Kerrisdale Capital is short shares of D-Wave Quantum Inc (NYSE:QBTS), a $2 billion quantum annealing company whose stock has surged more than 600% since investors began chasing anything remotely associated with next-gen computing last year. While the broader market looks to gate-model quantum systems as the industry’s future, D-Wave continues to promote its fundamentally different annealing architecture. Shares currently trade at over 57x consensus 2026E revenue – a ridiculous multiple for a company that has never generated more than $9 million in annual recurring revenue, has no clear path to profitability, and sees stagnating customer growth as its approach is increasingly recognized as a commercial dead end. D-Wave is riding quantum hype, but with a core technology that cannot stay afloat.
Read hedge fund letters here
D-Wave’s business is built around quantum annealing, a niche offshoot of quantum computing developed decades ago and largely abandoned by the industry. Despite marketing its systems as optimized for complex real-world optimization problems, quantum annealing has failed to demonstrate a clear performance edge over classical solvers. Encoding problems for D-Wave’s hardware requires cumbersome and lossy reformulations that inflate problem size and introduce instability, while the system’s limited qubit connectivity forces users to spread logical variables across fragile chains of physical qubits. Strip away carefully worded press releases and as a former D-Wave engineer admitted to us, “there is no proof that any optimization problem is solved faster” using D-Wave’s quantum systems. Multiple academic studies have shown little to no scaling advantage over classical methods making it no surprise that a wide range of D-Wave customers we interviewed in key verticals like logistics, manufacturing, and pharmaceuticals reported seeing zero benefit from the technology.
To compensate for the shortcomings of its quantum technology, D-Wave leans heavily on so-called “hybrid” solutions which combine its annealer with classical hardware and algorithms to tackle industrial-scale problems. But in a glaring red flag, the company refuses to disclose to customers the relative contribution of each. Why? Because according to former insiders who developed and deployed these solutions, “hybrid” in practice means “almost entirely classical.” We believe the quantum component is minimal – often cosmetic – and its added value, beyond marketing puff pieces, was debated even internally.
Meanwhile, D-Wave’s gate-model pivot appears both reactionary and stalled. After years of dismissing gate-based systems as impractical, the company abruptly reversed course in 2021 amid a surge of investor interest in gate-model competitors like IBM and Google. Over three years later, D-Wave has released no detailed architecture papers, no fidelity data, and no performance benchmarks. D-Wave’s absence of a gate-model roadmap leaves it clinging to a fading technology while the rest of the industry passes it by.
Even D-Wave’s most recent highly publicized research lacks commercial relevance. Last month, the company declared quantum supremacy on a “useful, real-world” problem, an assertion dismissed by physicists we spoke with as misleading. The benchmark was a toy problem engineered to align with D-Wave’s hardware constraints, bearing little resemblance to real-world magnetic materials. D-Wave is not a leading quantum company – it’s a struggling provider of uncompetitive optimization solutions. As investors wake up to this reality, the stock’s rally will unwind, and its valuation will collapse under the weight of physics, finance, and fact.
Executive Summary
Quantum annealing is a commercial dead end. D-Wave’s long-running struggle to commercialize its niche offshoot of quantum computing reflects fundamental technical limitations. Serious adoption hurdles tied to physical constraints, algorithmic limitations, and persistent underperformance against classical optimization software will only continue to produce anemic financial results and erode shareholder value.
Encoding and scaling issues severely undercut D-Wave’s practical utility. Real-world optimization problems involve multiple overlapping constraints. Advanced classical solvers can manage these with relative ease, while D-Wave’s annealer requires cumbersome and lossy reformulation into an obscure mathematical format. This process increases problem size, adds auxiliary variables, and consumes scarce quantum resources. On top of this, D-Wave’s limited qubit connectivity forces logical variables to be spread across chains of physical qubits, further reducing capacity and introducing instability. Studies have consistently shown a lack of meaningful scaling advantage for D-Wave’s systems. While marketed as a revolutionary tool, D-Wave’s hardware behaves more like an overengineered accelerator with no proven edge in solving large or complex industrial problems.
