Forget the qubit-count press releases you saw in isolation. This study puts every major quantum computing milestone of 2026 in one place: eight of the largest announced systems, the logical-qubit records that actually matter for fault tolerance, at least $4 billion in fresh capital, and a national roadmap that finally gives the industry a deadline. Every figure below is traced to a company announcement, a named publication, or an industry compilation. Claims we could not verify did not make the cut.
The headline is a shift in what counts. For a decade, the quantum race was scored in physical qubits. In 2026, the scoring changed: the year's most important number is 94 logical qubits — protected, error-corrected information — demonstrated by Quantinuum on a 98-qubit machine. And the year's most important event was not a chip at all: it was a $1.68 billion IPO that turned quantum computing into a public-market sector.
Key findings
- Two systems crossed 250 physical qubits in 2026: IonQ's 256-qubit Superion 256 (launched September 8) and Pasqal's 256+-qubit Vela neutral-atom QPU — the largest announced machines of the year.
- The logical-qubit race went past break-even: Quantinuum's 98-qubit Helios demonstrated 94 logical qubits under error detection and 48 under full error correction — roughly one physical qubit per logical one, a fraction of the overhead of traditional surface codes.
- Quantum went public in 2026. Quantinuum listed on Nasdaq as QNT in June, pricing 28.5 million shares at $60 to raise $1.68 billion at a ~$15 billion valuation — the largest pure-play quantum listing to date. Xanadu, Infleqtion and IQM also completed listings.
- At least $4 billion in fresh quantum capital in 2026: PsiQuantum's $1.5B raise (May), Quantinuum's $1.68B IPO (June), Pasqal's €340M package, Oratomic's $300M Series A, and OQC's £260M — plus IonQ's closed $1.8B acquisition of SkyWater Technology.
- Revenue is real now. IonQ reported Q2 2026 revenue of $80.1 million, up 287% year on year, and raised full-year guidance to $450–460 million after closing the SkyWater deal.
- Governments drew a line in the sand. The U.S. Department of Energy's October 2026 national roadmap targets a scientifically relevant, error-corrected quantum computer by 2028. IBM says Starling (200 logical qubits, 100 million operations) arrives in 2029.
Journalists and researchers are free to quote these findings with attribution to Calder Brief. The full dataset is licensed CC-BY.
The hardware scoreboard
These are the largest quantum systems publicly announced as of October 16, 2026, across the four leading modalities. Note the caveat that makes the chart honest: neutral-atom systems count trapped atoms, not gate-programmable qubits, and IonQ's and IBM's numbers describe different kinds of hardware. The chart shows the scale each platform is reaching, not a like-for-like ranking.
Physical qubits (or trapped atoms for neutral-atom systems) of the largest publicly announced systems as of Oct 16, 2026. Neutral-atom entries are not directly comparable to gate-programmable qubit counts. Sources: company disclosures, press releases, and industry compilations; see Sources.
| # | System | Company | Modality | Physical | Announced | Note | Source |
|---|---|---|---|---|---|---|---|
| 1 | Superion 256 | IonQ | Trapped ion | 256 | Sep 8, 2026 | Sixth-gen platform; control chips on standard semiconductor lines (SkyWater); designed for volume production; first unit pre-sold in Q1; server-rack footprint | SiliconANGLE |
| 2 | Vela QPU | Pasqal | Neutral atom | 256+ | H1 2026 target | Targeting quantum advantage demonstration H1 2026; 140-qubit system delivered to Italy's CINECA | Industry compilation |
| 3 | Phoenix | Atom Computing | Neutral atom | 1,180 atoms | 2025–26 | Trapped-atom array count, not gate-programmable qubits; largest reported neutral-atom system | MomoView / QI-QCR |
| 4 | Nighthawk | IBM | Superconducting | 120 | Cloud access Jan 13, 2026 | 218 tunable couplers; runs circuits of up to 5,000 two-qubit gates; later iterations target 7,500 gates by end 2026 | The Quantum Insider |
| 5 | Cepheus-1-108Q | Rigetti | Superconducting | 108 | GA expected end Q1 2026 | Twelve 9-qubit chiplets — the industry's largest modular quantum system; 99% median two-qubit gate fidelity on the 108-qubit system | GlobeNewswire / Rigetti |
| 6 | Willow | Superconducting | 105 | Dec 2024 | First below-threshold surface-code demonstration — errors fell as the code scaled | Industry compilation | |
| 7 | Helios | Quantinuum | Trapped ion | 98 | Mar 2026 (demo reported) | 94 logical qubits under error detection, 48 under full error correction; ~1:1 physical-to-logical ratio via Iceberg codes; first international deployment announced (Singapore, Mar 2026) | Industry compilation |
| 8 | Emerald | IQM | Superconducting | 54 | 2025 | Completed public listing in 2026 | Industry compilation |
Tracked total: 8 leading systems across superconducting, trapped-ion, and neutral-atom modalities. IBM's Nighthawk was unveiled in November 2025 and entered cloud availability on January 13, 2026; it counts because 2026 was its deployment year.
