Hype versus value
Lesson 14 of 23 in the course “Quantum”. Free, with animations and a quiz.
When quantum computers become useful
The honest answer: piece by piece and gradually. Right now, in 2026, chips are noisy, and usefulness is shown on narrow scientific problems. IBM expects the first confirmed advantages by the end of 2026. For 2029, IBM and Quantinuum promise error-corrected machines: IBM’s Starling with 200 logical qubits and Quantinuum’s Apollo. And by 2033 DARPA wants to check whether such machines deliver more value than they cost. Remember: these are company plans, not guarantees.
How to tell a real advantage
After the Sycamore story the industry changed the rules. IBM, with Algorithmiq, the Flatiron Institute and BlueQubit, launched an open quantum advantage tracker. Quantum teams publish results, and classical-algorithm experts try to catch up. If they can’t, the advantage counts. It’s like a bug bounty, but for science. See a big headline? Check whether it survived that kind of attack.
Governments are checking too
DARPA, the agency that once stood at the origins of the internet, launched the Quantum Benchmarking Initiative. Its question is simple and tough: can any approach deliver a quantum computer whose value exceeds its cost by 2033? In November 2025, 11 companies made it to the second stage, including IBM, Quantinuum, IonQ, QuEra and Xanadu. Independent experts take their plans apart bolt by bolt. For you it’s a handy filter: being in QBI is a sign the technology is taken seriously.
“Quantum AI”: handle with care
Two buzzwords together sell great, but there’s a catch. Neural networks learn from terabytes of data, and for a quantum computer to work with them, the data first has to be loaded into qubits. That loading is slow in itself and often eats the whole gain. There’s no proven big speedup for ordinary AI tasks, and ChatGPT and its rivals run on GPUs. Quantum machine learning is a lively research area, not a finished product.
Where the real money is today
The biggest quantum business right now is, funnily enough, protection from quantum computers. Banks, government agencies and cloud providers are moving to post-quantum cryptography, and that needs people and budgets. The second area is quantum sensors and clocks: navigation without GPS, medical diagnostics, mineral exploration. The third is research and cloud access to chips for pharma, chemistry and finance. If you’re looking for a career, start with these three.
Red flags
The word “quantum” sounds smart, so scammers love it. Ads for a “Quantum AI” platform that supposedly trades crypto for you often use fake celebrity videos. It’s a known scheme, and financial regulators warn about schemes like it. Add “quantum” tokens promising to moon and bracelets “for quantum energy.” The rule is simple: real quantum companies sell chip access, software and services to businesses, not guaranteed income to you.
Check yourself
- How do you check a big “quantum advantage” headline?
Correct answer: See whether classical algorithms tried and failed to beat it. An advantage counts when classical-algorithm experts tried to match the result and couldn’t. That’s what the tracker is for. - An ad: “Quantum AI trades crypto, +300% a month, recommended by a famous billionaire.” What do you do?
Correct answer: Scroll past: it’s a typical scam pattern. Guaranteed income, a celebrity in the ad and the buzzword “quantum”: three red flags at once. Real quantum companies don’t work like this. - Where can you work in quantum right now without being a physicist?
Correct answer: Companies moving to post-quantum cryptography. Banks and governments already need people to inventory encryption and move to the new standards. IT and security people are needed, not just physicists.