The main limiter of quantum computing is not technological competition, but fundamental laws of physics. Qubits are extremely sensitive to the external environment: any thermal fluctuation, vibration, or electromagnetic interference can destroy the fragile quantum state required for computation. This process, known as decoherence, is the key problem all developers face.
Physical and logical qubits: what's the difference
A physical qubit is a real quantum system: a superconducting circuit, a trapped ion, or a photon. Due to environmental influence, it "noises": for superconducting qubits, the error rate of two-qubit operations typically ranges from 0.1–1%. For comparison, market leaders have already demonstrated 99.9% accuracy. Quantinuum, on its H1-1 ion platform, showed 99.914% accuracy, while its flagship 98-qubit processor Helios achieved 99.921%. In superconducting systems, similar results have been achieved by IQM (CZ-gate accuracy of 99.91%) and IBM with its Egret and Heron processor lines.
A logical qubit is a protected unit of information where the quantum state is distributed across many physical qubits through quantum error correction (QEC). The most studied approach is the surface code, which ensures locality of operations and a high theoretical error-tolerance threshold (around 1%).
The past two years have brought significant breakthroughs in this area:
- Google Quantum AI's Willow processor demonstrated exponential suppression of logical errors: as the lattice of physical qubits grows, the error rate more than halves at each scaling step.
- Quantinuum's Helios, using barium-137 ions, demonstrated 48 logical qubits with error correction at a ratio of roughly two physical qubits per logical qubit.
- Researchers from Harvard, MIT, and QuEra used a 448-neutral-atom processor, executing circuits with dozens of logical qubits, and in certain configurations—up to 96.
Previously, it was believed that reliability required redundancy of up to ~1000 physical qubits per logical qubit. However, high-speed qLDPC codes could potentially reduce these costs by an order of magnitude—down to dozens of physical qubits per logical qubit.
What this means in practice
Mass-scale universal fault-tolerant quantum computers still remain only on roadmaps. IBM plans to build the Starling system with 200 logical qubits by 2029, capable of executing 100 million operations. But even if successfully implemented, it will be a highly specialized coprocessor in the cloud alongside a classical supercomputer, not a replacement for traditional data centers.
It's important to understand: the number of physical qubits alone says little. What matters is the quality of operations, the rate of error occurrence, and, above all, the number of logical qubits the system can sustain with a sufficiently low error probability.
Q-check: where quantum advantage is real, and where it isn't
Will a quantum computer replace an ordinary laptop?
No: a quantum computer operates on a different principle, handles a narrow class of tasks, and requires cryogenic infrastructure with temperatures of 10–20 mK.
Where is the advantage formally proven?
On specialized benchmark tasks—boson sampling and generation of random quantum states. Separately, there are algorithms with theoretically proven speedups, such as Shor's algorithm for factorization.
Where is the advantage expected?
In simulating nature: pharmaceuticals (precise molecular modeling), materials science (searching for superconductors and new batteries), chemistry (creating catalysts). Optimization and finance are actively researched, but practical advantage has not yet been proven.
Where is there no advantage?
In everyday tasks—working with text, storing files, running interfaces and games. Classical processors remain more efficient and cheaper.
More qubits—more benefit?
No. A large processor is useless for long algorithms if errors accumulate faster than the system can correct them.
What's next
If the technology is so error-sensitive and requires complex infrastructure, a logical question arises: why do corporations and governments continue to invest billions in it? In the next issue, I'll break down the industry's investment logic—where the line lies between a bet on the future and a bubble.
My view: despite impressive demonstrations, commercially significant quantum computing is still years away. Investors should distinguish real technological breakthroughs from marketing claims—the key metric here will not be the number of qubits, but the stability of logical qubit operation as scale increases.