Latest / Quickly Quantum / Ep 10: IBM's $10 Billion Quantum Bet Survives a 25% Stock Drop
Transcript
- IBM's stock cratered by as much as twenty-five percent in a single trading day this week — and in the very same investor letter where the CEO called the quarter disappointing, he doubled down on a ten-billion-dollar bet on quantum computing that won't pay off for years. Today on Quickly Quantum: does that kind of reaffirmation actually mean anything, or is committing to a five-year plan the cheapest thing you can say when the money hasn't gone out the door yet? Before that, in the headlines: NVIDIA claims its new AI decoder just slashed quantum error rates by up to three hundred forty-seven times, D-Wave is packing up its stock listing and moving to Nasdaq, and Rolls-Royce just signed on to use a quantum computer to help design jet engines. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Wednesday, July 15, 2026. Let's get into it. Let's start with the headlines, and NVIDIA is first up with a number that made me sit up. The company says its open-source Ising decoder family — that's the software that reads error signals off a quantum chip and figures out what actually went wrong — just got a new add-on for something called color codes, and in simulation, it cut the logical error rate by up to three hundred forty-seven point seven times. Quick gloss: the logical error rate is how often your error-corrected qubit — built from many physical qubits working together to catch and fix mistakes — still fails despite all that correction. Lower is better, and three hundred forty-seven times lower is not a rounding error. NVIDIA ran this at code distance thirty-one, essentially a stress test using a bigger, more error-resilient version of the code, and on top of the accuracy jump, the new decoder also ran seven point three times faster than the previous best classical decoder, called Chromobius. Speed matters here almost as much as accuracy, because in a real fault-tolerant machine the decoder has to keep pace with the qubits close to real time, or errors pile up faster than you can catch them. NVIDIA's own AI infrastructure account put it this way on X: quote, 'Quantum error correction just took a major leap forward. Color codes have always been limited by decoding performance. Today, NVIDIA Ising Decoding has changed that — delivering over 300x improvement in logical error rate and over 7x decoding speedup,' end quote. Now, this is a simulation result, not a demonstration on physical hardware, worth remembering before anyone gets carried away — but if it holds up on real chips, it revives color codes as a serious contender in the fault-tolerance race, a family of codes that had fallen out of fashion because nobody could decode them fast enough. Now, sticking with the corporate side of quantum: D-Wave Quantum is switching stock exchanges. The company's been trading on the New York Stock Exchange, but it's voluntarily moving its listing to Nasdaq, effective after market close on July twenty-fourth, with trading resuming under the same ticker, Q-B-T-S, on July twenty-seventh. Why do this? D-Wave frames it as aligning with a more tech-focused investor base — Nasdaq is where a lot of the growth and tech names live. As posted on X by @AilsaForshaw, D-Wave's CEO, Doctor Alan Baratz, said: 'Nasdaq is the marketplace for companies shaping the future of technology. As the first commercial quantum computing company and a leader in delivering real-world quantum performance, D-Wave is well aligned with the innovation Nasdaq represents.' Here's the context worth adding: this move comes while D-Wave's stock is trading well below its fifty-two week high, so switching venues by itself doesn't touch the fundamentals — it's not new revenue, it's not a new contract, it's a change of address. There can be real value in it: a different exchange sometimes attracts a different mix of index funds and active tech-stock traders, and there's brand-perception value in sitting next to Nasdaq's tech-heavy roster instead of the broader NYSE mix. But let's be clear-eyed about what this is and isn't. This is a corporate-plumbing story, not a quantum-computing story. The actual technology D-Wave sells — quantum annealing systems built for optimization problems — hasn't changed one bit because of where the ticker lives. Worth knowing, not worth overreacting to. Next, a story about making quantum security real-world rather than lab-only. Aliro Technologies and a company called zerothird say they've demonstrated a live, entanglement-based quantum key distribution network — QKD for short, a way of using quantum physics to detect if anyone's eavesdropping on an encryption key as it's shared — running directly on production-grade Cisco routing hardware. This happened at Cisco's Photonics Center in Vimercate, Italy, according to Quantum Computing Report; we haven't independently confirmed this one ourselves yet. Aliro's own account described it on X: 'Aliro Orchestrator is now serving as the quantum network management layer at the Cisco Photonics Center in Vimercate, Italy, operating and managing a live entanglement-based quantum key distribution network built on zerothird's eQKD system.' Why does 'production-grade' matter so much here? Because most quantum-networking demos to date have run on specialized lab benches, built for the experiment and nothing else. Running the same functionality on the same commercial routers that already sit in real enterprise networks is the difference between a proof of concept and something an IT department could plausibly deploy without ripping out its existing infrastructure. That's the real headline, if it holds up: not that entanglement-based QKD works — we've known that for a while — but that it might work on the hardware you already own. Quantum-safe networking has mostly been a story about post-quantum