Quantum Computing

Quantum computing has moved from theoretical to early commercial deployment in 2020-2025. Major players (IBM, Google, Quantinuum, IonQ, Rigetti, PsiQuantum, Atom Computing) are racing to scale qubit counts and error rates. Cryptography implications are pressing — RSA-breaking quantum capability could arrive within 10-20 years, forcing transition to post-quantum cryptography.

1,121
IBM Condor processor qubit count
~1M
Noisy physical qubits estimated to break RSA-2048 (Gidney, 2025)
$30B+
Public + private quantum computing investment to date
2024
First NIST post-quantum cryptography standards

Key insights

⚛️

Qubit count is the headline race, fidelity is the real metric

Headline qubit counts have grown rapidly — IBM 1,121 (Condor), Atom Computing 1,180, Google's Willow 105 with major error-correction breakthrough. But error rates limit useful computation. Today's processors are 'noisy intermediate-scale quantum' (NISQ) — too noisy for general-purpose quantum applications. Logical qubits (error-corrected combinations of physical qubits) are the metric that matters; estimated 1,000-10,000 physical qubits needed per logical qubit.

🔐

Cryptography stakes are existential

Shor's algorithm — discovered 1994 — would break RSA encryption and elliptic-curve cryptography in polynomial time on a sufficiently large quantum computer. Estimates for RSA-2048 keep falling: older work pointed to about 20 million noisy qubits, and Craig Gidney's 2025 paper put it under 1 million, running for under a week. Today's machines are still around a thousand times too small, and their error rates are not yet good enough. But the gap is narrowing. NIST released first post-quantum cryptography standards in August 2024 (ML-KEM, ML-DSA, SLH-DSA). Migration to post-quantum is now underway across major systems.

💼

Commercial applications are emerging but narrow

Quantum advantage demonstrated in specific narrow problems (Google Sycamore 2019; subsequent benchmarks contested). Useful commercial applications: optimization (logistics, finance), molecular simulation (drug discovery, materials science), machine learning components. None has produced clear 'kill app' that exceeds classical performance at meaningful cost. The 'quantum winter' concern — that commercial value lags hype — is being discussed but not yet materializing as serious.

Quantum processor qubit count — leading systems

Number of qubits in flagship processors

Show data table
Quantum processor qubit count — leading systems
Qubits
IBM Condor (2023)1121
IBM Heron R2 (2024)156
Atom Computing1180
QuEra Aquila256
Quantinuum H256
Google Willow105
IonQ Forte36
Rigetti Ankaa-384

Key Finding: Qubit count has grown ~10× in 5 years. But fidelity is the real bottleneck.

Quantum computing investment 2015–2024

USD billions in public + private quantum tech investment

Show data table
Quantum computing investment 2015–2024
USD billions
20151
20184
202015
202225
2024 (cumul)34

Key Finding: Investment surged ~2020-2024 with several very well-funded companies (PsiQuantum, IonQ).

Methodology & caveats

Qubit types

Superconducting (IBM, Google): cryogenic circuits. Trapped ion (IonQ, Quantinuum): individual atoms suspended in EM fields. Photonic (PsiQuantum, Xanadu): light particles. Neutral atom (QuEra, Atom Computing): cold atoms in optical traps. Each technology has different scaling characteristics, error profiles, connectivity. None is yet clearly dominant; the race is ongoing.

Error correction

Quantum error correction creates 'logical qubits' that are robust against individual physical-qubit errors. Surface codes are the leading approach; require ~1,000-10,000 physical qubits per logical qubit. Google's Willow (December 2024) showed first below-threshold scaling — error rates dropping as code distance increases. This is the precondition for practically useful fault-tolerant quantum computing.

Timeline uncertainty

Predictions for cryptographically-relevant quantum computing range from 5-30 years. NIST/NSA recommended migration to post-quantum cryptography by 2035, treating mid-2030s as plausible timeline. 'Harvest now, decrypt later' attacks already collect encrypted traffic for future decryption — making post-quantum migration urgent for high-value long-lived secrets.

Why qubit counts mislead

A processor with more qubits is not necessarily better. Google's 105-qubit Willow beat IBM's 1,121-qubit Condor on the measure that counts, error correction, and IBM itself moved to smaller, higher-quality chips such as the 156-qubit Heron. Technologies are also hard to compare: trapped ions are slower but connect every qubit to every other, while superconducting chips are fast but sparsely linked. Claims of "quantum advantage" should be read with the classical algorithm they beat, because classical methods have repeatedly caught up.

What would change the picture

The clearest milestone would be a sustained demonstration of many logical qubits running useful algorithms with low error rates. Resource estimates matter just as much. Gidney's 2025 paper cut the qubits estimated to break RSA-2048 by about 20 times from his 2019 estimate, without new hardware. So the timeline can shorten through better algorithms as well as bigger machines, which is why migration to post-quantum cryptography cannot wait for a working attack.

Cite this page

GlobeChart. “Quantum Computing.” Last updated 2026-10-31. https://globechart.com/technology/quantum-computing/

Please also credit the original publisher named under each chart: Company announcements, IBM Quantum Network; McKinsey Quantum Technology Monitor 2024, BCG. Spotted an error? Tell us.