Coherence Report is the only company providing an accurate deep-dive on D-Wave’s Q2, business model… and are the only group that has accurately forecasted their revenue since 2025.
D-Wave has a steady supply of news. Partnerships, press releases, technical claims, and marquee customer or two. What it does not have are revenue, consistent usage, projects in production, or meaningful and ongoing engagements. Understanding why is absolutely essential to anyone holding this stock.
Subscribe to our free newsletter for weekly updates on our work, or follow us on X.com at @coherence_repor.
The AT&T Headline
In late July, investor interest surged with news that AT&T would expand its use of D-Wave’s quantum annealing technology “across network operations”. There were claims about success, speedup, etc. but investors need to understand what the project actually was.
There were two projects:
- PoC #1: Technician routing and dispatch. The problem was routing thousands of technicians and dispatching them to assigned jobs across AT&T’s network. The result was the headline one: the workload dropped from roughly an hour to under 15 seconds, though neither company disclosed the classical baseline it was measured against.
- PoC #2: Outage optimization. A more exploratory effort focused on how AT&T responds to and manages network outages. It produced no comparable result, and AT&T’s own researcher described the findings as statistically inconclusive.
Neither company has published what the classical alternative was, or how it compared. This happens often in quantum computing releases: discussion about what quantum did, but not comparing it to alternatives that are commonly available in the marketplace.
Hear it from an AT&T executive himself:
“The results, I think, these are consistent comparable result. I do not claim anything better or worse because these are statistically not significant.”- Govinda Dhungana, Qubits 2026
The Revenue Itself
Q2 2026 revenue was $3.1m, flat against $3.1m a year earlier. This is a standard quarter for them, considering that every quarter without a “lumpy” system sale has been between $1.8M and $3.7M revenue since Q1 2024. And organic revenue has been decreasing since, which shows that existing customers are using their technology less.
This is not an industry problem. Compare D-Wave’s results to the following peers, this quarter:
- IonQ reported $80.1M in revenue, up 287% year over year and about 26x D-Wave’s quarterly figure
- Rigetti grew 185% year over year to $5.4M
- Quantinuum was $8m, up 280% y/o/y
Compare this to D-Wave’s 0% growth and $3.1m revenue.
To learn about how other quantum computing providers rank against each other in terms of technology and commercial traction, see our Quantum Industry Scorecard.
Also, if we believe that the AT&T deal and others are bringing in meaningful sales, yet revenue is zero… that means that current customers are abandoning the technology.
We’ve been predicting this, in published reports, for months. The problem with D-Wave is that customers believe their annealing technology is useless, but they get benefits from signing the initial Proof-of-Concept deal. That’s why you get initial excitement, but then the project never amounts to anything meaningful for either party (D-Wave or the client).
See our videos on technical explanations on why quantum annealing is useless.
About Margins and RPO
Margins also moved the wrong way alongside revenue. Gross margin fell to about 55% in Q2 2026 from about 64% a year earlier, a drop of roughly 8 percentage points. Revenue flat, margin down.
D-Wave’s remaining performance obligations fell from $42.4 million to $40.7 million in Q2, even as the company booked $2.1 million in new business. Backlog rises by what you book and falls by what you recognize as revenue, so adding $2.1 million in bookings and still ending $1.7 million lower means about $3.8 million left the backlog. Yet reported revenue was only $3.1 million, a $0.7 million gap between what exited the pipeline and what became revenue.
What this means: revenue and cancellations was faster than new bookings in the quarter. Here, $2.1m came in, but $3.8m left. Keep that in mind when getting excited about the AT&T partnership news.
The Truth About their Bookings
D-Wave’s remaining performance obligations fell from $42.4 million to $40.7 million in Q2, even as the company booked $2.1 million in new business. Backlog rises by what you book and falls by what you recognize as revenue, so adding $2.1 million in bookings and still ending $1.7 million lower means about $3.8 million left the backlog. Yet reported revenue was only $3.1 million, a $0.7 million gap between what exited the pipeline and what became revenue.
What this means: revenue and cancellations was faster than new bookings in the quarter. Here, $2.1m came in, but $3.8m left. Keep that in mind when getting excited about the AT&T partnership news.
Spending is Climbing Faster than Sales
R&D was more than $28M in Q2, roughly 9x quarterly revenue. IonQ spends far more in absolute terms, about $160.6M, but only about 2x its own revenue, because its revenue is much larger. D-Wave’s ratio is high because the denominator is not growing. The Q2 operating loss was $53.3M against $3.1M of revenue.