The logical-qubit race
Physical qubits are the vanity metric. Logical qubits — protected from noise by error correction — are the metric that decides whether a quantum computer can ever run useful programs. The year's standout result belongs to Quantinuum: 94 logical qubits under error detection and 48 under full error correction on its 98-qubit Helios machine, reported in March 2026. With Iceberg codes running at roughly one physical qubit per logical one, the overhead that has haunted the field for decades is starting to collapse.
Demonstrated logical qubits (entangled where noted). Error-detection logical qubits cannot correct faults, so they are not equivalent to full error-correction logical qubits. Separately, IonQ demonstrated in September 2026 an end-to-end real-time error-correction decoder running on a standard off-the-shelf CPU, tested on benchmark circuits simulating up to 408 logical qubits and 31.5 million operations with ~0.02% added stretch time. Sources: industry compilations (2024–26); IonQ via Dollar Stock Club, Sep 2026.
The field's direction is clear: error-correction breakthroughs are where the momentum is. IBM unveiled its Loon processor as the default path to its fault-tolerant Starling system, Classiq launched a Fault Tolerance Engine that translates logical programs into physical resource plans (September 2026), and IBM plus AMD published real-time classical decoding of errors using qLDPC codes — delivered a year ahead of schedule.
Quantum's capital-market year
2026 is the year quantum computing graduated from venture-funded research to a public-market sector. Quantinuum's June IPO — 28.5 million shares at $60, raising about $1.68 billion at a valuation near $15 billion — was the largest pure-play quantum listing to date and priced above its range on heavy demand. Xanadu, Infleqtion and IQM also completed listings, so the leading hardware companies now report quarterly. Private money kept flowing too: PsiQuantum closed $1.5 billion in May at a $10.5 billion valuation.
Fresh capital in 2026, USD. Euro and pound amounts converted at approximate 2026 rates. Excludes IonQ's $1.8B SkyWater acquisition (closed July 31, 2026), which was a purchase, not a raise. Sources: Sacra/PsiQuantum filings, company releases, industry reports; see Sources.
| # | Company | Event | Amount | Date | Detail | Source |
|---|---|---|---|---|---|---|
| 1 | Quantinuum | IPO (Nasdaq: QNT) | $1.68B | Jun 2026 | 28.5M shares at $60; ~$15B valuation; traditional underwritten IPO, not a SPAC | ainvest / QZ |
| 2 | PsiQuantum | Funding round | $1.5B | May 2026 | $10.5B valuation; prior round $1B at $7B in Sep 2025 | Sacra research |
| 3 | Pasqal | €340M financing package | ~$0.4B | 2026 | €170M equity (Parkway Ventures, Quanta, LG, CMA CGM, Temasek, Aramco's arm) + €170M convertible; dual Paris–Nasdaq listing planned via Bleichroeder Acquisition Corp. II | walaw.press |
| 4 | Oratomic | Series A | $300M | 2026 | Thesis: utility-scale QC may need ~10,000 neutral-atom qubits, not millions | Global Quantum Technologies Market report |
| 5 | OQC | Funding round | £260M (~$0.35B) | 2026 | Europe's largest private quantum computing round | Global Quantum Technologies Market report |
| 6 | IonQ | Acquisition (SkyWater) | $1.8B | Closed Jul 31, 2026 | Vertically integrated quantum platform, fab through cloud; six Superion tapeouts in H1 2026; design cycle cut from 9 months to 2 | SiliconANGLE |
Tracked total: at least $4.1 billion in fresh raises and listings (converted at approximate rates), plus the $1.8B IonQ–SkyWater acquisition. Xanadu, Infleqtion and IQM completed listings in 2026 without disclosed quantum-specific totals included here.
