Quantum computing, without the hype.

Quantum computers exist, you can rent time on one this afternoon, and for almost every business problem you have they are currently the wrong tool. Those three statements are all true simultaneously, and holding them together is most of what it takes to think clearly about this field.

What a quantum computer does is not “the same thing, faster”. It is a different computational model, which for a small number of specific mathematical structures admits algorithms with no efficient classical equivalent. Factoring large integers is the famous one. Simulating quantum systems — molecules, materials, chemistry — is the one most likely to matter commercially first, because it is the case where the machine and the problem share a structure. Outside that narrow set, a classical computer is faster, cheaper and available.

The machines available today are described in the literature as NISQ — noisy intermediate-scale quantum. Noisy is the operative word. Qubits lose their state within fractions of a second, gates have error rates that compound through a circuit, and the number of operations you can perform before the result becomes noise is small. Error correction is the field's central engineering effort and it is progressing genuinely and unevenly; it is not solved.

So there are exactly two things we think a business should do about quantum computing in 2026. One of them has a deadline, and it is not the one anybody markets.

Quantum machine learning is a real research field with real results and, as of now, no demonstrated advantage over classical methods on the kind of data businesses actually hold — customer records, images, text, sensor streams. That absence, after more than a decade of concerted effort and substantial funding, is the single most informative fact about the field, and it is the one most often omitted.

The results that are genuinely promising involve quantum data — measurements from physical quantum systems — rather than classical data loaded into a quantum computer. That distinction matters enormously, because loading classical data into a quantum state is itself expensive enough to erase the speedup in most proposed schemes. It is a real and underappreciated bottleneck rather than an engineering detail awaiting a fix.

None of which means the field is empty. It means the honest description of the AI overlap in 2026 is: an active research area, worth watching, with no production application for a business whose problem is a language model or a forecasting task. If your machine learning problem is hard, the answer is better data or a better classical method, not a quantum computer. We would rather tell you that than sell you a pilot.