Latest / Quickly Quantum / August 6, 2026 — Microsoft's Quantum Chief Doesn't Care Who Doubts Him
Transcript
- Today on Quickly Quantum: what happens when the man running Microsoft's quantum hardware program tells you, on the record, that he genuinely doesn't care whether most of the physics world believes his results? Before that, in the headlines: IonQ just posted its fifth straight record quarter and raised its own revenue forecast, D-Wave says it's found a gate that could cut quantum error rates by a factor of ten, and DARPA is quietly shifting its quantum bets from lab demos to factory floors. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Thursday, August 6, 2026. Let's get into it. And by the end of the show, you can decide for yourself whether Microsoft's confidence is earned, or just very well-funded. Let's start with a number that made me sit up: eighty million dollars. That's IonQ's second-quarter revenue — eighty-point-one million, up two hundred eighty-seven percent from a year ago and twenty percent above the guidance IonQ itself had set. It's the company's fifth consecutive record quarter, and it raised its full-year 2026 outlook to a range of two hundred eighty to two hundred ninety million dollars. The growth is coming from Tempo system deployments in South Korea and Switzerland, and IonQ also signed a defense-focused agreement with Anduril, pushing further into national-security applications across its whole stack — computing, networking, sensing. Days after the quarter closed, IonQ finished its one-point-eight-billion-dollar acquisition of chipmaker SkyWater, giving it its own in-house semiconductor fab. CEO Niccolo de Masi called it, quote, "the strongest quarter in our company's history." Now, before you get too swept up: IonQ is still deeply unprofitable. Its adjusted EBITDA loss widened to a hundred twenty million dollars as spending accelerates around the SkyWater integration, and this guidance doesn't even include SkyWater's revenue yet. On X, independent investor @avgvalueinvestor put the growth in perspective, writing, "$IONQ just did more revenue in one quarter than 2021 to 2024 combined!" That's a real number worth sitting with — and also a reminder that revenue and profit are two very different scoreboards. Ninety-nine-point-nine percent fidelity. That's the headline number in D-Wave's new Nature paper, describing a two-qubit gate for what's called a dual-rail erasure qubit — a design where errors are easier to catch because most of them show up as a qubit simply disappearing rather than flipping silently. The gate runs in five hundred nanoseconds, and D-Wave's own simulations suggest it could cut the logical error rate — the error rate of a fully error-corrected qubit built from many physical ones — by a factor of ten with each added layer of correction. That's notable partly because D-Wave has spent its whole history on annealing hardware, a narrower machine built for optimization problems, and this is its clearest pivot yet toward general-purpose, gate-model quantum computing, with a stated goal of a hundred logical qubits by 2032. CEO Alan Baratz framed the stakes plainly: "Gate-model quantum computing's greatest remaining challenge is not simply building more qubits. It is building systems that can correct errors efficiently as they scale." Chief development officer Trevor Lanting added that fault tolerance means "systematically solving a series of difficult scientific and engineering challenges, with each success bringing us closer to a scalable system." On X, physicist and investor @QInvestorNotes dug into the numbers, noting that in a surface code simulation, this gate's error structure cuts logical error twenty-seven-fold for every two steps of added code distance, versus fourteen-fold for a conventional noise model with five times lower raw error. The catch: those multipliers come from simulation built on top of the gate result, not a demonstrated full error-correction cycle running at scale yet. Sixty-two kilometers of ordinary, aerial internet fiber — strung on poles, swaying in the wind, occasionally used as a bird perch — just carried entangled photons from Gaithersburg, Maryland to the University of Maryland without losing the entanglement. That's the headline from a new NIST study done with the Joint Quantum Institute and the company Qunnect. Entanglement is the "spooky action at a distance" Einstein was skeptical of himself: two particles share a single quantum state, so measuring one instantly tells you about the other, no matter the distance. The reason this matters economically is that nobody wants to build a whole separate fiber network just for quantum applications, so this experiment tested whether entangled photons can survive real commercial fiber instead. NIST physicist Oliver Slattery didn't sugarcoat the conditions: "It's about as bad a connection as you can possibly have." That it worked anyway is the point. What it isn't yet: a working quantum network. Useful quantum networking still needs quantum memories and repeaters — devices that store and relay entangled states — and neither was part of this demonstration. DARPA's turning its attention from lab demos to factory floors. The agency just launched a program called "It's About Time," aimed at building a pilot manufacturing pipeline for tactical-grade optical clocks — ultra-precise timekeepers that use the frequency of light instead of microwaves, useful for things like GPS-free navigation. It builds on two earlier DARPA efforts, ROCkN and OASIC, which took optical clocks from fragile lab prototypes toward field-tested hardware. The twist: DARPA tapped IonQ, through its Vector Atomic subsidiary, to actually build the pipeline, with a production facility expected by mid-2027 and deliverable units about a year after that. Program manager Mukund Vengalattore didn't mince words about why this matters: "For too long, the quantum community has driven the state-of-the-art exclusively within the laboratory, but these innovations rarely make it into the field." This is the same instinct behind IonQ's own SkyWater fab purchase — the government increasingly treats manufacturing and supply chain, not just the next physics breakthrough, as the real bottleneck standing between quantum tech and anything deployable. One more brief note, on PsiQuantum. The pitch there has stayed consistent: rather than chasing exotic new materials, PsiQuantum bets that reusing existing semiconductor fab infrastructure is the fastest road to a million-qubit photonic quantum computer — a real strategic contrast to the superconducting and trapped-ion approaches everyone else is running. Today, CEO Victor Peng took that pitch to Bloomberg TV, walking