Analyst reaction to D-Wave Quantum’s announcement today of an expanded product development roadmap aimed at helping organizations address a range of artificial intelligence/machine learning (AI/ML) workloads was decidedly positive, but there was a consensus that much needs to happen before quantum computing can be widely adopted.
The Palo Alto, California-based company said it is “strengthening the connection between quantum optimization, AI and machine learning” with the enhancements to its Leap quantum cloud service, adding that the move comes “at a time when the entire AI industry is facing a computing crisis.”
The amount of computing required to satisfy an ever-growing number of use cases, and the associated energy costs, are rising rapidly, according to the company. D-Wave’s new product aims to leverage annealing quantum computing’s “unique ability to solve optimization problems to help customers discover better, faster, and more energy-efficient AI and ML workloads,” it added in the release.
Roadmap Focus
According to the release, the new roadmap will focus on three areas:
Quantum distribution for generative AI: The company said its development in this area is focused on “designing new, modern generative AI architectures that use quantum processing units (QPUs) samples from a quantum distribution that cannot be generated traditionally.” Initial focus on use cases including molecular discovery. Restricted Boltzmann Machine (RBM) architectures leveraging D-Wave’s QPUs are being used for applications “that could lead to reduced energy consumption in training and running AI models, from cybersecurity and drug discovery to high-energy physics data analysis.” GPU (Graphics Processing Unit) Integration with Leap: D-Wave said it will incorporate additional GPU resources for training and supporting AI models in parallel with optimization workloads. In addition, it said, “Efforts are underway to further reduce the latency between QPUs and classical computing resources, which is a key step towards enabling hybrid quantum technologies for AI/ML.”
Potential impact
“D-Wave’s advances in quantum computing for AI are exciting, but it’s still very early days to assess the impact and value of quantum computing for real-world AI use cases,” said Bill Wong, research scientist at Info-Tech Research Group.
Currently, he says, “Most companies are not prepared for, nor do they expect, AI breakthroughs that can be achieved through the use of quantum computing. This will remain one of the key challenges for quantum computing: finding AI use cases that can greatly benefit from this accelerated computing platform, when traditional (i.e., GPU-based) computing requirements can be handled at a much lower cost.”
“Potential use cases could focus on developing quantum-resistant cryptography, where traditional computing platforms cannot accommodate the required resources. D-Wave is doing cutting-edge research, and I am also looking for use cases that can drive adoption of this unique platform,” Wong said.
“The idea has always been that AI and quantum would work synergistically, but at this point it’s more about AI influencing quantum,” said Heather West, a quantum computing analyst at IDC.
She said a key part of the announcement was about annealing quantum computing, which is specifically designed to solve optimization problems, because quantum “needs a larger customer base to really make an impact.”
Asked whether D-Wave’s announcement today would bring the quantum discussion into the mainstream, she said: “I think that’s fair. D-Wave has taken a customer-centric approach to developing quantum systems.”
She said that many quantum hardware vendors “are going to talk about their qubits, they’re going to talk about the different components of their quantum systems, they’re going to talk about potential use cases. But D-Wave is going to talk about use cases that are being explored and are becoming increasingly valuable. They’re really driving a customer-centric focus on quantum, and that’s what differentiates them.”
Sid Nag, a vice president analyst at Gartner who focuses on scalable computing, said the announcement represents “an alternative to GPUs, even though D-Wave hasn’t explicitly said so.”
But he cautioned that “a lot of specialized cloud providers are popping up and competing with the hyperscalers, and this is a product of the trend that AI is getting bigger and bigger.” It’s unclear how big it will be in terms of actual growth. There will be a trough of disillusionment at some point. [a Gartner term referring to a phase ‘where, after initial hype and inflated expectations, interest begins to wane.’] will begin.
In D-Wave’s case, “they’re going after a very specific market,” Nag added.
