D-Wave, Gate-Model Quantum Computing, and its Lessons in Adapting a Patent Portfolio

D-Wave Systems, one of the biggest and oldest producers of quantum computing hardware, announced in October 2021 that it would be developing a gate-model quantum computer that it plans to launch in 2023 or 2024 in addition to continuing to develop its quantum annealer. In order to understand why this is a big deal, one must understand what it’s been doing differently so far. There are plenty of articles comparing the company’s former model to the device it proposes to build, but it boils down to this: Its former quantum annealer was limited to solving optimization problems, but the new quantum computer will be more similar to traditional computers, be able to tackle a wide range of problems, and be more similar to what other major companies, like IBM and Google, are developing. However, from a larger perspective, D-Wave’s pivot is interesting in what it teaches us, and what it might later teach us, about the adaptability of large patent portfolios. 


Like many of the biggest quantum computing companies, D-Wave has invested heavily in patents, with 282 patent assets over the past 10 years. However, whereas companies like Google, Microsoft, and IBM have accelerated their filings heavily over the past 5 years, D-Wave has actually slowed down. Over the past five years, D-Wave has 166 patent assets, 17 more than Microsoft and 19 more than Google. If you look over only the past three years, the trend is even more pronounced, with D-Wave only having 85 patent assets, 20 less than Microsoft and nearly 50 less than Google. However, the questions facing D-Wave as it transitions to a new technology are not only how many total patents it has, but whether those patents cover the new technology.

While it cannot be said for certain that any of D-Wave’s previous patents will cover their planned gate-model quantum computer, there are a few reasonable inferences that can be made. First, 34 of its patent assets deal with annealing techniques, which are unlikely to be very applicable. Second, it has 78 patent assets in qubit manipulation. Qubit manipulation is something needed in both quantum annealing and gate-model quantum computing, but the nature of many of D-Wave’s patents in this category is such that they are unlikely to be cross-applicable. Therefore, while a few of the patents may support the new system, it is probably safe to eliminate those 78 as a general rule. On the other hand, it has 20 patent assets dealing with qubit creation and storage, which can likely be assumed to be applicable, assuming that D-Wave still plans on using similar superconducting technology. D-Wave even has a few patents specifically dealing with the creation of gates, but these patents appear to be very old, likely from before D-Wave even knew it was going to focus on quantum annealing. D-Wave has patent assets in plenty of other categories, but even without considering whether those are relevant, the elimination annealing and qubit manipulation documents brings it down by about 100 assets over the past ten years. 

While D-Wave may have good coverage for its quantum annealer so far, it is unquestionably behind in the pursuit of coverage of a gate-model system compared to its major competitors, especially within the superconducting qubit space. Those watching this space should either look for D-Wave’s patent filing trend to reverse course and build quickly over the next 2 years, or should look out for an announcement that D-Wave is abandoning its gate-model ambitions. For more on D-Wave’s patent developments, follow the Quantum Computing Patent Forecast!