Latest / Quickly Quantum / July 17, 2026 — IBM's $67 Billion Wipeout — and Its $10B Quantum Bet
Transcript
- Today on Quickly Quantum: IBM just posted the worst trading day in its hundred and fifteen year history — a twenty five percent plunge that erased about sixty seven billion dollars in market value — and CEO Arvind Krishna's answer, right in the middle of the wreckage, was to double down on quantum. Before that, in the headlines: quantum stocks give back a third of their post executive order gains, imec drops the technical paper behind its EUV quantum chip, and a Dutch built quantum computer just landed at a German aerospace institute. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Friday, July 17, 2026. And if you thought the AI story and the quantum story were separate — think again, because today they collided head on. Let's get into it. Now, let's start with the quantum stock pullback everyone's watching. IonQ, D-Wave, Rigetti, and Quantum Computing Inc. have all given back roughly a third of the gains they'd banked since President Trump's June twenty second quantum executive orders. IonQ's the poster child here — it peaked above sixty dollars a share after the orders landed, and this week it slid down to around thirty nine. Now, before you panic, the timing matters: this pullback lines up with broad risk off selling tied to tensions in the Strait of Hormuz, not any company specific bad news. Analysts are calling it a valuation reset, not a fundamentals problem. And the fundamentals, for what it's worth, still look fine — IonQ guided second quarter revenue of sixty five to sixty eight million dollars, with remaining performance obligations, that's contracted future revenue not yet booked, up five hundred and fifty four percent year over year. And here's the thread worth pulling: those binding federal deadlines for post quantum cryptography migration, twenty thirty and twenty thirty one, are still sitting there as a durable revenue driver no matter what the stock does day to day. Worth watching whether the fundamentals catch up to the swings. Now, from stock swings to something more exotic, literally. Researchers have laid out a new theoretical route to universal, fault tolerant quantum computing using non-Abelian anyons — exotic quasiparticles, disturbances in matter rather than real particles, whose defining trick is that braiding them around each other produces a logic operation that depends only on the shape of the braid, not the messy details of the path. That built-in resistance to noise is the whole appeal of topological quantum computing. This isn't the first time we've heard the word anyon — Quantinuum previously demonstrated creating and braiding these things on its H2 trapped ion processor. What's new is the theoretical framework extending that toward running arbitrary algorithms with the same flexibility as a classical computer. Harvard's physics department flagged the paper on X. The skeptic's note writes itself: this is modeling and theory, not a new chip on a bench, and practical universal topological quantum computers are still years out. But if the theory holds, it's a genuinely different path to error resistance than the error correction everybody else is racing to build. Now, sticking with hardware manufacturing for a second. Belgium's imec has published the full technical paper behind the quantum test chip it first showed off back in May at ITF World. The headline back then was that imec used High-NA EUV lithography — the same extreme ultraviolet tool making next-generation logic chips — to pattern silicon quantum dot spin qubits with gaps around six nanometers, aimed at building quantum hardware right on standard chip manufacturing lines instead of specialized lab processes. The new arXiv preprint fills in the details behind that milestone. As @lithos_graphein, a semiconductor lithography expert who runs the Chip War Dashboard, put it on X: quote, 'imec just published their paper on the quantum test chip they made with EUV lithography. They did a press release some time back on this first-of-its-kind process, and now they made the full work public on arxiv,' end quote. We haven't independently confirmed the details beyond that post yet, and mainstream outlets haven't picked it up. But if you care about quantum riding on the same manufacturing rails as your phone's processor, this is the paper trail worth bookmarking. Now, over in the small-cap end of the quantum stock world: Quantum Computing Incorporated, ticker QUBT, has officially closed its acquisition of NHanced Semiconductors for seventy three point one million dollars. The deal was first announced back in June, and it's meant to expand QUBT's push into actual semiconductor manufacturing, not just quantum software or algorithms. That matters because the federal government is now backing a broader quantum semiconductor buildout — the same wave that's funding IBM's much larger Anderon foundry project — and QUBT wants a seat at that table too. Now, a dose of perspective here: QUBT is a much smaller, more speculative player than IonQ, Rigetti, or D-Wave, and it's genuinely unclear how quickly this acquisition turns into real revenue rather than just a bigger balance sheet. Still, it's one more data point in a pattern we keep seeing this year — quantum companies racing to control their own chip fabrication rather than depending on someone else's foundry. Now, from foundries to fiber optics. An Israeli company called Quantum Source Labs, working with Israel's defense research directorate — that's DDR&D — has demonstrated a single-atom source of entangled photon pairs built to survive actual deployment conditions, not just a clean lab bench. As the company put it on X, the achievement is on-demand generation of entangled photon