AT&T's Quantum Speedup: One Hour to Fifteen Seconds

Today on Quickly Quantum: a Fortune 50 telecom just took a network optimization job that used to eat an hour of compute time, and got the answer back in under fifteen seconds — using a quantum computer, on a live operational problem, not a lab benchmark. Before that, in the headlines: a Swiss trapped-ion startup lands seed money to go two-dimensional, Pittsburgh's building a new hybrid quantum-classical supercomputer with Rigetti and HPE, Quantinuum pushes its error-correction numbers even lower, and researchers propose using sound waves to wire up a quantum chip. Welcome back to Quickly Quantum, your daily brief on the quantum frontier. It's Tuesday, July 28, 2026. Let's get into it.

Let's start in Switzerland. ZuriQ, a spinout of ETH Zürich, just raised twenty-five point five million dollars in seed funding to scale what's called a two-dimensional trapped-ion quantum processor. Now, quick gloss for anyone new to this: trapped-ion computers hold individual charged atoms in place with electromagnetic fields and use laser pulses to control them — it's one of the most accurate qubit technologies out there, prized for low error rates. The catch has been geometry. Today's leading trapped-ion machines, the ones from IonQ and Quantinuum, run on one-dimensional ion chains. ZuriQ wants to arrange them in a two-dimensional grid instead, which in theory gives you far more connections between qubits and a cleaner path to scaling up. @0noisee, a startup-funding tracker on X, flagged the raise as it happened. Now, I'll note the obvious caveat: this is seed-stage money — no qubit count, no working chip, no hardware milestone yet, just a well-pedigreed team and an architecture bet. But a credible European entrant challenging the American trapped-ion duopoly is worth watching.

Now, sticking with infrastructure: the Pittsburgh Supercomputing Center is building a brand-new hybrid quantum-classical supercomputer, teaming up with Rigetti Computing and HPE. The idea here is straightforward — instead of treating a quantum processor as some separate, exotic experiment down the hall, you wire one directly into a classical high-performance computing cluster, so researchers can hand off the parts of a problem that quantum handles well and let the classical machine do the rest. HPCwire, the trade publication, put it plainly on X: a hybrid quantum-classical testbed with HPE and Rigetti. And @QubitReport, another quantum-focused account, posted its own headline on the news: a new hybrid quantum-classical supercomputer coming to Pittsburgh. This fits a broader pattern — hybrid integration, not standalone quantum machines, is where most researchers think the near-term useful work actually happens. No timeline yet on when the quantum piece goes live, and no qubit specs disclosed, but it plants Rigetti's hardware inside national HPC infrastructure, which is a meaningful foothold.

Here's a fun one out of the University of Warwick. Researchers are proposing something they call Quantum Phononic Links — using engineered vibrations, essentially sound waves traveling through specially strained germanium-on-silicon material, to carry quantum information between qubits that are physically far apart on the same chip. Think of it as a tiny acoustic wire. Why does this matter? Because long-range wiring is one of the biggest unsolved problems in scaling a quantum computer past a few hundred qubits — packing in millions of qubits means finding a way to connect ones that aren't sitting right next to each other, and today's approaches mostly don't stretch that far. The researchers, publishing in APL Quantum with collaborators at NRC Canada, describe phonons acting as a 'quantum bus' between distant qubits, fully compatible with existing semiconductor manufacturing. Now, the honest caveat: this is a theoretical and materials proposal, not a working multi-qubit demonstration. Prior phonon-bus work has topped out at two-qubit proofs of concept. So file this under promising physics, not a lab result yet.

