roughly “half a degree above absolute zero,” said Brun. Cooling itself can be a big drain on resources and costs, and as the machines themselves grow, so do the cooling needs. Plus, the qubits in a quantum machine are deeply impacted by the “noise” of the atoms around them, he said, which can only be accounted for by making the machines bigger.
But once science has pushed past those barriers, quantum computing comes with big promises that go far beyond research uses. Quantum simulation can potentially benefit industries like chemicals and pharmaceuticals, and interest from banks like JPMorgan Chase and HSBC show that the financial sector has a use for quantum optimization, too. Plus, quantum computing could accelerate the potential of AI to quickly make sense of lots of data, according to Deloitte. Amazon may want to be an early adopter.
“The idea is using it to solve optimization problems,” said Brun. “If that pays off — and that we don’t know — then we would actually have very wide applications for all kinds of businesses.”