Latest / Quickly Quantum / September 6, 2026 — Physics or Plumbing: What's Actually Blocking Quantum Computing?
Transcript
- Today on Quickly Quantum: is the race to build a useful quantum computer still fundamentally a physics problem, or has it quietly turned into an engineering problem — pipelines, standards, workforce, the unglamorous stuff — the same shift semiconductors went through two generations ago? It's a claim that sounds almost too convenient coming from inside IBM's own orbit, and today we're spending the whole show on it instead of chasing headlines, because if it's true, it reorders what every lab, agency, and investor should be prioritizing right now. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Sunday, September 6, 2026. This is one of our Sunday think pieces — one subject, the whole hour, no quick hits — because the claim on the table this week is the kind that, if it's right, changes what labs and investors should be funding next, and if it's wrong, it's a comforting story dressed up as strategy. Let's get into it. So here's where this comes from. Back in January, a federally funded group called the Engineering Research Visioning Alliance — ERVA, set up with NSF money to identify national engineering priorities — published its twelfth report, warning the United States risks losing its quantum lead. The report organizes its recommendations around four pillars: quantum and materials, quantum and biology, quantum and computing, and quantum and artificial intelligence. Its co-chair is Brian Gaucher, an IBM veteran, and he's spent this year on podcasts and in interviews pushing the report's central argument, most recently in a September 5th interview with The Quantum Insider. His pitch, in short: the foundational physics of quantum computing is basically sound, and what's actually missing is the industrial machinery — shared pilot lines, agreed standards, metrology, trained people — needed to turn lab demonstrations into manufacturable systems. It's the exact transition the semiconductor industry made decades ago, going from garage experiments to fabs that could reliably stamp out identical chips, at scale, at a predictable cost. Now, we've talked before about IBM needing quantum to be the answer to its next act, and whether the market's actually buying that long-term bet. Gaucher's framing is worth hearing on its own terms first, because it's not a small claim — it's a bet that the scoreboard everyone's been watching, error rates and qubit counts, isn't where the real fight is anymore, and that's a genuinely provocative thing for an insider to say out loud. In his own words, in that Quantum Insider interview, Gaucher isn't declaring victory — he's declaring a closing window. "The physics breakthroughs will slowly be completed over time. But we still haven't shown quantum advantage yet," he says, meaning nobody's yet proven a quantum computer solving something a classical machine genuinely can't. And here's the part that reframes the whole debate: he locates the urgency in supply chains, not equations. "The urgency really comes from supply chains forming that can manage all of this, creating that standard that will work so we're not all building disparate types of hardware and systems. And we get the right people involved who are skilled to doing this, whether it's a neutral atom or a superconducting qubit." That's a genuinely interesting pivot — he's essentially saying the platform wars, superconducting versus trapped-ion versus neutral atom, are a distraction from the bigger question of whether any of them can be manufactured at scale with consistent quality, the way any given transistor design became almost irrelevant once fabs learned to mass-produce whichever one won. It's worth remembering he co-chaired the report making this exact case, so this is an argument to weigh, not a neutral diagnosis handed down from nowhere — he has every incentive to frame his field's next chapter as an engineering victory lap. But it's still a coherent argument: if you believe the physics roadmap holds, the next fight really is about factories, not formulas. Now, not everyone in the field would sign off on "the physics is basically done." The broader technical literature on decoherence — that's when a qubit's fragile quantum state gets scrambled by noise from its environment — still treats it as a live, unresolved research problem, not a settled constant you engineer around. Physics Today's standing explanation is blunt about it: quantum computers are highly susceptible to errors because the coherent states that give them their power are easily damaged by uncontrolled interactions with the environment, which is exactly why quantum error correction exists in the first place, as a defense mechanism against a problem physics hasn't closed out. And current industry analysis backs that up with hard numbers — as of this year, the state of the art is still noisy, limited to roughly fifty to two hundred error-prone physical qubits, with gate error rates sitting in the range of one in a thousand to one in a hundred, and qubits losing their coherence in microseconds to milliseconds. That's not a manufacturing footnote you can wave away with better tooling. That's the central technical constraint the whole field is still fighting, day to day, in the lab, and it sits awkwardly next to a claim that the physics side of the ledger is mostly wrapped up. There's a middle position here, though, and it might be the most honest one on offer. Analysis from IEEE's Computer Society, looking