The most underreported number of the year may be IonQ's revenue. Q2 2026 revenue hit $80.1 million — up 287% year on year — and the company raised full-year guidance to $450–460 million, ahead of its September 8 Investor Day. Quantum computers are now a product category with quarterly results, not just a physics project.
2026 in one timeline
Who promises what, by when
Roadmaps are not results, but they are data: they show where the money and engineering are going. Here is what the major players publicly promise next.
| Target | Who | Promised |
|---|---|---|
| First processor storing information in efficient error-correcting (qLDPC) memory | IBM | 2026 |
| 100 logical qubits (neutral atoms) | QuEra | 2026, reported on track |
| 256 qubits at 99.99% two-qubit fidelity | IonQ | 2026 roadmap |
| 1,000+ programmable qubits via multi-chip modular cryogenic systems | IBM | 2027 |
| Quantum advantage demo (neutral atoms) | Pasqal | H1 2026 target |
| Scientifically relevant, error-corrected quantum computer | U.S. DOE national roadmap | 2028 |
| Starling: 200 logical qubits, 100 million operations | IBM | 2029 |
| Superion 10K: CMOS-on-chip generation for fault tolerance | IonQ | In development |
| Universal fault-tolerant quantum computing | Quantinuum | 2030 |
| 2M physical / 80,000 logical qubits | IonQ | 2030 roadmap |
One telling pattern: 2026's roadmaps stopped promising bigger machines and started promising protected computation. The qubit-count race is ending; the fault-tolerance race has begun.
Methodology
Scope. The dataset covers hardware announcements, logical-qubit demonstrations, funding rounds, listings and acquisitions, government programs, and published roadmaps from January 1 through October 16, 2026. Two entries (Google Willow, QuEra specs) were announced before 2026 but remain the relevant baselines for the 2026 scoreboard; IBM's Nighthawk was unveiled in November 2025 but entered cloud availability on January 13, 2026.
Inclusion rules. Every hardware entry is traced to a company announcement, press release, or named publication. Funding entries require a disclosed amount and date. Rumored or unconfirmed rounds were excluded. "Industry compilation" rows draw on The Quantum Insider / Quantum Computing Report data compiled in open industry research notes; those rows are marked as such rather than as primary announcements.
Comparability. Neutral-atom systems report trapped-atom counts; these are not directly comparable to gate-programmable qubit counts, and the page labels them as atoms where appropriate. Error-detection logical qubits are not equivalent to error-correction logical qubits; both are labeled. IonQ's September 2026 decoder result was demonstrated on simulated circuits, which the page states explicitly.
Corrections policy. If a figure here is wrong or outdated, we will correct the dataset and note the change in this section with a date.
Sources
Company announcements and releases: IonQ (Superion 256 launch, Sep 8, 2026; NVIDIA NVAQC partnership; Q2 2026 revenue; SkyWater acquisition), Rigetti (Cepheus-1-108Q roadmap update, Jan 9, 2026, via GlobeNewswire), IBM Newsroom (modular cryogenic systems milestone, Aug 19, 2026), Quantinuum (QNT IPO via ainvest.com and Quantum Zeitgeist), Pasqal (€340M package via walaw.press), PsiQuantum (Sacra research report). Press: SiliconANGLE (Superion 256 launch), The Quantum Insider (Nighthawk cloud availability, Jan 13, 2026), BusinessWire (IonQ–NVIDIA), Dollar Stock Club (IonQ decoder demo), The Motley Fool (IonQ roadmaps), Quantum Computing Report (DOE national roadmap, Oct 2, 2026), dev.to/The Daily Flare (roadmap summary). Industry compilations: MomoView quantum deep dive (sourced from company disclosures and The Quantum Insider / Quantum Computing Report, 2024–2026); the lynnlangit/learning-quantum and jonboy648/quantum-worldline open research notes. Market data: "The Global Quantum Technologies Market 2026–2036" report summary (giiresearch.com).