Saturday, May 23, 2026

Show Me the Code: The Long Vindication of D-Wave

This is the story of D-Wave, built from D-Wave’s own filings and investor materials. The tone runs bullish, and it should: this is my first pass at the bull case, told from the company’s own side of the record, where every story wears its best clothes. That is the point of this piece. In future notes, I will challenge this version of the story, dig deeper into the science underneath it, and revise the thesis where the facts demand it. This is not a valuationEstimating what a company is actually worth, typically from its projected future cash flows., and it is not financial advice. The story, as I tell it here, is my own opinion. Do your own research, and decide for yourself whether you agree with me.

I have been reading D-Wave’s filings, and the thing I keep coming back to is the sentence its CEO chose to open his most recent earnings call. Quoting Linus Torvalds, Alan Baratz said: “Talk is cheap. Show me the code.”¹ It is a strange line for a hardware company, but he did not mean software. He meant proof. He meant that in a quantum-computing sector swollen with promises about machines that will matter in fifteen or thirty years, D-Wave is the company that already has customers running real problems on real machines today. That is the whole pitch, and it is also, conveniently, the discipline a careful reader should bring to the pitch itself: do not tell me the story, show me the number underneath it.

So let me do both. Because the D-Wave story is, I think, the most genuinely interesting one in quantum computing, and it is interesting for a reason almost nobody leads with: this is the oldest company in the field, the one that was written off, and it has just spent eighteen months turning vindication into a second act. The stock tells you something happened. Two years ago it traded near a dollar and flirted with delistingBeing removed from a stock exchange, often for trading too low or breaking listing rules..² As I write, it sits around $29, for an equity valueThe total value of a company's shares. Closely related to market cap. somewhere between eight and nine and a half billion dollars.² That is an enormous number to hang on a company that booked $24.6 million of revenueThe total money a company brings in from sales, before any costs are subtracted. last year.³ The gap between those two figures is exactly the space a story has to fill, and D-Wave’s story has more load to bear than most.

The heresy

D-Wave was founded in 1999 in British Columbia, by Haig Farris, Geordie Rose, Bob Wiens, and Alexandre Zagoskin, which makes it older than the iPod, older than Tesla, older than the smartphone.⁴ From the start it made a contrarian bet that would define both its glory and its long exile. While the rest of the eventual field chased the universal, gate-model quantum computer (the kind that, in principle, runs any quantum algorithm, the kind IBM and Google and later IonQ would build), D-Wave built something narrower and stranger: a quantum annealer.

Here is the one piece of jargon worth defining, because the entire story rests on it. A gate-model machine is a general-purpose quantum computer. An annealer is a special-purpose one. It does essentially a single thing: it finds the lowest-energy configuration of a system, which turns out to be the same mathematical shape as a vast class of optimization problems. How do you schedule ten thousand shifts, route a thousand delivery trucks, lay out a factory line. Annealing is built, by its physics, to answer questions like those, and it is far more tolerant of noise than a gate-model machine, which is why D-Wave could ship working hardware while the universal-computer crowd was still fighting error rates in the lab.

And ship it did. In 2011, D-Wave announced the D-Wave One, a 128-qubit machine it called the world’s first commercially available quantum computer. Lockheed Martin bought one. So did a Google and NASA partnership, the University of Southern California, and Los Alamos.⁴ By any fair telling, D-Wave invented the category of the commercial quantum computer.

For this, it was treated as something close to a fraud. For the better part of a decade, the academic and commercial establishment dismissed D-Wave’s annealer as not “real” quantum computing, a niche curiosity at best, a marketing exercise at worst. The criticism had teeth: skeptics published classical algorithms that matched D-Wave’s machines, and the question of whether the thing was doing anything genuinely quantum hung over the company for years.⁴ Baratz still bristles at it. “Many still view D-Wave through an outdated lens,” he told investors this month. “It’s time for a vision check.”¹ When a CEO is still fighting a characterization, the characterization mattered. D-Wave spent its first two decades as the field’s heretic: the pioneer nobody would canonize.

The numbers carried the scars. The company went public in August 2022 through a SPACA shell company that raises money through an IPO, then merges with a real business to take it public quickly. merger and listed on the NYSE as QBTS.⁵ Revenue then sat almost perfectly flat for years, $8.8 million in 2023 and $8.8 million again in 2024, against operating losses that ran north of $77 million a year.⁵ By 2023 the stock had collapsed toward a dollar.² Here is the life-cycle paradox that makes D-Wave so hard to place: by age it is a mature company of twenty-seven years, but by its financials, an accumulated deficitThe running total of all the losses a company has piled up since it began. of $626.9 million,⁵ revenue you could fit inside a mid-sized car dealership, it is a start-up that simply refused to either grow up or die. It had been “early” for an entire generation of computing.

