AT&T Hits 240x Speedup With D-Wave Quantum Computer in First Real-World Quantum Advantage
For years, quantum advantage has been the industry’s favorite moving target. Every time a quantum computer beats a classical one at something, someone argues the something does not matter. Google’s Sycamore generated random numbers faster than a supercomputer in 2019, and critics said the task was useless. Chinese researchers claimed advantage on a different benchmark in 2021, and the debate restarted. The pattern was consistent: quantum computers kept getting better at problems that made for good papers but not good business cases.
This time, the something is running network operations for AT&T, and the numbers are hard to argue with. The task was not chosen because it makes quantum look impressive. It was chosen because it is one of the hardest computational problems the carrier faces every day, and classical computers were not keeping up.
D-Wave announced on July 27 that AT&T is now using its quantum annealing technology to shrink a network optimization workload from roughly an hour of classical processing to under 15 seconds. That is a 240x speedup on a task that directly affects how one of America’s largest carriers routes traffic, balances loads, and allocates resources across its live network. This is not a contrived benchmark designed to make quantum look good on a carefully selected problem. It is production infrastructure serving millions of customers every day.
AT&T has signed an agreement to expand its use of D-Wave’s quantum computing technology across its live network operations, covering traffic routing, load balancing, and resource allocation. The carrier is also evaluating D-Wave’s forthcoming gate-model machines for quantum security and communications, which would represent a longer-term bet on additional quantum capabilities beyond the annealing approach currently in production.
For D-Wave, the deal validates a strategy that has sometimes looked lonely. While IBM, Google, and a wave of venture-backed startups poured billions into building universal gate-model quantum computers, D-Wave stayed focused on quantum annealing and real-world optimization problems. The company took criticism for years from researchers who argued annealing was a dead end and that only gate-model machines would ever deliver commercial value. The AT&T deployment does not settle that debate, but it does prove that annealing has a commercial path that gate-model does not yet have, and that path runs through some of the largest companies in the world.
“The speed we’re seeing with D-Wave challenges what’s currently possible,” AT&T’s Lucus Haugen said in a statement. “It has the potential to help us optimize faster, increase efficiency and scale more real-time operations, making quantum a practical tool for advancing how we run and modernize our network.”
Why This One Is Different
Quantum advantage claims have a checkered history. Google’s 2019 Sycamore demonstration was criticized because the task it solved had no practical value: generating random numbers that a quantum computer could produce faster than a classical one, but that nobody actually needed. IBM argued it was a synthetic benchmark designed to play to quantum’s strengths, and that a classical supercomputer could eventually match it with enough time and clever engineering. The argument was technically valid and commercially irrelevant: nobody was going to spend weeks of supercomputer time to generate random numbers they did not need. But the criticism stuck.
Since then, every quantum advantage claim has been met with the same question: does this actually matter outside the lab? The AT&T deployment answers that question directly. The task is not synthetic. Network optimization is a real, computationally expensive problem that directly affects AT&T’s business. It involves routing millions of data flows across tens of thousands of nodes in real time, accounting for capacity constraints, latency requirements, and fault tolerance. Classical approaches use heuristics and approximations because exact optimization across a network of AT&T’s scale is computationally infeasible. Those approximations work well enough most of the time, but they leave performance on the table.
Shaving an hour of compute down to 15 seconds means the carrier can respond to network conditions in near real-time rather than running periodic batch optimizations that are always slightly out of date by the time they finish. In a network context, an hour-old optimization is already stale. Traffic patterns shift in minutes. A 15-second optimization cycle means AT&T can essentially run continuously optimized routing, adjusting to conditions as they change rather than reacting after the fact.
The second reason this deployment is different is the comparison baseline. The quantum solution is not beating a theoretical classical ceiling or a research paper’s estimate of what a classical computer might achieve with unlimited time and resources. It is beating what AT&T was actually running in production. The comparison is against the carrier’s existing infrastructure and software stack, which has been optimized by teams of engineers over years. A 240x improvement over a production system that was already well-tuned is much more meaningful than a 240x improvement over a naive baseline.