at Google's recent below-threshold error-correction results — meaning the codes finally suppress errors faster than they create them — argues the real story is that success just relocates the bottleneck rather than closing it outright. Once error-correcting codes actually work as intended, every part of the pipeline that touches syndrome data, timing, and control electronics becomes critical, which is exactly why the analysis calls quantum error correction "a full-stack problem" now, rather than a purely physics one or a purely engineering one. In other words, maybe Gaucher and the decoherence researchers aren't even fully disagreeing — they might just be describing two stages of the same climb, standing on different parts of the same mountain and describing what they each see. The physics clears one threshold, and the thing waiting on the other side looks a lot like the manufacturing, control-systems, and standards problem Gaucher's talking about. That would make "physics versus engineering" less of a fight with a winner and more of a relay race, with the baton already changing hands in places the public debate hasn't quite caught up to yet. And then there's the piece nobody in this debate fully owns: people. Even if Gaucher's engineering framing is exactly right, the capacity to execute on it doesn't exist yet, and that's arguably the more urgent shortage. The Quantum Insider's own economic reporting puts the global quantum computing workforce at roughly sixteen thousand five hundred professionals in 2025, while its economic-impact projections call for two hundred fifty thousand quantum-sector jobs by 2030 and eight hundred forty thousand by 2035 — worth flagging that these are industry-affiliated projections, not an independent audit, so treat the trajectory as a forecast with a rooting interest, not a certainty. The gap is sharpest in error correction specifically: Riverlane's QEC Report 2025 estimates only around eighteen hundred to twenty-two hundred quantum error-correction specialists exist worldwide, against a projected need of five thousand to sixteen thousand by 2030, with somewhere between half and two-thirds of quantum job openings currently going unfilled. A Forbes analysis cited alongside that data makes the pointed argument that people, not capital, may be quantum's real bottleneck — noting that no single country can meet its quantum workforce needs domestically, which undercuts any tidy national pilot-line strategy on its own. If that's right, it doesn't really settle whether the underlying problem is physics or engineering. It suggests the field doesn't have nearly enough humans who are genuinely fluent in both. So where do I land on this? I think Gaucher's framing is more useful than it first sounds, and also doing a bit more work than the evidence fully supports. He's right that nobody wins by treating quantum computing as a perpetual physics experiment forever — at some point you do need pilot lines, agreed standards, and metrology, the unglamorous machinery that turned semiconductors from a science project into an industry capable of shipping billions of identical chips. But the decoherence literature isn't quietly waiting for factories to catch up; gate error rates in the one-in-a-hundred range and qubits decohering in microseconds are not a solved problem wearing a manufacturing costume, they're an open research frontier that determines whether there's anything stable to engineer around in the first place. The full-stack read is the one I trust most here — error correction crossing a real threshold doesn't end the physics conversation, it hands the baton to control electronics and decoders, which is engineering, but engineering that only works because the physics held up under it. And the workforce numbers cut through the abstraction: a field with somewhere around two thousand qualified error-correction specialists worldwide, against demand many times that by 2030, doesn't have the luxury of neatly filing this under physics or engineering — it just needs more people who understand both, and neither camp in this debate is training them fast enough. For quantum computing specifically, here's the practical stake: if Gaucher's pilot-line strategy actually gets funded, it could shorten the road from today's noisy, small-scale machines to genuinely useful ones — but only if it's paired with, not substituted for, continued work on the error rates themselves. Time for the Hype Check. I'll put this one at a six — Gaucher's thesis is a real, well-argued reframe worth taking seriously, but it understates how unsettled the underlying physics still is, and a stakeholder's diagnosis of his own field's next phase deserves exactly that kind of pushback. On IBM specifically, this doesn't resolve whether the market's buying its long quantum bet, but it does sharpen the question — because if the fight really is shifting toward pilot lines and standards, IBM's early manufacturing muscle could matter more than its next physics headline. So here's the open question nobody in this debate has actually answered: would a semiconductor-style national pilot-line strategy resolve the platform disputes between superconducting, trapped-ion, neutral-atom, and photonic camps, or would it just lock in whichever platform happens to get funded first? If today's episode gave you a new way to think about the physics-versus-engineering fight, follow Quickly Quantum wherever you're listening so tomorrow's episode lands automatically. 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!