The change in fortune

Every story turns, and D-Wave’s turn has a date: March 12, 2025. That day, the journal Science published a paper from D-Wave with a deliberately provocative claim. Using an Advantage2 prototype, the company had performed a simulation of magnetic materials that, it argued, lay beyond the reach of classical computing entirely.⁶ The framing matters. Google and others had claimed “quantum supremacyA demonstration that a quantum computer performed a specific task no classical computer could do in any reasonable time, even if the task itself is not useful.” before, but on contrived, useless benchmarks like random circuit samplingA contrived benchmark task used to claim quantum supremacy. It has no practical use of its own.. D-Wave claimed it on a problem with real scientific meaning, and the comparison it offered was vivid: the calculation took minutes and under a dollar of electricity, while reproducing it on Frontier, one of the world’s most powerful supercomputers, would have taken close to a million years and more than the planet’s annual electricity consumption.⁶

You do not have to take the claim at face value (I will come back to why) to see what it did. After a decade of being told its machine was a toy, D-Wave had published, in the most prestigious venue in science, an argument that its toy did something no classical computer could. The heretic produced a miracle, and the congregation had to at least look up.

The financials turned with it. RevenueThe total money a company brings in from sales, before any costs are subtracted. jumped 179% in 2025, to $24.6 million.³ The balance sheet, which had nearly killed the company, was transformed: D-Wave raised aggressively into the rising stock and ended 2025 with $884.5 million in cash, up nearly 400% in a year.³ Management began saying, repeatedly and pointedly, that D-Wave could become the first independent public quantum company to reach sustained profitability, and to do it with less capital than any of its peers.¹ A company that had been one bad quarter from the exit was suddenly, by its own description, fully funded to the finish line.

Why annealing wins, where it wins

Before the reinvention, I want to be clear about what is genuinely strong here, because it is the bedrock under everything else. The “Why,” the reason this company has a durable reason to exist, is narrow but real: for optimization problems, D-Wave’s annealer is not just competitive, it is in a class of its own, and the structure of the technology suggests it will stay there.

Three things hold this moatA durable advantage that protects a company from competitors, like the moat around a castle. together. First, annealing is structurally resilient to errors. A gate-model machine needs elaborate, expensive error correctionTechniques that combine many shaky physical qubits into fewer reliable ones, so a long calculation stays correct. before it can do anything useful, and that overhead is crushing. An annealer tolerates noise by its nature and produces good answers without it. Second, D-Wave is the only player at scale. After twenty-seven years, it is essentially a monopoly in annealing; the company notes that the few competitors now attempting it are working with two-, three-, or four-qubit systems, while D-Wave runs machines with more than 4,000 qubitsThe basic unit of a quantum computer. Like a 'bit' in a normal computer, but instead of being only 0 or 1 it can be 0, 1, or a blend of both at once. and is designing toward 100,000.⁸ Third, and most underrated, is operational maturity: D-Wave is the only company that has run cryogenic quantum systems at commercial-grade uptimeThe percentage of time a system is running and usable (for example 99.9%). A sign of commercial-grade reliability., with years-long operation and 99.9% service-level availability on its Leap cloud.⁶ Its on-chip cryogenic control is the unglamorous crown jewel here, letting it steer 4,000 qubits with roughly 200 control wires, where rival superconducting approaches need three to five wires per qubit⁸ and head toward, in Baratz’s words, football-field-sized installations.¹

The benchmark below is the kind of “show me the code” evidence that makes the moat concrete rather than rhetorical.

Source: QED-CThe Quantum Economic Development Consortium, an industry group that publishes neutral benchmarks. benchmarking cited by D-Wave.⁶ The point is not the exact figures but the order of magnitude: on the problems annealing is built for, gate-model machines are not in the same race.

And crucially, the revenueThe total money a company brings in from sales, before any costs are subtracted. that flows from this is real commercial demand, not grant money. Over 73% of D-Wave’s Q1 2026 revenue came from commercial enterprises (the company is pointed about contrasting this with IonQ and Rigetti, whose revenue it characterizes as dominated by government grants).⁶ The customer list is not a research roster: Ford Otosan scheduling a thousand vehicles per run, NTT DOCOMO improving mobile-network performance by 15%, Pattison Food Group lifting scheduling efficiency 80%, BASF, Mastercard, BBVA, a clutch of Forbes Global 2000 names.⁶ This is the part of the story that clears every hurdle cleanly. D-Wave owns optimization, optimization is a genuine market, and that alone is a defensible business.