pairs that survive real-world fiber conditions, with no active stabilization needed, meaning the system doesn't require constant fine-tuning to keep the entanglement — the quantum link between two photons — intact. That's the piece that's usually missing from lab demos: robustness. Entangled photon sources like this are a building block for quantum networks and quantum key distribution, the method of using quantum physics to detect eavesdroppers on a communication line. Now, this is a single-source claim straight from the company and its defense partner, so we haven't independently confirmed it, and there's no word yet on independent replication or a deployment timeline. But if it holds up, it's a real step toward quantum communication that works outside a climate-controlled lab. Now, a quick one for the error correction nerds — and I say that with love, because this is the plumbing that decides whether quantum computers ever scale. A team at Rice University introduced a new topological subsystem code, a mathematical scheme for detecting and fixing errors in quantum hardware, that only needs weight-three or weight-four local check operator measurements — jargon for saying you only have to look at three or four neighboring qubits at a time instead of some sprawling, hard-to-measure cluster. That's a meaningfully lower overhead template, built on something called an anticommuting quantum spin liquid, and the pitch is that it could apply across very different qubit platforms, not just one company's hardware. This is purely theoretical for now, with no hardware implementation reported, but reducing the complexity of these stabilizer measurements — the routine checks that catch errors before they spread — is one of the field's most persistent bottlenecks, so a cleaner template is worth banking even before anyone builds it. Now, one more, and it's a nice small signal of how global this race has gotten. The Netherlands has delivered its first-ever exported quantum computer, a Dutch-built machine now installed at Germany's DLR aerospace research institute. Details on the qubit count and architecture are thin in the reporting, but the bigger picture isn't: this is part of a broader European pattern of sovereign quantum hardware programs trading hardware across borders, rather than every country waiting for the Americans or the Chinese to hand them a machine. It's a small story next to what's coming next, but small stories like this are how you know an industry is maturing — when shipping a quantum computer to a neighboring country stops being a headline event and starts being routine business. Speaking of headline events, though — let's talk about the one that actually shook the market this week, because it happened to a company that's also the biggest name in quantum. Our main story today, and it's a big one: call it Big Blue's Quantum Alibi — the thesis being that when IBM's core business took its worst hit in over a century, quantum computing became the story management reached for to keep investors from bolting entirely. Here's what happened. On Tuesday, July fourteenth, IBM shares fell just over twenty five percent in a single session, closing around two hundred seventeen dollars. That's the worst single day in at least fifty eight years of trading records for the company, and possibly the worst in its hundred and ten years on the New York Stock Exchange. The previous record holder was Black Monday back in October nineteen eighty seven, when IBM fell twenty three point seven percent — but that was a market-wide crash. This time, there was no broader collapse to hide behind. This was IBM's problem alone, and roughly a quarter of a two hundred seventy two billion dollar company evaporated before lunch — call it about sixty seven billion dollars, gone in a morning. The trigger itself was unusual. IBM pre-announced its second quarter results a full week ahead of its scheduled July twenty second earnings call — something companies generally only do when the numbers are bad enough that sitting on them becomes a legal problem. Adjusted earnings came in at two dollars ninety three cents a share against a consensus of three-oh-two, and revenue landed at seventeen point two billion dollars against expectations of seventeen point eight six billion — a six hundred sixty million dollar miss. On its own, that's not historic-collapse material. What spooked the market was the why. CEO Arvind Krishna told investors that in the final weeks of June, enterprise clients abruptly redirected their quarterly spending away from software and mainframes and toward servers, storage, and memory — racing to lock in supply-constrained hardware before expected price hikes. Krishna's own words: quote, 'this quarter we faltered,' end quote. The mechanics matter beyond IBM: the global memory market has been squeezed since late twenty twenty five, as Samsung, SK Hynix, and Micron shifted fabrication capacity toward the high-bandwidth memory that feeds AI accelerators — leaving ordinary memory for servers and PCs in short supply. IBM's mainframe business, its z17 line, fared worse than the company's already-lowered expectations, and software growth slowed to five percent from eleven percent the quarter before. The pain spread fast — Salesforce fell around four percent that day, Microsoft nearly three, and a software sector ETF dropped two. At least one securities firm has already announced an investigation into whether IBM misrepresented the pace of new deal signings. And this wasn't even IBM's first bruising this year — back in February the stock fell thirteen percent in a day after Anthropic demonstrated an AI tool that modernizes COBOL, the decades-old programming language IBM's legacy business still runs on. Now, here's where quantum enters the