D-Wave Quantum touts tech breakthrough that lets gate models scale

D-Wave Quantum has announced a breakthrough that addresses a key challenge in developing superconducting gate-based quantum computers: how to gather a ton of quantum bits (or qubits) in the same place while keeping them all cool enough to function.

A particularly tricky problem of heat

We want quantum computers to be able to solve complex problems. Complex problems require these machines to utilize a lot of qubits. Those qubits, in a superconducting system, need to be housed in an extremely cold environment to operate.

But connecting and communicating with all those quantum processing units (or QPUs) via individual wires would result in too much heat, not to mention adding to the cost of the system.

D-Wave says it’s solved this problem through multiplexing (using one wire to communicate with a number of other chips) and bump bonding (stacking a QPU and a control chip together), as well as controlling qubits by magnetic fields.

“This industry-first milestone advances the development of commercially viable gate-model quantum computers by significantly reducing the wiring required to control large numbers of qubits without degrading qubit fidelity,” per the press release. “Using superconducting bump bonding and advanced cryogenic packaging techniques, D-Wave built a multichip package that integrates a high-coherence fluxonium qubit chip with a multilayer control chip.”

Annealing vs. gate-based

D-Wave is the major player in annealing quantum computing, an approach that solves more specialized optimization problems. The company has already been able to apply this on-chip cryogenic control technology to its annealing systems.

But gate-based quantum computers, which aim to address even more complex and broad queries, are the dominant approach, pursued by the likes of Rigetti Computing and IonQ as well as D-Wave.

 “We wanted to make sure that we had kind of the scalable control piece sort of nailed down, because we think that to get to broad quantum utility with gate-model architectures requires scaled, error-corrected architectures, which requires a lot of physical qubits,” said Dr. Trevor Lanting, chief development officer at D-Wave.

“This is basically proof that we can use the technology that exists that we’ve developed, and more or less in a very straightforward way, to control gate-model architectures.”

He added that D-Wave’s superconducting approach to quantum computing allows the firm to leverage preexisting manufacturing and packaging processes that have been developed, rather than having to build up a technology base from scratch. 

A solid first step

During the conference call that followed the release of Q3 earnings in November, CEO Dr. Alan Baratz highlighted gate-model development as a priority for D-Wave.

“ Up until now, our investment in gate has been light, mostly because we haven’t had the funds to be able to grow that investment all that much. Now with the roughly $830 million in the bank, we have the resources to be able to invest more in that program, both internal investment and through acquisition to accelerate the program,” he told Sherwood News.

“ We have one customer who has said, ‘When you have a gate-model system, I want it.’”

At the time, Baratz told us that what was ultimately announced today would mark the “first step” in the company’s gate-model program.

“ From there, we will go to a small logical qubit, a small surface code logical qubit to demonstrate that we can now use this technology to build error-correctable logical cubits,” he said. “And our hope would be to have that before the end of next year, and then we’ll start scaling to larger surface code.”