The reinvention

Which brings us to January 2026, and the move that turned a vindication into an ambition. D-Wave acquired Quantum CircuitsA sequence of quantum gates applied to qubits, read left to right in time and ending in measurement. The recipe a quantum computer runs., Inc., paying $250 million in cash plus stock,⁹ and in doing so it did something that should make you sit up: the annealing company, the one that had bet its entire identity on not building a universal computer, bought its way into building one.

This is the archetype shift, and it is the most important strategic fact about D-Wave today. For most of its life D-Wave was a Bully inside a single room (the undisputed master of annealing) while being treated as an Underdog in the broader category. The Quantum Circuits deal is a bid to walk out of that room and become a Bully of the entire house. Management’s phrasing is not subtle. The acquisition deck declares that D-Wave is now “the world’s leading quantum computing company” and that, with both platforms, it intends “to singlehandedly lead the market.”⁹ The logic: customers will need annealing and gate-model machines to cover the full range of computational problems, and D-Wave says it is the only company that can sell them both.

What makes this more than a press release is who came with the deal. Quantum Circuits was co-founded by Dr. Robert Schoelkopf of Yale, and Schoelkopf is not a marketing hire. He is a member of the National Academy of Sciences and the inventor of the transmonThe most widely used superconducting qubit design, invented by physicist Robert Schoelkopf's group., the qubitThe basic unit of a quantum computer. Like a 'bit' in a normal computer, but instead of being only 0 or 1 it can be 0, 1, or a blend of both at once. design that most of the superconducting industry, including IBM and Google, actually uses. He also pioneered the “dual-rail” qubit, which is the technical heart of D-Wave’s gate-model bet.⁹

You can see D-Wave assembling a cast of founders across three eras. Geordie Rose made the original contrarian bet on annealing and is long gone.⁴ Alan Baratz, CEO since 2020,⁵ is the Experience-narrative operator, the career technology executive (not a physicist) brought in to convert science into commerce, and his combative, take-no-prisoners style (“stop spreading competitive misinformation and start doing your homework”)¹ is the voice of a company done apologizing for itself. And now Schoelkopf arrives as the Charisma founder, the scientific eminence whose presence is meant to make the gate-model leg credible overnight. The dual-rail pitch is genuinely elegant: it claims to fuse the speed of superconducting qubitsA qubit made from tiny electrical circuits chilled to near absolute zero, where they lose all electrical resistance. (gate operations up to 1,000 times faster than trapped ionsA qubit made from a single electrically charged atom held in place by electromagnetic fields and controlled with lasers. or neutral atomsA qubit made from an uncharged atom held in place by focused laser beams ('optical tweezers').) with the high fidelity usually reserved for those slower technologies, by building error detection directly into the qubit.¹ The qubits flag roughly 90% of errors as they occur, which D-Wave says lets it build a logical qubitA reliable 'qubit' built by bundling many error-prone physical qubits together with error correction. These are the units that actually matter for useful computing. from up to ten times fewer physical qubitsAn actual piece of qubit hardware. On its own it is fragile and makes frequent errors. than rivals need.¹

The strategic prize is the size of the market this opens. Leaning on Boston Consulting Group’s framework, D-Wave maps a quantum opportunity that grows from a few billion dollars near-term into the hundreds of billions over fifteen to thirty years, with optimization (annealing’s home, where D-Wave stands alone) as a large slice, and the rest requiring gate-model machines.⁶

BCG sizes the long-term (15 to 30 year) quantum total addressable market at $450 to $850 billion, with $90 to $170 billion of that accruing to hardware, software, and services providers.⁶ The dual-platform claim is, in essence, a claim on all three rows rather than just the first.

The reversal: the proof is not the whole story

A narrative with no friction is a sales pitch, and Baratz’s “show me the code” creed cuts both ways. So let me show you the parts of the code that give me pause, because they are real.

Start with the revenueThe total money a company brings in from sales, before any costs are subtracted., the supposed proof of commercial arrival. That 179% growth in 2025 is genuine, but roughly half of it was a single system sale to the Julich Supercomputing Centre, recognized mostly in one quarter.⁶ The very next quarter, Q1 2026, revenue fell 81% year over year to $2.9 million, precisely because no comparable system closed in the period.⁷ BookingsThe value of new orders signed during a period, which may be recognized as revenue only later. tell the same volatile tale from the other side: full-year 2025 bookings were actually down 22% from 2024,³ and then a single quarter, Q1 2026, delivered $33.4 million of bookings, nearly twice the entire prior year.¹ A bull sees an inflection. A skeptic sees a business so lumpy that any single quarter, up or down, is noise. The honest answer is that we do not yet have enough quarters to know which it is.