cryptography, algorithms designed to resist a quantum computer rather than quantum hardware itself; this is a rare entry from the other camp, the physics side of the security race. Now, from security to engineering: Quantinuum, Rolls-Royce, Riverlane, and the UK's EPCC supercomputing centre have launched a multi-year partnership aimed at speeding up computational fluid dynamics — the physics of how air and gas flow through machinery — specifically for gas turbine design. This is reported by Quantum Computing Report, and we haven't independently verified it beyond that source yet. Here's why it's worth attention regardless: Rolls-Royce is a real, recognizable industrial partner with a genuine engineering problem, not a quantum company manufacturing a headline. Jet engine and gas turbine design already leans heavily on classical supercomputers to simulate turbulent airflow, and those simulations are punishingly expensive — small improvements in accuracy or speed translate directly into fuel efficiency and design-cycle time, which is real money for an aerospace manufacturer. The plan is what's called a hybrid quantum-classical workflow, meaning the quantum computer doesn't replace the supercomputer, it handles a specific sub-piece of the calculation that classical machines struggle with, while the classical side does everything else. Riverlane brings error-correction expertise, and EPCC brings the high-performance computing infrastructure to run the hybrid pipeline at scale. Nobody's claiming a quantum advantage here yet — this is a research collaboration, not a product launch — but industrial partnerships with a named end-user and a named application are exactly the kind of story that tells you where near-term quantum value might land first: not in cracking encryption, but in shaving costs off things that already cost a fortune to simulate. Next, a workforce story. Photonic quantum computing company Xanadu is expanding its partnership with Lockheed Martin, integrating its open-source programming toolkit, PennyLane, into Lockheed's internal Quantum Talent Pipeline — a training program to build a bench of employees who can actually write quantum code. This is reported by Quantum Computing Report; we haven't independently confirmed it elsewhere yet. Lockheed Martin's own account posted on X: 'Lockheed Martin will explore quantum concepts with Xanadu's PennyLane, defining the future of aerospace and defense at the quantum level.' Xanadu, for its part, posted on X about the momentum continuing at the Chicago Quantum Exchange's Q-Ready Workshop, noting a session on quantum algorithms and resource estimation using PennyLane happening today, July fifteenth. Workforce stories don't usually make for thrilling television, but they matter more than they sound, because the industry's most quoted bottleneck isn't qubits, it's people — there aren't enough engineers who understand both quantum mechanics and how to actually program the hardware. Defense contractors like Lockheed Martin have an obvious motive: quantum computing shows up on national-security roadmaps, from cryptography to materials science to logistics optimization, and if you're not training your workforce now, you're behind when the hardware matures. It's not flashy, but training-pipeline partnerships like this one are the unglamorous infrastructure the rest of the industry's flashier headlines quietly depend on. And last in the headlines: the National Science Foundation has picked Connecticut as one of twelve regions nationwide to receive a Regional Innovation Engine award, unlocking a reported fifteen million dollars in federal funding tied to quantum technology. Connecticut has pledged up to one hundred twenty-one million dollars of its own money toward quantum startups and workforce development, which puts the ratio at roughly eight state dollars for every one federal dollar — a signal that the point of this award isn't the federal check itself, it's the much bigger state-level commitment it's designed to unlock. Southern Connecticut State University, which is leading workforce development for the effort, posted on X: 'Two million dollars is coming to New Haven to help shape Connecticut's quantum workforce. As part of the NSF's fifteen million dollar Regional Innovation Engines award, Southern will lead workforce development for the Connecticut quantum engine, expanding research, experiential learning, and career pathways for students across the state.' Now, fifteen million dollars is a rounding error next to the kind of money the biggest players in this industry throw around — and that's exactly the pivot I want to make, because our main story today is about a company treating quantum spending on a completely different scale. IBM just told investors, in the same breath as admitting a genuinely bad quarter, that it's not backing off a ten-billion-dollar, five-year quantum commitment. Fifteen million in regional seed funding and ten billion in corporate capital are two very different bets on the same underlying technology, and today we're digging into which of those commitments you should actually trust more. Our main story today, and I'm calling this one the reaffirmation that costs nothing — until it costs everything. Here's what happened. On July fourteenth, IBM released preliminary second-quarter results, and they were rough. Revenue came in at seventeen point two billion dollars, up just one percent from a year earlier, which missed Wall Street's estimates. Software revenue grew five percent, consulting revenue was basically flat, and infrastructure revenue — IBM's hardware business — fell seven percent, dragged down by weak sales of its Z mainframe line. Diluted GAAP earnings per share dropped two percent to two dollars and twenty-seven cents, and CEO Arvind Krishna called the quarter disappointing in a letter to investors. The market's verdict