A quantum computer computes with qubits — elements that obey the laws of quantum physics. It will not replace a laptop or speed up everything, but for particular problems — modeling molecules and materials, some optimization and cryptography problems — it promises what ordinary supercomputers cannot reach. For business today, this means access through manufacturers’ clouds, experiments and getting encryption ready for the future.
How it works
An ordinary bit is either 0 or 1. Before measurement, a qubit is in a superposition — a combination of 0 and 1 with particular weights, called amplitudes. Several qubits can be entangled: their states become linked so that one can no longer be described without the other. With each new qubit, the number of states the machine describes doubles: 50 qubits have more than a quadrillion. An ordinary computer needs enormous memory just to write down such a state.
But a quantum computer does not try all the answers at once. Its strength is interference, like waves on water: the algorithm is built so that the paths to wrong answers cancel each other out, while the paths to the right one add up. On measurement the superposition disappears, and the desired result comes out with high probability. That is why only problems with a special structure gain anything: Google Willow ran a verifiable algorithm 13,000 times faster than a supercomputer.
The main difficulty is fragility. Superconducting qubits live in refrigerators at hundredths of a degree above absolute zero, and any heat, vibration or stray field knocks their state off — errors pile up. The solution is error correction: many physical qubits are combined into one “logical” qubit, and faults are constantly checked and fixed. IBM promises 200 logical qubits by 2029 — a step toward machines that compute for a long time without errors piling up.
Where it already works
- Cloud access to quantum processors — for experiments and team training.
- Research in chemistry and materials: modeling molecules, which is hard for ordinary computers.
- Google Willow: a verifiable algorithm 13,000 times faster than a supercomputer.
- Data protection: the move to post-quantum encryption starts ahead of time.
What to consider
- For ordinary business tasks there is no practical gain yet — first check whether the problem can be solved classically.
- Timelines are manufacturers’ roadmaps: IBM promises 200 logical qubits by 2029.
- The number of qubits by itself means little: what matters is the error rate and logical qubits.
- Data that must stay secret for years should be prepared for post-quantum encryption now.