This was due to their investment in gate technology, which investors should be extremely skeptical about. Coherence Report recently rated D-Wave’s technology and roadmap in last place (Pretender tier), and believes the company has little to no chance of ever successfully profiting from what they are spending today.
For a deeper dive into D-Wave’s investment case, and where the business is headed, make sure to download our free D-Wave Investment Report.
Other Notable Deals and News
Some recent headlines, and our notes:
$1.5M National Science Foundation deal: See our writeup here: D-Wave Touts a Recycled Relationship to Show a $1.57 Million Yale Grant
$20m Florida Atlantic University sale: D-Wave’s announced system sale to Florida Atlantic University deserves scrutiny, both for its timing and for how little of the headline figure may convert to cash. It is D-Wave’s first sale to a US university in nearly a decade, and two details stand out. On geography, the sale coincides with D-Wave moving its headquarters to Boca Raton, the CEO’s longtime home city. On standing, FAU is a legitimate campus with some prior quantum activity, but was never a recognized national hub, and the deal’s own language talks about “establishing” FAU as a center rather than backing an existing one.
We are still investigating, but the pattern fits a CEO advancing a project at his hometown university. And a reported $20M agreement is not $20M received. Recognized revenue can fall well short of the announced number, and as of Q2 2026 it appears limited to roughly $300K of installation and site-prep work. A deal structured to support the narrative rather than meet a real need is also one that could quietly shrink or unwind, and the revenue that actually lands is the only part worth counting.
Why Customers Pay for Projects
“I think quantum gives us an opportunity to kind of, at least, study quantum itself. And then, you know, compare how it performs against the currently running classical models.” – Govinda Dhungana (AT&T), Qubits 2026 , January 27, 2026
It’s important to understand why companies engage with D-Wave (or most other quantum computing providers); in order to understand why you cannot follow headlines and press releases.
D-Wave does not sell projects to companies that hope to solve a problem (even if that is their stated goal). D-Wave’s customers are, for the most part, people paid to experiment with someone else’s money. For example:
- University researchers chasing grants,
- Government units with mandated innovation budgets,
- Corporate innovation arms inside large enterprises.
These groups engage even when they expect little to come of it, because they are incentivized to do so. A quantum pilot looks good in a press release, advances careers, produces publishable work, and helps justify next year’s budget. Everyone involved is rewarded for starting the project; but no one is penalized when it quietly ends.
The problem is what happens when a proof-of-concept has to move from that audience to the person who owns a real operating budget, the executive moving billions in inventory or running a network. That person does not care that the tool is quantum. They care whether it takes out cost or time versus what they already run, and against mature classical software the case usually isn’t there. That gap is why so few of these projects graduate from announcement to recurring revenue.
For a review on other factors affecting how customers perceive D-Wave’s technology, see the report from D-Wave insider Josh Speyer here.
Here is a list of other D-Wave projects that never came to fruition. Remember, every single project got just as much as excitement as AT&T did, and none of them went to production in a meaningful way. They certainly did not add any significant revenue to D-Wave’s quarter.
| Date | Customer | What They Used | What They Solved |
| May 2011 | Lockheed Martin | D-Wave One | “Most challenging computation problems,” later specified as building cyber-physical systems |
| Oct 2011 | Lockheed Martin | D-Wave One (installed at USC-Lockheed Martin Quantum Computation Center) | Began applied research |
| May 2013 | Google, NASA, USRA | D-Wave Two (Quantum Artificial Intelligence Lab) | Machine learning, later described as web search, speech recognition, planning/scheduling, air-traffic management, robotic planetary missions, mission-control support |
| May 2013 | D-Wave Two (acceptance testing) | Optimization problems at least 10,000x faster than classical solvers; highest scores on standard SAT-competition problems | |
| Jun 2013 | D-Wave Two | Problems solved 11,000x faster on average, 33,000x faster on the harder half, 50,000x faster on the top quartile (per D-Wave’s Colin Williams to the NYT) | |
| 2014 | Lockheed Martin | D-Wave quantum annealing | Verification and validation of complex software, such as flight control systems (published paper) |
| Nov 2015 | Lockheed Martin | D-Wave 2X (second system upgrade) | Continued optimization work |