frame. In the same investor letter admitting the company faltered, Krishna reaffirmed the ten billion dollar, five-year quantum investment and the twenty twenty nine target for Starling, IBM's fault-tolerant quantum computer — meaning a machine built from logical qubits, error-corrected qubits assembled from many noisy physical ones, robust enough to run real algorithms without drowning in errors. IBM also reaffirmed it's putting one billion dollars in cash into the proposed Anderon quantum wafer foundry, a project backed by up to a billion dollars in CHIPS Act incentives. And Krishna's language wasn't hedged at all: quote, 'Quantum computing is no longer decades away, it is upon us,' end quote. The very next day after the crash, IBM shipped Nighthawk R2, the second generation of its square-lattice quantum processor, right on schedule toward a promised seventy five hundred two-qubit gates by the end of this year, up from five thousand on the original chip — and it also pushed out Qiskit two point five, its quantum software framework, with a new compiler and faster performance. So here's my read, and I want to separate two different stories that are being told as one. Story number one: IBM had a genuinely bad quarter, and that's an AI story, not a quantum story. The AI infrastructure boom is not lifting all boats — it's actively draining budget from everything adjacent to it. Zoom out past IBM and the picture is grim: Oracle's down roughly a third year to date, Microsoft's off about twenty percent, and Accenture has been cut in half. Enterprise clients are hoarding memory and hardware ahead of shortage-driven price hikes, and that means software and infrastructure budgets everywhere are getting raided to pay for it. That's the real, near-term, verifiable trend, and it has nothing to do with qubits. And this isn't even the first time the market has done this to IBM this year. Twice in six months now, investors have repriced Big Blue on the suspicion that AI is coming for its legacy franchises — first in February, when Anthropic's COBOL modernization tool spooked the mainframe business, and now this week, over an AI-driven hardware and memory squeeze. Different mechanism, same underlying fear. Story number two is the one IBM wants you to focus on instead: quantum as the reassurance play. And the skeptic's note on this one is exactly right — quantum is being used defensively, as the long-term story to point to when the near-term numbers are ugly. None of the quantum roadmap actually changed this week. The ten billion dollar, five-year commitment, the twenty twenty nine Starling timeline, the Anderon foundry money — all of that was already public before Tuesday. What changed is the context it's landing in: a company whose stock just had its worst day in over a century, using quantum as proof that the long game is still intact. It's also landing the same week Washington's quantum executive orders and CHIPS Act equity stakes — the same policy wave that's been juicing IonQ and Rigetti's stock, and the same one that just gave some of that back — have made quantum an explicit pillar of IBM's turnaround pitch to a spooked market. Now, is that cynical? A little. But it's also not wrong on the merits. IBM isn't a peripheral player pretending to be relevant in quantum — it put the first quantum computer on the public cloud back in twenty sixteen, open-sourced Qiskit the year after, and Qiskit remains the closest thing gate-based quantum computing has to a standard, with a developer community nobody else is close to matching. The roadmap discipline has actually been the through-line here — milestone after milestone landing on schedule, right through Nighthawk and the experimental Loon chip last November, which IBM says demonstrates every hardware component fault tolerance requires. That cadence didn't skip a beat this week either — Nighthawk R2 and Qiskit two point five both shipped right on schedule, crash or no crash. So which read is right — is quantum the real IBM story wrapped inside a bad quarter, or is quantum the alibi for a bad quarter that has nothing to do with it? Honestly, both can be true at once, and that's kind of the point. The AI hardware crunch is real, and by Micron's own CEO's estimate it's going to keep squeezing memory supply well past twenty twenty seven. IBM's quantum bet is also real, funded, and on schedule. The danger is letting one story provide emotional cover for the other — letting 'quantum computing is upon us' distract from 'this quarter we faltered' in the very same letter. Watch what happens July twenty second, when IBM's full second quarter results and updated guidance land — that's the actual test of whether this was a one-quarter anomaly or something structural, and no amount of quantum roadmap talk changes that answer either way. Time for the Hype Check. I'm giving this one a four. Real dollars, a real foundry commitment, a genuinely disciplined roadmap — none of that's hype on its own. But the way it's being deployed this week, as the reassuring headline draped over a real earnings miss, is exactly the kind of narrative substitution this show exists to flag. The quantum story didn't get better on July fourteenth. IBM just needed you looking at it. These are threads we'll keep pulling on — especially once IBM's full numbers land July twenty second. If the show's earning your time, hit follow wherever you're listening, and if you've got a colleague who thinks quantum's still science fiction, this is a good week to send them an episode. That's Quickly Quantum for today. New episodes every day. This is an AI-voiced podcast, created and built by a real person using today's cutting-edge technology. And remember: nothing on this show is financial advice. I'm Brian Lampert — see you tomorrow.