Now, over at Quantinuum — this is the same error-correction grind that keeps producing small, steady gains rather than one big leap. New research applied an integer-linear-programming technique — think of it as a solver that hunts for the most efficient way to arrange a fault-tolerant circuit — to build error-correction circuits, then tested them on Quantinuum's System Model H2 hardware. The result: running a Steane error-correction gadget on a two-block group algebra code — the notation is [[24,10,4]], don't worry about memorizing it — across ten thousand shots produced a logical block error rate of roughly zero point zero zero zero one four, or about zero point zero zero zero zero one four per logical qubit. The catch: about one point six percent of the shots had to be thrown out through post-selection — a common technique where you discard runs showing telltale signs of an error, but one that doesn't scale cleanly once you're running millions of shots on a large fault-tolerant machine. Still, this is the slow, unglamorous work — automated circuit optimization — that the error-correction roadmap actually runs on.

Now a quick one on preparedness. A new white paper finds Hong Kong's banking sector has a long road ahead on quantum security — banks there scored poorly on quantum-readiness assessments, according to The Quantum Insider — we haven't independently confirmed this one yet. The underlying worry isn't new but it's real: post-quantum cryptography is the new encryption standard designed to resist a future quantum computer breaking today's codes, and the fear driving urgency is 'harvest now, decrypt later' — adversaries scooping up encrypted financial data today, banking on being able to crack it once sufficiently powerful quantum machines exist. That urgency keeps showing up in sector-by-sector readiness reports worldwide, and this one adds another data point: most industries, banking included, look behind schedule on the shift.

Let's close the headlines with PsiQuantum, and this is the same story that keeps evolving in small steps. Quick refresher: PsiQuantum is building utility-scale, photonic quantum computers — using particles of light instead of trapped atoms or superconducting circuits — with active construction underway at Moreton Bay outside Brisbane and at the Illinois Quantum and Microelectronics Park near Chicago. Today's delta: Australia's US Ambassador posted on X highlighting continued progress at both sites. The Illinois Economic Development Corporation added color of its own, posting on X that Illinois' position as a global leader in quantum computing depends on a strong supply of local talent, and crediting PsiQuantum's investment in STEM programs for students on Chicago's South Side. Now, no new technical milestone here — this is a status highlight, not breaking hardware news. And it doesn't move the read from last week: PsiQuantum hasn't shipped a working large-scale machine yet, and construction progress and community investment are good signs, but the proof still has to come from working hardware.

Our main story today: call it the fifteen-second problem. AT&T — yes, the actual phone and internet company most of you get a bill from — just expanded its commercial deal with D-Wave Quantum to bring quantum annealing hardware deeper into how it runs its live network. Here's the headline number: in pilot testing, D-Wave's quantum annealing hardware took a network optimization workload that used to take AT&T about an hour to process, and cut it down to under fifteen seconds. That's a two-hundred-forty-times speedup, on a real operational task, not a research paper's synthetic benchmark. Now, quick gloss for anyone who hasn't followed quantum annealing before: it's a specialized type of quantum computer, different from the general-purpose 'gate-model' machines you hear about from companies like IBM or Google. Annealers are built specifically to solve optimization problems — the kind where you're hunting for the best arrangement out of an enormous number of possible combinations, like the most efficient way to route trucks, allocate bandwidth, or lay out a network. D-Wave has sold this approach to industrial customers for years; AT&T is expanding, not starting, that relationship. Under the new agreement, AT&T plans to layer quantum optimization into its 'agentic AI' systems — the industry term for autonomous software agents that make decisions and take actions without a human in the loop at every step. AT&T already leans on these agents for things like outage detection and automated fault isolation; the company says that in 2025 alone, these agentic systems helped cut total customer downtime by twelve million hours. The plan now is to hand the gnarliest combinatorial pieces of those workflows — the actual routing and scheduling math — over to D-Wave's quantum hardware. That includes real-time rerouting during fiber cuts or storms, optimizing field technician dispatch routes, managing bandwidth across 5G and fiber networks as AI traffic grows, and improving the math behind where AT&T builds new network infrastructure. There's also a longer-horizon piece: AT&T says it's evaluating D-Wave's upcoming gate-model quantum systems too, following D-Wave's acquisition of Quantum Circuits Incorporated and its dual-rail superconducting architecture roadmap. The interest there is squarely in security — AT&T's Chief Data Office is looking at how future fault-tolerant quantum machines could support quantum key distribution and quantum-resilient network encryption. That security angle is a longer bet — dual-rail superconducting qubits are a different hardware approach than D-Wave's core annealing business, and fault-tolerant, gate-model machines capable of running real quantum key distribution protocols remain a future ambition, not a near-term rollout. Now, here's why this landed as big as it did. This news broke the same week D-Wave voluntarily moved its stock listing over to the Nasdaq and rang the opening bell there. And the market reaction was immediate: D-Wave shares jumped more than twenty percent, up three dollars and thirty cents to close at nineteen dollars fifty-one cents. That rally didn't stay contained to D-Wave, either — it dragged IonQ, Rigetti, and Quantum Computing Inc. higher right along with it. Which tells you something about where investor sentiment sits right now: any concrete, named, Fortune-50 customer story sends the whole quantum-stock basket up together.