| Dec 2015 | Google, NASA | D-Wave 2X | Quantum annealing 10⁸x faster than simulated annealing on a single core for hard optimization problems |
| Mar 2017 | Volkswagen | D-Wave quantum computer | Optimized traffic flow for Beijing taxis, reducing travel times across the city |
| Mar 2017 | Google, NASA, USRA | D-Wave 2000Q (upgraded) | Quantum-enhanced optimization and sampling algorithms for AI and machine learning |
| Oct 2019 | Volkswagen, CARRIS | D-Wave quantum computer | Routed 9 buses in real time in Lisbon during Web Summit, reducing passenger travel times |
| Sep 2020 | Menten AI | D-Wave Advantage hybrid solvers | Designed de novo proteins; peptide candidates synthesized and advanced to live-virus testing against COVID-19 |
| Sep 2020 | Save-On-Foods | D-Wave hybrid quantum algorithms | Cut weekly grocery optimization task from 25 hours to 2 minutes; in-store pilots underway |
| Sep 2020 | Accenture | D-Wave hybrid solvers | Helped a banking client pilot currency arbitrage, credit scoring, and trading optimization |
| ~2021-22 | SavantX, Fenix Marine Services | D-Wave Advantage | Optimized terminal operations at Pier 300, Port of LA; doubled cargo-handling equipment productivity, +60% deliveries per crane per day |
| ~2021-22 | Multiverse Computing | D-Wave hybrid solver service | Optimized a BBVA portfolio to 15% risk with a 60% return |
| ~2019-20 | DENSO | D-Wave | Optimized multimodal logistics and factory vehicle movement |
| Sep 2023 | Pattison Food Group | D-Wave annealing | Automated e-commerce driver scheduling across 100+ stores, reducing weekly manual scheduling effort by ~80% |
| Dec 2024 | NTT DOCOMO | D-Wave annealing | Improved mobile network resource utilization by 15% |
| Dec 2024 | Ford Otosan | D-Wave annealing | Reduced manufacturing production scheduling time by 83% |
| Mar 2025 | D-Wave (internal) | Advantage2 | Simulated nonequilibrium magnetic spin dynamics, framed as “an important, useful, real world problem” in materials science |
| Apr 2025 | Ford Otosan | Production-grade D-Wave hybrid application | Sequenced Ford Transit vehicles, cutting scheduling of 1,000 vehicles from 30 minutes to under 5; plans to extend to paint shops, assembly lines, buffer zones |
| Q1 2025 | Japan Tobacco (pharmaceutical division) | D-Wave Advantage with AI | Drug discovery; generated molecular structures described as better candidates than purely classical methods |
| Mid-2025 (Qubits conference) | NTT DOCOMO | D-Wave hybrid-quantum solution (production) | Optimized mobile network performance across 90M+ subscriptions |
| Mid-2025 (Qubits conference) | Davidson Technologies, QuantumBasel, SAS, Staque, Pusan National University, USC | D-Wave quantum optimization | Addressed computationally complex problems, no specific process improvements stated |
The Truth About Quantum Annealing
The bull case on the stock treats annealing as a general optimization engine. The reality is narrower, and explains why customers lose interest and move on from D-Wave shortly after engagement.
D-Wave’s quantum hardware only handles a narrow slice of optimization problems and can’t fit real enterprise-scale ones. Its “hybrid” solver covers the gap by doing most of the work on classical computers and sending only small pieces to the quantum chip, so the quantum contribution to most advertised use cases is minor, and by the company’s own account and outside studies, standard classical software matches or beats it on hard problems.
None of this makes annealing vaporware, but it does explain why customers pay for the initial engagement, experiment, and then abandon the technology.
Subscribe to our free newsletter, or follow us on X.com at @coherence_repor.
Dilution
Shareholders have funded much of this. Share count rose more than 130% from Q1 2024 to Q2 2026. The cash balance, often cited as a strength, was built largely through equity raises rather than operating cash. That cash is real and buys time. It also means any per-share view of future progress has to account for a much larger share base than three years ago.
The Core Question
There is a potential bull case, just one that Coherence Report does not think is likely. D-Wave has a large cash position, heavy R&D investment, a widening technology portfolio, and some genuine technical milestones. The bear case is that revenue is flat while peers compound, margins are slipping, dilution has been heavy, backlog is shrinking, near-term conversion is slow, and the excitement is driven by announcements rather than results.
Both can be true at once. Which one dominates comes down to a single question, and it is the question every partnership, PoC, grant, and conference slide should be measured against: when do the announcements start turning into sustained revenue growth? Until the answer shows up in the top line, the safer assumption is the one the numbers already support. The story is getting bigger, but the business is not.