D-Wave’s approach is also different from what most people picture when they think of quantum computing. The company uses quantum annealing, a technique optimized for optimization problems rather than general-purpose computation. It is less flexible than the gate-model quantum computers IBM and Google are building, but for the specific class of problems AT&T is targeting, routing and resource allocation, it is exactly the right tool. Quantum annealing finds low-energy states in complex systems, which maps directly onto finding optimal configurations in a network with millions of variables.
The tradeoff is real. A gate-model quantum computer can theoretically run any quantum algorithm, including Shor’s algorithm for factoring and Grover’s algorithm for search. But the gate-model machines available today are still noisy and error-prone, with qubit counts in the low hundreds and coherence times measured in microseconds. They are not ready for production optimization workloads at AT&T’s scale. D-Wave’s annealing machines have thousands of qubits and are purpose-built for optimization, which means they skip the generality and go straight to the problem class that has the most immediate commercial value. For AT&T’s use case, that is the right tradeoff. For other use cases, it will not be, and the market will sort that out over time.
The Business Context
The announcement coincided with D-Wave’s Nasdaq debut, and the market noticed. D-Wave’s stock jumped on the news, and the broader quantum computing sector got a lift as well. The timing is hard to ignore: a major enterprise deployment announced on the same day as a public listing is the kind of coordination that signals confidence from both the company and its largest customer.
For AT&T, the investment makes sense on purely economic grounds. Network optimization is a compute-intensive task that grows more complex as networks add nodes and traffic patterns become less predictable. Classical approaches scale poorly with network size because the number of possible configurations grows exponentially. Heuristics and approximations keep the problem tractable, but they leave money on the table in the form of underutilized capacity, suboptimal routing, and slower response to congestion events. If quantum can deliver a 240x speedup on current workloads, the value proposition for larger, more complex networks down the line is substantial. AT&T spends billions annually on network infrastructure. Even a single-digit percentage improvement in utilization from better optimization could represent hundreds of millions in saved capital expenditure.
The deal also positions AT&T as a first mover in enterprise quantum adoption at a time when most quantum computing headlines are still about research milestones rather than production deployments. For a carrier in a competitive market where network performance is a differentiator, being able to say you are running quantum-optimized infrastructure is more than a press release. It is a signal to enterprise customers that you are building for the next decade, not just the next quarter. Competitors will almost certainly follow, but AT&T gets the first-mover narrative and whatever operational advantages come from being the first to deploy at scale.
What This Means for Quantum Computing
The industry has been waiting for a clean enterprise win. Not a lab result. Not a proof-of-concept with a friendly partner who has equity in your company. A paying customer using quantum computing to solve a real problem at a scale where the improvement is measurable and meaningful to the business. AT&T and D-Wave just delivered that.
The effect on investment is likely to be immediate. Venture capital has been cautious on quantum hardware in 2026 after a few high-profile startups struggled to hit their technical roadmaps and the broader tech downturn made investors skeptical of capital-intensive deep tech. A deployment with a Fortune 500 company that shows clear, quantifiable ROI changes the narrative from “someday, maybe” to “now, at least for optimization problems.” Expect quantum annealing companies in particular to point to this deal as validation. D-Wave’s competitors in the annealing space, and there are several, will rush to announce their own enterprise partnerships. Some of those will be real. Some will be press releases in search of a product. The market will sort them out.
The broader gate-model quantum efforts from IBM, Google, IonQ, and others are not directly challenged by this announcement, since they are targeting a different class of problems with a different approach that has a longer time horizon. But they benefit from the rising tide. Every enterprise quantum deployment makes the technology feel less speculative to the CIOs and CTOs who control the budgets for the next round of quantum adoption. The conversation shifts from “should we have a quantum strategy” to “who is our D-Wave, and what optimization problems should we throw at them first.” AT&T is also evaluating D-Wave’s gate-model machines for quantum security applications, which suggests the carrier sees annealing as the near-term play and gate-model as the longer-term horizon for different use cases.
The 240x number will get the headlines, and it should. That is the kind of improvement that makes engineers stop scrolling and procurement departments schedule meetings. But the more important number might be one: the number of times a major U.S. carrier has previously used quantum computing in live production operations. That number just went from zero to one, and once the first domino falls in an industry as competitive as telecommunications, the rest tend to follow faster than anyone expects.
