Where the money went: quantum VC, 2021–2026, with the methodology caveats attached
The shape of the curve
2021 first-wave peak (Quantinuum $300M, PsiQuantum $450M)
2022 cooling begins; large EU rounds continue (IQM €128M)
2023 trough: ≈ −40% vs 2021 (MIT QIR estimate)
2024 $1.59B publicly announced (QIR count) — the selective year
2025 $4.9B private VC (TQI) — >2× prior record; US $2.7B
2026 China Q1 ≈ its full 2025; Quantonation II: €220M
Two counts, two methodologies: MIT’s Quantum Index Report tallies publicly announced rounds and warns of reporting gaps; The Quantum Insider’s totals run higher. We show both because the honest headline is the trend, not any single number: a boom, a two-year correction, then a record-breaking re-acceleration.
What the 2025 wave actually bought
The anchor was PsiQuantum’s $1B Series E at a $7B valuation (September 2025) — the largest single private round in the field’s history, on top of Australia’s earlier ~$620M public commitment in equity, grants and loans. The pattern across the top rounds is consistent: photonic and superconducting hardware, fault-tolerance roadmaps, and national-strategy co-investment. Governments are not a side note — public money increasingly arrives blended into private rounds.
Geographic concentration is loosening: 2025 was US-dominated ($2.7B of $4.9B), but China’s Q1 2026 nearly matching its full 2025 total suggests the capital map of 2026–2027 will look different from 2024’s.
The software sliver
Every source agrees software’s share is a fraction of hardware’s; they disagree on the fraction. QIR’s 2024 count logs $621M for quantum software against $1.59B for quantum computing rounds under its taxonomy; industry surveys like QED-C have put hardware’s share of total investment near 87%. Taxonomies differ (is a full-stack company hardware or software?), so we publish the range rather than pick a flattering point inside it. Under any count, the layer that determines whether the machines are worth using — algorithms, verification, benchmarking — receives a small minority of the capital building the machines themselves.
The reading, without narrative
- Capital has decided the hardware will exist. $4.9B in one year is not a bet on a lab curiosity.
- Almost nobody is paid to ask whether the algorithms beat classical baselines on real instances. That verification gap is not a moral failing; it’s an unfilled market position — the one our ledger occupies.
- Record funding raises the cost of hype: the more capital in, the more expensive every unverified claim becomes for whoever believes it.
What we don’t know
Consolidated 2026 full-year figures (the year is running). How much announced government money converts to deployed capital, and on what schedule. And whether the 2025 re-acceleration is durable or a rate-environment artifact — a question the 2026–2027 data will answer, and we’ll update this map when it does.
Educational content, dated 2026-07-23. Figures as reported by the cited sources; methodologies differ and are flagged where they do.