through the company's path to what it calls utility-scale, fault-tolerant quantum computing. There isn't a lot of new substance in today's appearance itself — it's mostly the company's own account promoting the segment — so the read here hasn't moved: PsiQuantum has previously put its own commercial-scale machine around 2030, and the company still hasn't shipped a working large-scale system, so the proof remains something only working hardware can provide. Credit where it's due, though — PsiQuantum's own post included the line, "Feynman said nobody understands quantum mechanics, marketing departments missed that memo." That's funnier than most quantum PR, and I'll take it. Our main story today: what happens when the science says no, but the roadmap says yes anyway. WIRED just published a profile of Chetan Nayak, the physicist running Microsoft's quantum hardware program, and the blunt question underneath it is whether Nayak cares that most of the physics community doesn't buy Microsoft's biggest quantum claim. His answer, in effect: not really. And that disagreement is splitting the field. Quick refresher for anyone new to this: back in February of 2025, Microsoft published a paper in Nature claiming it had built the first topological qubit — a fundamentally different way of storing quantum information, designed to resist errors by nature rather than through brute-force correction, built on exotic particles called Majorana states that were predicted decades ago and never definitively observed before. It was a massive claim. And the physics community never fully bought it. In June, University of St. Andrews physicist Henry Legg published his own peer-reviewed critique — in Nature, the same journal — arguing that Microsoft's own tuning software, quote, "yielded inconsistent and misreported outcomes." And Microsoft's most persistent public critic, University of Pittsburgh's Sergey Frolov, has gone further still, at one point calling the entire effort, quote, "essentially a fraudulent project." Here's why this is more than an academic squabble: the stakes are real money and real government policy. Microsoft has accelerated its own roadmap toward a working quantum system by 2029. The U.S. government has poured two billion dollars into quantum research, chasing its own target of a scientific quantum system by 2028. And Nayak's team advancing into the final phase of DARPA's Quantum Benchmarking Initiative — an independent government evaluation of quantum hardware claims — is the piece of outside validation Microsoft keeps pointing to. So you've got a company racing toward a government-adjacent deadline, a government funding that race, and a meaningful chunk of the physics community saying the underlying science hasn't been proven. Worth sitting with before we go further: Microsoft's original Nature paper wasn't independently peer-reviewed to the standard critics wanted, and Microsoft keeps much of its supporting data private for commercial reasons — meaning outside physicists are being asked to judge a roadmap built partly on evidence they literally cannot see. And this isn't the first stumble here, either: two earlier Microsoft-backed Majorana papers were retracted from Nature before this one ever published. So when Nayak says he doesn't care what the skeptics think, the fair question is — should he? Let's hear how Nayak actually makes his case. His go-to metaphor is people standing next to an airplane arguing, quote, "is flight possible or not." To him, the debate is settled — the DARPA evaluation, built partly on data Microsoft hasn't made public, is proof enough, and the ongoing skepticism is just science doing what science does. Henry Legg isn't buying the metaphor. His peer-reviewed critique argues Microsoft's own tuning software "yielded inconsistent and misreported outcomes" — which, if that holds up, means the foundational 2025 result doesn't hang together the way it was originally presented. And Sergey Frolov goes furthest of all, arguing Microsoft lacks the "longstanding evidence" that rivals like IBM and Quantinuum have built without ever needing to prove Majorana particles exist in the first place — meaning those companies' qubits work and are useful regardless of how this particular physics debate resolves, while Microsoft's whole architecture is staked on winning it. Frolov says the field has gotten so used to the back-and-forth that people now "chuckle" when Microsoft's Majorana work comes up at physics conferences. Here's where I land, and it's not a comfortable place. The problem isn't that Microsoft made a bold claim — bold claims are how physics moves, and skepticism is supposed to be normal, healthy, boring science. The problem is the asymmetry: Microsoft is racing on a public deadline, backed by public and private money, while validating its own work with data outside scientists can't fully audit. Nayak's DARPA answer only works if you trust an evaluation whose details haven't been published either. That's not proof of fraud — nobody credible outside Frolov's harshest language is claiming that — but it's a real gap between what's being claimed out loud and what's actually been shown. And there's history that makes me more cautious, not less: two earlier Microsoft-backed Majorana papers were retracted from Nature before this one. That's not a company that's earned the benefit of the doubt on this specific claim — that's a company asking for it again. Time for the Hype Check. Here's the case: an accelerating roadmap tied to a government-adjacent 2029 deadline, a validation process built on data the public can't see, a peer-reviewed critique this year specifically challenging the software behind the original result, and a track record that already includes two retracted papers on this exact topic — stack all of that up, and Nayak's confidence looks less like settled science and more like a company that's decided the roadmap has to be true whether or not physics agrees, which is why I'm putting this one at a 3. If Microsoft actually hits 2029, the physicists who spent two years publicly challenging the data end up looking stubborn; if it doesn't, it's taxpayers and Microsoft shareholders left holding a bet that skipped the usual proof. Either way, this is a fight worth revisiting once more of that DARPA data actually becomes public. If you want the rest of this week's quantum stories waiting in your feed, follow or subscribe to Quickly Quantum wherever you're listening. This has been Quickly Quantum, an AI-voiced podcast, created and built by a real human using today's cutting-edge technology. Nothing you heard on this show is financial advice. I'm Brian Lampert, and I'll catch you all tomorrow — take care!