was blunt: IBM's stock fell by as much as twenty-five percent on the news. That's not a wobble, that's a gut punch. Krishna didn't dodge it — he said IBM failed to execute effectively during the quarter, with large customer deals not closing on the timeline expected. He also pointed to structural noise: some customers shifted spending late in the quarter toward servers, storage, and memory to lock in supply-constrained infrastructure before anticipated price increases, while broader cybersecurity concerns delayed other purchasing decisions. Not everything was bad — Red Hat revenue accelerated to eleven percent growth, recent acquisitions like HashiCorp and Confluent helped software, and IBM's distributed infrastructure business, boosted by Power systems and storage, posted its strongest performance, up thirty-seven percent. IBM also touted its new Lightwell cybersecurity platform, backed by a five-billion-dollar commitment and more than twenty thousand engineers, which went generally available on July eighth with major banks — including Bank of America, Citi, Goldman Sachs, and JPMorgan Chase — already signed on as early adopters. But here's the part that makes this a quantum story and not just an earnings story. In that same disappointing letter, Krishna spent real space reaffirming IBM's quantum roadmap, writing: 'Finally, quantum computing is no longer decades away, it is upon us, and we are investing aggressively.' He pointed to IBM's letter of intent, signed in May, with the U.S. Department of Commerce to build Anderon — a dedicated quantum wafer foundry, a factory purpose-built to manufacture the specialized chips quantum computers need, planned for Albany, New York. That project totals two billion dollars: up to one billion in CHIPS Act incentives, matched by a billion dollars of IBM's own cash, part of a broader two-point-oh-one-three-billion-dollar federal quantum push spread across nine companies including D-Wave, Rigetti, PsiQuantum, and Quantinuum. And Krishna reiterated the bigger number behind all of it: more than ten billion dollars in quantum investment over the next five years, covering research, manufacturing, capital spending, acquisitions, and ecosystem building — all in service of IBM's stated goal of delivering the industry's first large-scale, fault-tolerant quantum computer, meaning one that can reliably detect and correct its own errors well enough to run commercially useful calculations, by 2029. So here's my read on this. Reaffirming a plan is, structurally, the cheapest sentence in corporate English — it costs nothing to say 'we're still committed' in a letter, and it only costs something once the checks actually go out the door. That's not a knock on IBM specifically, it's just the mechanics of how these announcements work, and it's exactly why the skepticism here has to be sharper than usual. Start with the caveat IBM itself put on the number: these are preliminary, unaudited figures, and IBM has said final results could differ slightly from what it released. There's a full earnings call scheduled for July twenty-second, with complete results and updated full-year guidance. That call is the real test — not this letter, which reads a lot like Krishna managing a bad-news cycle by pairing it with a good-news story he'd already announced. The quantum reaffirmation isn't new information; it's the same ten-billion-dollar, five-year plan and the same 2029 fault-tolerance target IBM's been saying for a while, wrapped around a disappointing quarter to give investors something to hold onto. Now, I want to be fair to the other side of this, because it's not pure spin. The Anderon foundry commitment is real, already-structured money — a billion dollars of IBM's own cash matched against a billion in CHIPS Act incentives, part of a two-billion-dollar federal quantum push that also includes D-Wave, Rigetti, PsiQuantum, and Quantinuum. You don't sign a letter of intent with the Department of Commerce for theater. So this isn't purely rhetoric sitting on top of a bad quarter — there's a real, physical building with real committed dollars behind at least part of the ten-billion-dollar figure. But here's the tension the story doesn't resolve: IBM's own CEO, in the same letter, admitted the company failed to execute — deals didn't close, customers delayed purchases, and the stock dropped twenty-five percent because Wall Street didn't buy the excuses. That's a company currently struggling to execute on its near-term, revenue-generating business, telling investors to trust its execution on a moonshot that by definition won't generate revenue for years. Krishna's own words are the best evidence for both sides of this: quantum computing is, quote, 'no longer decades away, it is upon us' — spoken by the same executive explaining, in the same breath, why his existing infrastructure business just missed its numbers. So, time for the Hype Check. I'm putting this one at a five. Half credit, and here's why: the Anderon foundry money is real and already committed, which earns genuine points — this isn't a company promising jam tomorrow with nothing behind it. But reaffirming a plan you already announced, inside a letter designed to soften a genuinely bad quarter, is doing a lot of unearned rhetorical work, and 'we remain committed' is not a milestone. What would move my score, up or down, is what comes out of the July twenty-second call. Mark July twenty-second on your calendar — that earnings call is the real tell on whether IBM's quantum talk was strategy or substance, and we'll be right back on it the day it happens. If the show's earning your time, hit follow wherever you're listening, and if you've got a colleague who's curious about this stuff but doesn't know where to start, sending them this episode is the best compliment you can pay us. 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.