D-Wave hides how hybrid solutions are almost entirely driven by classical algorithms. Given the severe limitations of its quantum systems, D-wave’s commercial strategy revolves around selling “hybrid” quantum-classical solutions for industrial optimization problems like vehicle routing and workforce scheduling. Based on multiple interviews with D-Wave customers and former employees, the company deliberately obfuscates how its hybrid solutions function in practice. Based on comments from former D-Wave insiders, the hybrid approach is driven “almost entirely” by advanced classical algorithms, with the inclusion of quantum processing representing little more than a marketing gimmick.
D-Wave’s gate-model progress appears stalled. For most of its history, D-Wave dismissed gate-model quantum computing as impractical. That conviction abruptly reversed in October 2021, when the company announced plans to build a gate-based system – widely seen as a defensive reaction to investor capital and technical momentum shifting toward gate-model leaders like IBM, Quantinuum, and Google. Over the past three years, however, D-Wave has released no performance metrics, published no peer-reviewed data, and provided no detailed product roadmap. Without a credible gate-model program, D-Wave’s already niche relevance in the quantum landscape risks further narrowing.
Quantum supremacy claim is commercially overstated. D-wave recently declared it had achieved quantum supremacy on a useful, real-world problem. According to three quantum physics experts we interviewed, however, the benchmark was in fact a “toy problem” carefully designed to match D-Wave’s strengths and lacks industrial relevance. While the result may be of interest for condensed matter physicists engaged in academic research, it falls well short of real-world quantum advantage.
Share price is divorced from fundamentals. Despite pulling back from recent highs, the stock remains up over 600% since gate-model quantum computing hype reignited last year. The excitement over quantum’s AI-like potential centers on the promise of gate-model systems – not D-Wave’s entirely different annealing architecture. Yet shares trade at a ludicrous 57x 2026E inflated consensus revenue (and 152x our estimate), even as customer growth stalls and years of cash burn loom ahead. The rally reflects misplaced quantum hype, not a business grounded in commercial reality.
Company Overview
Co-founded in 1999 by Dr. Geordie Rose (former CTO) and a team of physicists, D-Wave pioneered the commercialization of quantum computers through its specialized quantum annealing approach – a branch of quantum computing that is starkly different from the gate-model architectures pursued by all other leading competitors in the field such as Google and IBM. Unlike competitors, who are focused on building universal, programmable quantum computers capable of running diverse algorithms, D-Wave’s machines have been designed from the outset to solve a specific class of optimization problems using quantum effects to identify low energy solutions. The company went public in 2022 via a SPAC merger and is headquartered in Burnaby, British Columbia, with additional offices in Palo Alto, California.
D-Wave occupies a contentious position in quantum computing circles. One former senior executive bluntly characterized the company as the "redheaded stepchild" of the industry. This reputation stems from longstanding controversies, beginning with early skepticism that D-Wave's annealing approach constituted genuine quantum computation. The company further alienated members of the scientific community through aggressive marketing claims – including premature declarations of speedups over classical computing in carefully selected benchmarks that later failed to hold up under scrutiny. Though Rose departed in 2014, multiple industry sources confirmed to us that D-Wave continues to grapple with the legacy of his hype-heavy marketing approach.
Quantum annealing is a method of quantum computing designed to solve optimization problems – situations where the goal is to find the best solution among many possibilities. D-Wave’s quantum annealers approach this by modeling a problem as a network of binary variables and searching for the configuration with the lowest “energy,” a concept borrowed from physics. Instead of trying every option one by one like classical computers might, quantum annealing uses quantum superposition and tunneling to explore many solutions at once and “tunnel” through barriers that would trap classical methods in suboptimal answers. While powerful in theory, quantum annealing is limited in scope: it works best for a narrow class of problems and must be carefully tuned to each instance. For a more detailed overview of quantum annealing, see Appendix I.
Hybrid Solutions
D-Wave’s quantum annealing technology faces core limitations that restrict its ability to solve large-scale real-world problems (discussed further later in this report). To work around these constraints – and revive struggling commercial prospects – D-Wave introduced hybrid quantum-classical solvers in 2019. These systems combine advanced classical optimization algorithms with quantum sub-processing, enabling users to tackle problems that pure quantum annealing cannot handle effectively. When D-Wave advertises the ability to solve optimization problems with up to 2 million variables and constraints, it is exclusively through these hybrid solvers. These solutions now form the backbone of D-Wave’s go-to-market strategy of helping customers “confront real-world problems of growing complexity.”
