D-Wave is something of an anomaly in a quantum computing industry. While companies ranging from hyperscalers like Microsoft, Google, and Amazon Web Services to traditional enterprise system makers like IBM, Dell Technologies, Cisco Systems, and Hewlett Packard Enterprise are trying to get some revenue, and pure-play vendors, software and algorithm specialists, and cybersecurity firms are establishing themselves and looking toward the near future. But D-Wave has a growing business already in place that is bringing in millions of dollars.

Company executives years ago embraced annealing quantum technology as the shortest path to get to market, and initially made its Advantage and then Advantage2 quantum system accessible through its Leap quantum cloud service. Early last year, the vendor began selling systems to organizations, like Jülich Supercomputing Centre at Forschungszentrum Jülich (FZJ) in Germany, to deploy on-site. More recently, Florida Atlantic University (FAU) agreed to buy a 5,000-qubit Advantage system for $20 million.

It seems to be working. In the first quarter, this year, D-Wave showed closed bookings of $33.4 million, a 1,994 percent year-over-year jump and a 149 percent increase over the previous quarter. Included in all of that was not only the deal with FAU but a two-year, $10 million quantum computing-as-a-service deal with a Fortune 100 company.

Banging The Gate-Model Drum

More recently, D-Wave has been making more noise around its dual-platform strategy, in particular its plans for gate-model quantum systems leveraging the superconducting modality that others like IBM, Google, and IonQ favor. The company made a significant move in that direction when it closed its $550 million deal for ten-year-old Quantum Circuits and its dual-rail error correction technology.

As we wrote earlier this year, the dual-rail qubit architecture embeds error detection directly into the qubits, which D-Wave executives say not only improves the quality of the qubits themselves but means that a logical qubit could be created with fewer physical qubits, which should accelerate the development of fault-tolerant quantum systems. While other vendors are focusing on the number of qubits, D-Wave is aiming to reduce errors at the hardware level by embedding error correction onto the qubits themselves to errors can be found during computation at the single-qubit level.

The Dual-Rail Way

This way, the dual-rail qubits can identity about 90 percent of errors has they pop up, so fewer physical qubits are needed. D-Wave executives also boast of reaching – with error detection – 99.9 percent two-qubit fidelities, so physical errors occur about once in every 1,000 operations. The company is leaning on the Lambda metric for measuring how quickly a computer's errors decrease as more error-correction capabilities are added. According to D-Wave, the norm for most quantum companies is a value of 2, which means each error correction cuts the number of errors by half. D-Wave is looking at a Lambda value of 10, which executives say will reduce errors more quickly by a factor of 10X for each increment, which will lead to fault-tolerant quantum computing that needs significantly fewer physical qubits.

The dual-rail technology dovetails well with D-Wave’s superconducting expertise and its desire to manufacture the systems at scale, according to D-Wave chief development officer Trevor Lanting.

“We saw an immense amount of synergy between this core qubit technology, which really we see as game-changing because it really shrinks the physical requirements to get the full tolerance, but it doesn't take us in a 90-degree direction,” Lanting tells The Next Platform. “This is parallel to the direction we're going with both our annealing and our demo architectures, which is to use superconducting devices to build out these quantum computing systems.”

An Expanded Timeline

With the merger complete, D-Wave is rolling out an expanded timeline for its gate-model systems. When the deal was announced in January, D-Wave laid out a partial timeline that included the general availability of a 17-qubit, dual-rail transmon-based system this year, a 49-qubit system next year, and 181-quibit quantum computer in 2028. The system will deliver 2,000-fold error reduction factor over the physical error rate and will be the blueprint for scalable fault-tolerant architectures. The company also will complete the design for a 1,000-qubit that year.

Earlier this month, D-Wave announced that in 2030 it will complete a 10-logical-quibit system that can support the first fault-tolerant algorithms, and two years later will launch a 100-logical-qubit system that will be able to perform more than 1 million operations supporting initial chemistry and quantum AI applications.

Lanting says D-Wave is getting customer interest in the promise of post-NISQ (Noisy Intermediate-Scale Quantum) systems “the ability to do error correction and then build your software stack on top of that. On the near-term path, getting to that 100 logical-qubit million gate operation point, we're really seeing that as the phase where the gate model systems become commercial across the board. A lot of our customers don't want to wait to 2030 or 2032, because at that point, they will not have developed the expertise on the algorithm side, the expertise in really figuring out how they actually integrate these systems into their industrial workflows.”

Shifting Goalposts

The timeline for these systems becoming performant is speeding up, and organizations are realizing they need to have teams and plans in place for addressing the technology, he says.

“The goalposts are shifting over time in terms of what really is necessary for broad quantum utility, but the best evidence today based on what we know about algorithms and applications is that once you hit that 100 logical-qubit mark and you're able to support that scale of operations, like over a million successful operations, that's really unlocking initial applications, primarily in material simulation and quantum chemistry, that's the point at which we really see this technology transitioning to fully supporting those early commercial workloads,” he says.

D-Wave’s goal is to be a sort of one-stop quantum shop, with both annealing and gate-model quantum systems available. Lanting says there are synergies between the annealing and superconducting efforts, no plans now to directly integrate them.

“Our annealing systems are supporting optimization workloads,” he says. “We're seeing a lot of early signs that they could help with quantum AI applications. There's increasing evidence that gate-model systems just won't be able to attack optimization or a lot of machine learning tasks. ... The annealing systems are not going to be able to address the really juicy quantum chemistry problems and the novel materials discovery problems. That's the sweet spot for gate systems. Depending on the customer and the industry and the use case, they may engage with one or the other, or both.”

Federal Funding For The Effort

As D-Wave moves forward with its plans, it’s likely going to do so with another $100 million on hand. The company was one of several quantum vendors to sign a Letter of Intent (LOI) with the U.S. Commerce Department, which last month announced a $2 billion program being made available to them through the U.S. CHIPS and Science Act in exchange for the federal government taking an equity stake in the company.

D-Wave would use the money to help fund the development of a 100,000-qubit annealing system and 10,000-qubit gate-model computer. Other vendors, including IBM, Quantinuum, Rigetti, and PsiQuantum, sign similar LOIs. However, the money isn’t guaranteed. There is debate whether the investments adhere to the intent of the CHIPS Act, which was passed during the Biden Administration, and whether the U.S. government should be taking stakes in private companies.

Still, it would be welcomed, Lanting says.

“For us there's a couple of things,” he says. “One is this is R&D that is core to our roadmap, core to both our annealing and our gate systems. And this allows us to pull in some timelines for some of the development work that we're planning to do. The second thing, and maybe more importantly, is this is the first time we've got a direct endorsement from the U.S. government as D-Wave for technology development. That endorsement we see as incredibly important.”