I want to tell you about quantum computing the way a slightly careful friend would — excited by the research, honest about the timeline, and allergic to the hype cycle that's been spinning up around this space for about a decade now.
Quantum AI is real. It is probably going to matter. It is almost certainly not going to matter for your business in the next five years. Here's why, and how to think about it anyway.
What a quantum computer actually is
A classical computer stores information in bits — each one a 0 or a 1. A quantum computer uses qubits, which, through the magic of quantum mechanics, can be in a superposition of both states at once. When you have many qubits working together, you can, in principle, explore many possibilities simultaneously.
This turns out to be extremely useful for a specific set of problems — factoring large numbers, simulating molecules, certain kinds of optimisation. For most problems, though, it offers no speedup at all, and for some it's actively worse than classical computing.
The myth that quantum computers are just "faster classical computers" is one of the most persistent and unhelpful misunderstandings in tech.
Where quantum and AI intersect
Optimisation. Many AI systems have optimisation problems buried inside them — finding the best set of parameters for a model, the best route for a delivery fleet, the best scheduling of jobs in a factory. Some of these problems may be dramatically faster on quantum hardware, eventually. "Eventually" is doing some work there.
Simulating molecular systems. Drug discovery and materials science both involve simulating quantum systems — and what better way to simulate a quantum system than on another quantum system? This is probably where quantum computing will first pay its rent. Pharmaceutical AI teams are already experimenting.
Certain kinds of machine learning. A subfield called quantum machine learning is exploring whether quantum algorithms can train models faster, or produce models with different capabilities. Research is lively. Practical results are scarce.
Breaking current cryptography. The other side of the coin: quantum computers will eventually be able to break much of the public-key cryptography we rely on today. Post-quantum cryptography is being standardised now, and any organisation with data that needs to be secure for decades should be migrating. This is not a theoretical concern.
What the current state of the art actually looks like
Honest report, as of where we are today:
- Qubit counts are climbing. IBM, Google, IonQ, and others have systems with hundreds to thousands of qubits.
- But noise is brutal. Qubits are extremely fragile. Most of the computation is currently spent on error correction. The "useful" computational power is far below what the raw qubit count suggests.
- Genuine quantum advantage is rare. There are a few specific problems where today's quantum computers beat the best classical computers. For most real-world problems, classical still wins.
- Error-corrected fault-tolerant quantum computers are years away. Five to fifteen years is the usual honest estimate from researchers I trust.
In short: the field is advancing, but the marketing has been running ahead of the technology for a while.
What this means for you
If you run a non-quantum AI team today, here's my advice:
Don't panic about quantum. You don't need to be adding quantum to your roadmap this year. Anyone telling you that you do is probably selling something.
Do plan for post-quantum cryptography. This is the one quantum-adjacent thing that's actually urgent. If your organisation stores data that needs to remain confidential beyond about 2035, start the migration conversation now. NIST has standardised post-quantum algorithms; your security team knows what to do.
Pay attention to hybrid approaches. The most likely near-term wins involve hybrid classical-quantum algorithms, where a classical computer runs most of the workflow and hands the quantum hardware the specific bit it's good at. Keep an eye on this.
Follow the drug discovery and materials science applications. These are the most likely places to see real business value first. If you're in pharma or materials, there are serious groups worth watching.
A tempered optimism
Quantum computing is one of the most exciting frontiers in science. The fact that we're building computers that exploit the deep structure of physical reality — that's extraordinary, and we shouldn't lose sight of how wonderful it is.
But wonderful and practical are different things. The pattern with every major new technology has been: early hype (five years too soon), disillusionment (three years too pessimistic), then the quiet, steady payoff of the real thing. We're probably mid-disillusionment on quantum.
I'd love to be wrong about the timeline. I'd love to be writing a very different post in five years. In the meantime, don't let anyone rush you.
If you want to think through quantum's relevance to your roadmap — or the non-quantum AI you should actually be building this year — we're always happy to chat.