Based on multiple conversations with D-Wave customers and former employees, the company does not disclose precisely how much of its hybrid solution is powered by classical computing versus quantum hardware – making the system, by design, a literal black box. This lack of transparency is not coincidental. In our view, it reflects an uncomfortable truth: while customers are charged a premium for the supposed advantage of quantum computing in solving complex problems, D-Wave’s hybrid solvers rely overwhelmingly on advanced classical techniques, such as parallel tempering. The quantum component, according to multiple sources, contributes no differentiated value. One former D-Wave engineer we spoke with put it bluntly: apart from marketing, the quantum component’s added benefit is “highly questionable” and a source of internal skepticism and debate.
(No) Advantage Systems
D-Wave released its first commercial quantum annealer in 2011 with 128 qubits. The company's current Advantage system now features over 5,000 qubits and supports hybrid quantum-classical solvers. Prototypes of its second generation Advantage2 system with around 1,200 qubits are currently available with commercial rollout of a full-production 4,400 qubit version targeted before year-end. Advantage systems are available via the company’s Leap quantum cloud service and access can be purchased directly from D-Wave or through Amazon Web Services Marketplace. D-Wave removed access to its devices from Amazon Bracket in 2022, following a falling out between the two companies. According to a knowledgeable source, this stemmed directly from D-Wave’s frustration with certain Amazon technical staff who believed D-Wave’s systems provided no practical advantage over classical methods in solving customer problems.
Though impressive sounding, D-Wave’s oft-cited qubit counts can be misleading when comparing D-Wave's technology to gate-model quantum processors like those from IBM or Google. The two are fundamentally different architectures designed for distinct purposes. Gate-model quantum computers employ flexible, programmable qubits capable of universal quantum computation, where each additional qubit exponentially expands the computational space.
In contrast, D-Wave's annealing qubits serve as fixed components in a physical optimization system, functioning more like specialized sensors than general-purpose computing elements. In D-Wave’s architecture, qubits are not fully connected, meaning logical problems must be “embedded” into the hardware using extra qubits as bridges. According to experts consulted, this embedding process ties up large numbers of qubits – sometimes >90% of them – just to represent a problem, not solve it. When D-Wave promotes a 5,000-qubit+ Advantage system, only a fraction of those qubits typically contributes to meaningful computation after accounting for the substantial overhead.
In addition to classical optimization use cases, D-Wave promotes its annealing technology as having potential in quantum simulation, particularly for materials science applications. In March 2025, the company claimed to be the first in the world to demonstrate quantum supremacy on a “useful, real-world problem.” However, as we detail later in the report, expert perspectives from the scientific community raise serious doubts about the true significance of this claim.
Lack of Gate-Model Progress
For most of D-Wave’s corporate history, its leadership dismissed gate-model quantum computing as impractical in the near term, citing crippling challenges like qubit coherence, fidelity, and error correction. Founder Geordie Rose once derided gate-based systems as a “rotten idea,” insisting quantum annealing was the only viable path to commercial applications. Yet in October 2021, D-Wave pivoted, announcing plans to develop its own gate-model quantum computer. The reversal came amid mounting pressure: customers demanded algorithms beyond annealing, rivals like IBM and Google had made great technical strides in materials engineering and error-correction, and investors had poured billions into gate-based approaches.
Three-and-half years after announcing the program, the company has published no gate fidelity measurements and offered no updates on error correction strategies – all while continuing to heavily promote its annealing systems. An industry source with direct connections to D-Wave told us he believed D-Wave’s system had been “de-prioritized.” While competitors like IBM and Quantinuum routinely demonstrate progress through peer-reviewed publications, detailed hardware roadmaps, and verifiable performance metrics, D-Wave maintains a conspicuous silence on substantive details. The absence of these standard technical disclosures suggests D-Wave's gate-model efforts remain in early exploration phases at best.
As the broader industry coalesces around gate-model architectures as the future of general-purpose quantum computing, D-Wave remains tethered to a modality with limited scalability, constrained applicability, and declining investor interest.
Read the full report here by Kerrisdale Capital
