Latest / Quickly Quantum / The Week Quantum Money Got Real — And Real Small — September 19, 2026
Transcript
- Two hundred and fifteen million dollars. That's the size of the new prize the Department of Energy just put on the table this week for whoever builds the first fault-tolerant quantum computer — one that's error-corrected enough to actually do useful science. That's one of five big numbers running through the week in quantum: federal money finalizing into paperwork, a new benchmark saying the industry's still a hundred thousand times short of useful, and a real quantum computer now humming inside a Tennessee utility. Before we get into it — this is our weekly recap, so we're folding together everything from the week that was, the stories you already heard and a couple that slipped past without their own episode. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Saturday, September 19, 2026. Let's get into it. Let's start with the story we've been tracking since the spring. PsiQuantum, the photonics company betting its roadmap on light-based qubits — individual photons carrying quantum information instead of trapped atoms or superconducting circuits — finalized its hundred-million-dollar CHIPS Act award with the U.S. Department of Commerce on September 8th, roughly eleven days before this broadcast and just ahead of the week we're otherwise covering, so we're catching you up on it now. Now, if that number sounds familiar, it should: the letter of intent for this award actually went out back in May. What happened on September 8th is the paperwork catching up — the government and PsiQuantum turning a handshake deal into something legally binding. The money itself is earmarked for manufacturing research and development: optical switches, single-photon detectors, and the advanced packaging techniques PsiQuantum needs to build its machines at scale. And it's worth asking — the reporting doesn't fully answer this — how much of that hundred million is contingent on PsiQuantum actually hitting manufacturing milestones, versus money that's simply available now. That finalization fits the bigger pattern we've been watching this month: Washington converting quantum letters of intent into signed, executable checks. PsiQuantum wasn't alone; Anderon inked its own billion-dollar CHIPS deal around the same window. The open question we've flagged before hasn't gone away — PsiQuantum still hasn't shipped a working large-scale machine. A hundred million dollars in finalized federal paperwork buys the company runway and scrutiny in equal measure. Now for the reality check of the week. A new benchmark called QUOPS — that stands for quantum operations, built with contributions from Sandia National Laboratories, Quantinuum, and NVIDIA — set out to measure something nobody had pinned down cleanly before: the largest quantum circuits a machine can actually run before noise wrecks the answer. The result is blunt. Today's best hardware sits roughly a hundred thousand times below the scale you'd need to run a scientifically useful calculation — something that tells researchers something a classical supercomputer couldn't already tell them. That's a sobering number to land in the same week as a hundred million here and two hundred fifteen million there in federal quantum bets. But it's worth being precise about what the number actually measures. This is a snapshot against today's best physical hardware — the benchmark doesn't claim the hundred-thousand-fold gap is permanent, and it hasn't been adopted as an industry-wide standard yet. Nobody's publicly committed to reporting their own QUOPS score on a regular basis, at least not so far. It's an honest yardstick for where things stand today, the kind of number that keeps every roadmap slide a little more honest. For an industry announcing hundreds of millions of dollars in federal bets this same week, that's exactly the counterweight worth sitting with. Here's a story that flew under the radar this week, and it deserves more than a mention. EPB — the municipal utility that powers Chattanooga, Tennessee — switched on an IonQ Forte Enterprise quantum computer this week at its EPB Quantum Center. IonQ builds its machines out of trapped ions, individual charged atoms held in place by electromagnetic fields and steered with lasers — the company's core technology bet since its founding. What makes this deployment notable isn't just that it's live; EPB says it's combining commercial quantum computing and quantum networking under one roof for the first time anywhere in the country. Access rolls out in stages. Right now, it's EPB's own NIST-funded research fellows running experiments on the machine. Commercial access, including for the University of Tennessee at Chattanooga, opens when the service formally launches in early October — so don't mistake this week's power-on for the moment outside customers get to actually touch it, because that's still a few weeks out. And there's something genuinely striking about the geography here. This isn't a coastal research campus — it's a municipal utility in a mid-size American city running a quantum computer to work on its own grid problems. That's the direction this industry keeps bending, out of pure research demos and into places that pay an actual power bill. The caveats matter too. This is a single-system deployment, one utility, one machine, and EPB has floated a long-range economic-impact projection of one point one billion dollars by 2035 tied to this investment. That's EPB's own number, not an independently audited figure, and any organization forecasting its own decade-out payoff deserves the skepticism you'd give any company grading its own homework. Still, a working trapped-ion system tied into networking infrastructure, running inside a public utility rather than a research lab, is a real data point for the shift from demonstrations toward things people actually operate day to day. Sticking with federal money for a minute, because the Department of Energy dropped something structurally different this week from the CHIPS Act deals we just discussed. It's called the Quantum Genesis Q Competition, and the number is big: up to two hundred fifteen million dollars for any private-sector company that can deploy a fault-tolerant quantum computer — one that uses error correction to stitch reliable logical qubits together out of many noisy physical ones — with at least a hundred logical qubits capable of running hundreds of millions of fault-tolerant operations. Hit that bar, and there's bonus money for scaling further, up to a hundred fifty and two hundred logical qubits. Here's what makes this different from the PsiQuantum and Anderon awards — those are manufacturing grants, money to build fabrication capability. This is a prize competition, milestone-based, tied to an informal 2028 target for someone, anyone, to actually deliver a working fault-tolerant machine at that scale. Now, before you picture DOE cutting a two-hundred-fifteen-million-dollar check, read the fine print. Only two-and-a-half million dollars of that is actually guaranteed in this fiscal year's budget. The rest depends on future congressional appropriations coming through, and, more fundamentally, on some company actually hitting a hundred logical qubits running hundreds of millions of operations. Nobody has done that yet. So how do you even verify a claim like that? DOE hasn't detailed exactly how it plans to independently confirm a hundred-logical-qubit result before it pays out real money, and that's the process detail worth watching as this competition gets underway. What this tells me is Washington is done funding promises and wants to start funding proof — the money follows the achievement, it doesn't fund the attempt. Given how many big quantum announcements get walked back or asterisked later, that might be the more honest way to spend two hundred fifteen million taxpayer dollars. Last story of the week, and it's a quieter one, but it fits a pattern we flagged just over three weeks ago. This week adds another data point. Diraq and Dell Technologies started a technical collaboration to physically co-locate a Dell high-performance computing server cluster inside Diraq's lab in Sydney, right alongside its quantum hardware. The goal is straightforward: cut the latency between classical and quantum processing by putting the two literally in the same room instead of routing data across a network. It's part of a broader theme running through the week — the industry building out the plumbing that connects classical supercomputers to quantum processors. Now, I want to be careful with the word partnership here, because Quantum Computing Report describes this specifically as an active technical testbed, and no timeline was given for when a customer could actually use this setup. So today, that's lab infrastructure — nothing customers can buy yet. That doesn't make Diraq's roadmap proven, but it does mean the modality is no longer starving for attention or capital the way it was a couple of years ago. That's the week: federal dollars finalizing into paperwork, a benchmark humbling the whole industry, a utility in Tennessee running a live trapped-ion computer, a quarter-billion-dollar prize with almost none of it guaranteed yet, and a silicon-spin lab wiring itself straight into Dell's hardware. If any one of those stories landed for you, today's episode is a good one to send to a colleague who keeps asking what quantum computing actually is right now. Follow Quickly Quantum wherever you listen, and I'll be back with the daily show this week. 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! I also host Concrete Compute: a daily briefing on the AI buildout. The datacenters, the megawatts, and who actually pays for them. Find it wherever you get your podcasts.