Then there is the valuationEstimating what a company is actually worth, typically from its projected future cash flows. gap, which I raised at the start and which only widens on inspection. An equity valueThe total value of a company's shares. Closely related to market cap. near nine billion dollars² on $24.6 million of revenue³ is more than 300 times sales. Almost the entire market value of this company is a claim on a future that has not happened. That is not a criticism unique to D-Wave (it is true of the whole quantum sector), but it means the story is doing essentially all of the work, and the story has to be very good to justify it.

Now run the three-part test that any business narrative has to survive. Is D-Wave’s story possible? Unquestionably. The machines exist, the customers are real, the supremacy paper is peer-reviewed.⁶ Is it plausible? For the annealing-and-optimization story, yes, strongly: D-Wave is the only player, the performance edge is documented, the revenue is commercial. But the dual-platform story, “we will singlehandedly lead the entire quantum market,” is where plausibility strains. The gate-model leg that is supposed to unlock the larger half of the market is, today, an eight-qubit machine in a New Haven lab,⁹ with a roadmap to 100 logical qubitsA reliable 'qubit' built by bundling many error-prone physical qubits together with error correction. These are the units that actually matter for useful computing. by 2032.¹ Is it probable, the single most likely outcome among the plausible ones? Here I have to separate the two D-Waves. That D-Wave remains the dominant force in optimization and grows a real, defensible business there: probable. That D-Wave leapfrogs IBM, Google, IonQ, and Quantinuum, all better funded and years ahead in gate-model qubit countsThe raw number of physical qubits in a machine. A weak measure of power on its own, since quality and connectivity matter more., to lead the whole field: not probable, at least not yet. The most likely outcome is a company that owns a smaller-but-genuine optimization market and is a credible, not dominant, second-wave gate-model contender.

Two more wrinkles deserve naming. First, an irony: the company that loudly mocks rivals for “moving the goalposts” on their roadmaps is now publishing its own seven-year roadmap to 100 logical qubits, and every quantum roadmap in history has slipped, by D-Wave’s own observation.¹ Second, a genuine break risk that history makes specific: D-Wave’s supremacy claims have been challenged by classical algorithms before.⁴ The company itself notes there are “no other unspoofed claims” in the industry,⁶ which is a confident way of admitting that “spoofing,” a classical method catching up and erasing the quantum advantageA quantum computer solving a genuinely useful problem faster, cheaper, or better than the best classical computer., is the precise thing that could detonate the thesis. It has happened to D-Wave’s claims in the past. If it happens to the 2025 Science result, the heretic’s miracle becomes an asterisk.

The story, told in value drivers

A story earns its keep only if every claim in it points to a number. Here is the bridge from this narrative to the drivers a valuationEstimating what a company is actually worth, typically from its projected future cash flows. would need.

What would change this story

The story is not fixed, and it is worth being precise about what moves it.

A Break, the kind of event that decimates the narrative, would be a credible “spoofing” of the 2025 supremacy result by a classical algorithm, which would gut the central proof point and revive the old “not real quantum” charge. In the same category: classical and GPU-based optimization (the work NVIDIA and others are doing)¹ advancing fast enough that quantum’s edge in optimization stops translating into dollars, or an outright failure of the gate-model roadmap to leave the lab.

A Change, a fundamental reshaping that does not kill the story, would be the dual-platform bet actually paying off: dual-rail hitting its milestones and D-Wave emerging as a true full-spectrum player, completing the archetype’s leap from one-room Bully to category leader. The opposite Change is just as plausible and in some ways sturdier: the gate ambition quietly fades, and D-Wave settles into being “the optimization company,” a smaller identity but a more honest and more defensible one.