So how much of this is a genuine quantum breakthrough, and how much is a very well-timed press release? Let's dig in. On the enthusiastic side, @financial_mone, a quantum-investor account on X, flagged the numbers and the market response — the two-hundred-forty-times speedup and the stock pop across the sector. Their read, roughly: quantum's road to practical use just took a real step forward. And search coverage today showed the same one-hour-to-fifteen-seconds figure circulating across multiple financial outlets — Forbes and Yahoo Finance among them — which at least tells you this isn't a one-outlet rumor; the core claim is being reported consistently. And it's worth putting this in context: this is one of the more concrete, customer-facing 'quantum advantage' claims we've seen yet from a Fortune 50 company, on an actual live operational problem rather than a research benchmark run in a lab. That's a meaningfully different kind of proof point than another company publishing a paper about a toy problem. But here's where I want to slow down. Analysts covering this stock, even the bullish ones, keep flagging the same two things: D-Wave carries an extreme valuation for a company still posting losses, and the underlying comparison — quantum annealing hardware versus classical software on this specific workload — is a heuristic result measured against what AT&T hasn't specified as a rigorously optimized classical baseline. That distinction matters a lot to people in this field. Researchers have criticized this exact style of comparison before — quantum-versus-classical benchmarks where the classical side wasn't pushed as hard as it could've been. A two-hundred-forty-times speedup sounds enormous, and it might be completely real for this specific task. But it's not the same claim as a formally verified quantum-supremacy result, and the press materials don't spell out what the classical side of that comparison actually looked like. And then there's the timing. This announcement landed the same week D-Wave rang the Nasdaq opening bell — which doesn't make the AT&T relationship fake, but it does mean this was released, at least in part, as a marketing moment, not purely as a technical disclosure. Convenient timing, that. Now, to be clear about what the skeptics are NOT saying: nobody's disputing that AT&T and D-Wave have a real commercial relationship, and nobody's saying the fifteen-second result is fabricated. The workload got faster — that part checks out. The debate is about how to characterize why, and how much of that speedup would survive a more rigorous, apples-to-apples classical comparison. Time for the Hype Check. I'm putting this one at a six. Here's why: the customer relationship is real, the speedup figure is corroborated across multiple outlets, and a Fortune 50 company committing to expand — not just pilot — quantum hardware in live network operations is genuinely notable, more substance than most quantum press releases carry. But the comparison itself runs against an unspecified classical baseline, it's not an independently verified benchmark, and it dropped in the same week as a stock-listing bell-ringing that sent the whole sector's shares higher — so some of that number's punch is doing promotional work as much as technical work. Where this goes next matters more than today's headline: watch whether AT&T discloses actual production deployment results — not pilot numbers — once outage response and technician routing go live for real. That's the test that separates a marketing win from an operational one.

These are threads worth pulling this week — D-Wave's Nasdaq debut, its gate-model roadmap, and whether other Fortune 50 companies follow AT&T's lead into live network operations. If this episode helped make sense of the noise, hit follow wherever you're listening so tomorrow's show finds you automatically, and if you've got a colleague who'd get something out of this, send it their way. 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!