Council Post: Quantum And AI Are Becoming One Technology Bet

Dax Grant is the CEO of Global Transform and an authority on C-suite leadership. Global 100 CIO, 100 Women to Watch, CREA Global Award List.

getty

For a decade, quantum tech and artificial intelligence have run as parallel storylines in tech: one aiming to rewrite the limits of computation, the other already modifying how the world works. In 2026, those storylines are merging. The two fields have discovered they share both a bottleneck problem and, increasingly, a solution. AI is becoming the tool that makes quantum hardware usable, and quantum hardware is starting to look like the next lever for pushing AI past its own ceiling.

I spend significant time guiding leadership teams on which emerging technology bets are worth board attention. Two years ago, quantum technology and AI came up in separate meetings, on separate slides, championed by different people in the room. This year, for the first time, I’m hearing them discussed as a single line item—not because clients suddenly understand the physics but because the vendors they already buy from are now bringing the two together. The question I get most is whether we need to act yet, and the answer is shifting from no to not yet but watch closely.

The clearest evidence is in how quantum hardware is now being built. Quantum bits are notoriously fragile: a stray vibration or temperature fluctuation can corrupt a calculation in microseconds, which is why error correction, not raw qubit count, is the industry’s real scoreboard. That correction problem turns out to be exactly the kind of high-dimension pattern-recognition task machine learning is good at.​

Nvidia has built a business case around that overlap. Its NVQLink architecture, made broadly available via a new API at GTC 2026, is designed to connect quantum units directly to GPU supercomputers, enabling AI-based decoders to correct quantum errors in real time. In one demonstration on Quantinuum’s Helios QPU, an NVQLink-connected GPU decoder cut the logical error rate of an eight-qubit logical memory, a result that compounds toward longer, more useful computations. Seventeen quantum hardware companies and nine national labs have signed on to use the architecture, a sign that “AI-assisted quantum” is becoming standard infrastructure rather than a research curiosity.

My advice to clients evaluating this space: Stop counting qubits and start asking whose error-correction stack a vendor has bought into.​

Microsoft’s topological qubit program tells a similar story from the materials side. When the company unveiled its Majorana 2 chip this summer, it enlisted an agentic AI system called Microsoft Discovery to help, thereby enabling material combinations and automating fabrication testing, which increased qubit lifetimes by roughly a thousandfold compared with last year’s Majorana 1. Microsoft now expects to achieve a scalable quantum computer by 2029, cutting its original timeline in half. Directionally, I see that AI is no longer something quantum computers might someday run; it’s part of how they’re being engineered today.​

The traffic runs the other way too. IBM is building what it calls quantum-centric supercomputing, which it believes “will be a pivotal bridge to achieving quantum advantage.” Google’s Willow chip has separately shown that error-corrected logical qubits can now outlive their best individual physical qubits—evidence, Google says, that scalable error correction is now an engineering problem rather than an open scientific question.​

The commercial argument for caring about any of this is getting louder. McKinsey’s 2026 “Quantum Technology Monitor” report puts quantum computing’s prospective economic value at up to $2.7 trillion by 2035. The prior year’s edition of the same report identified AI and machine learning as one of four domains in which quantum technology is expected to have the most near-term impact: both by accelerating AI training and optimization operations and through the maturing domain of quantum machine learning, which uses quantum circuits to search pattern spaces conventional computers struggle with.​

Still, it’s worth resisting declaring victory. Quantum machine learning has yet to show a proven advantage over classical methods on a real commercial workload, and prediction markets tracking the field into the second half of 2026 expect “incremental engineering progress” rather than a breakthrough moment. My own read, after two years of watching vendor roadmaps slip and then compress again, is that the hardware is still outrunning the software.

That combination of real progress and real uncertainty is precisely what makes this time worth paying attention to. Unlike previous quantum computing cycles, this one has a concrete customer: AI systems that need better optimization, better materials discovery and better ways to search enormous solution spaces, and that are willing to pay for it. Nvidia, IBM, Google and Microsoft are not making these investments because quantum computing seems futuristic. They’re making them because each has concluded that quantum and AI now solve pieces of each other’s hardest problems.

In the boardrooms I sit in, the practical takeaway isn’t to bet the roadmap on quantum breakthroughs next quarter. It’s to recognize that quantum machines are no longer developing in isolation from the AI infrastructure they’re already building. Concretely, I recommend you ask cloud and AI vendors what they’re doing with NVQLink or similar integrations, and to put one small, reversible pilot on the roadmap, a real workflow with a named owner and a firm date. The organizations experimenting now will be the ones positioned to move quickly once the hardware catches up.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?