A Shift, a smaller adjustment, is the ordinary rhythm investors will actually live with: the lumpiness of system-sale timing, the bookingsThe value of new orders signed during a period, which may be recognized as revenue only later. cadence, the pace at which research-and-government customers convert to commercial, and whether the promised two to three system sales a year materialize.¹

I will end where Baratz began. Show me the code. After twenty-seven years, D-Wave finally has some: a peer-reviewed supremacy result, customers in production, a monopoly in a market that turns out not to be a niche, and a balance sheet that can fund the wait. That is a real change in fortune for a company the field had buried, and it is enough to make the optimization story probable on its own terms. The larger ambition, to lead all of quantum computing with two platforms, is a different and unproven bet, layered on top of a gate-model business that is, for now, eight qubitsThe basic unit of a quantum computer. Like a 'bit' in a normal computer, but instead of being only 0 or 1 it can be 0, 1, or a blend of both at once. and a roadmap. The most useful way to hold D-Wave is to keep those two stories separate in your head, the way the company, for all its bravado, cannot quite afford to. The patient heretic was right once, and it took a generation to prove. Whether it is right again is a question only the next several years of code can answer.

Sources & notes

1. D-Wave Quantum Q1 2026 earnings call, prepared remarks and Q&A (May 12, 2026). Source of the Linus Torvalds “show me the code” opening and the “outdated lens / vision check” framing; the “stop spreading competitive misinformation and start doing your homework” remark; management’s “fully funded path to profitability” claim; the BCG $100-220 billion optimization sizing; the dual-rail performance claims (roughly 90% of errors flagged, up to 10x fewer physical qubits, gate operations up to 1,000x faster); the gate-model roadmap to 100 logical qubits by 2032; Q1 2026 bookings of $33.4 million; and the “two to three system sales a year” guidance.
2. Market data for D-Wave Quantum (QBTS): a share price near $29 and an equity value of roughly $7.5 to $9.5 billion (varying by source and date), the stock’s earlier collapse toward $1, and 2023 delisting risk. Yahoo Finance, finance.yahoo.com/quote/QBTS, and companiesmarketcap.com, companiesmarketcap.com.
3. D-Wave Quantum fourth-quarter and full-year 2025 results, released February 26, 2026: revenue of $24.6 million (up 179% from $8.8 million in 2024), year-end cash and marketable securities of $884.5 million, and full-year 2025 bookings of $18.7 million (down 22% from $23.9 million in 2024). D-Wave press release via Business Wire, businesswire.com; The Quantum Insider, thequantuminsider.com.
4. Company history: D-Wave Systems was founded in 1999 near Vancouver, British Columbia, by Haig Farris, Geordie Rose, Bob Wiens, and Alexandre Zagoskin; it announced the 128-qubit D-Wave One in 2011 as the first commercially available quantum computer, with early customers including Lockheed Martin, a Google and NASA partnership, USC, and Los Alamos; and its early supremacy-type claims were repeatedly challenged by classical algorithms. Wikipedia, “D-Wave Systems,” en.wikipedia.org/wiki/D-Wave_Systems.
5. D-Wave Quantum FY2024 Annual Report (Form 10-K): the August 2022 SPAC merger and NYSE listing as QBTS; FY2023 and FY2024 revenue of $8.8 million each; operating losses in excess of $77 million per year; an accumulated deficit of $626.9 million; and Alan Baratz as CEO since 2020.
6. D-Wave Quantum Q1 2026 investor presentation: the March 2025 Science quantum-supremacy result on a magnetic-materials simulation (versus the Frontier supercomputer, with the minutes-and-under-a-dollar comparison); 99.9%+ availability of the Leap quantum cloud; the QED-C optimization benchmark (about 98% of optimal at problem size 320, against gate-model rivals); over 73% of Q1 2026 revenue from commercial enterprises and the named customers (Ford Otosan, NTT DOCOMO, Pattison Food Group, and others); the BCG $450-$850 billion quantum TAM with about 20% accruing to hardware, software, and services providers; and the “no other unspoofed claims” comparison.
7. D-Wave Quantum Q1 2026 Form 10-Q: first-quarter 2026 revenue of $2.9 million, down 81% year over year; and the New Haven, Connecticut gate-model R&D center acquired with Quantum Circuits.
8. D-Wave “Technology and the Quantum Competitive Landscape” deck (January 27, 2026): annealing qubit counts (more than 4,000 today, scaling toward 100,000 in the Advantage3 system) versus rivals’ few-qubit annealers; the control-line / wiring-complexity comparison; and the modality-by-problem mapping.
9. D-Wave / Quantum Circuits acquisition press-conference materials (January 2026): the roughly $250 million cash-and-stock acquisition; the “world’s leading quantum computing company” and “singlehandedly lead the market” claims; and Dr. Robert Schoelkopf (Yale, member of the National Academy of Sciences, inventor of the transmon and the dual-rail qubit) with the New Haven team. The Aqumen Seeker is an 8-qubit dual-rail processor per Quantum Circuits’ product announcement, prnewswire.com.