| 'in years, not decades' | Microsoft Quantum | Fault-tolerant topological prototype (DARPA US2QC) | In progress | Microsoft is in DARPA's final stage and has provided a Majorana 2-based system for DARPA evaluation (September 2026); no prototype date given. Source |
| 'over the next three years' (November 2025); 2028-2029 (August 2026) | Xanadu | Fully fault-tolerant quantum computer | Unclear | The CEO said in August 2026 that loss reductions give confidence in reaching fault tolerance in 2028 to 2029; no independent evidence yet. Source |
| 2023 | IBM | Condor, 1,121-qubit processor | Delivered | IBM introduced the 1,121-qubit Condor at its Quantum Summit in December 2023. Source |
| 2023 | IonQ | Modular quantum computers small enough to be networked in a data centre | Unclear | No source opened confirms or refutes delivery of networked modular systems by 2023. Source |
| 2023 | IonQ | #AQ 29 (algorithmic qubits) | Delivered | IonQ reported reaching #AQ 29 seven months early in 2023 and opened Forte at that level to direct-access customers (company-reported metric). Source |
| 2024 | Atom Computing | 1,000+ qubit second-generation gate-based system available | Unclear | We found no source confirming general or customer availability of this system in 2024; the company only says it has developed systems with over 1,000 qubits. Source |
| 2024 | Pasqal | Ship a 100+ qubit neutral-atom computer to end users | Delivered | Pasqal's hardware page lists an Orion system deployed at Jülich (Jade) in November 2024, and a 140-qubit system was later delivered to CINECA, Italy. Source |
| 2024 | QuEra Computing | Gen-1 error-corrected system: 10 logical qubits, >256 physical, transversal gates | Slipped | QuEra's own system list dates its first gate-based machine (Gemini, 260 physical qubits) to 2025; the first unit was installed at AIST, Japan, in May 2025. Source |
| 2024 / 2026 | Rigetti Computing | Scale to 1,000 qubits by 2024 and 4,000 by 2026 (2021 plan) | Slipped | Rigetti's largest system as of October 2026 is the 108-qubit Cepheus-1-108Q, and its current ~1,000-qubit goal is set roughly three years out. Source |
| 2025 | IBM | Kookaburra as a 1,386-qubit module, three linked into a 4,158-qubit system (2022 roadmap version) | Slipped | In December 2023 IBM redefined Kookaburra as a coupler/error-correction demonstrator, and its June 2025 roadmap moved Kookaburra to 2026 as a qLDPC memory plus logic module; no 4,158-qubit system was delivered in 2025. Source |
| 2025 | IBM | Loon experimental chip with long-range c-couplers for qLDPC codes | Delivered | IBM announced Loon in November 2025 as an experimental processor demonstrating the key components for fault tolerance (company-reported; not available to users). Source |
| 2025 | IonQ | Broad quantum advantage | Slipped | In 2025 IonQ still described broad quantum advantage as a goal it was working toward; no claim of having achieved it was found. Source |
| 2025 | Microsoft Quantum | Commercial Atom Computing + Microsoft logical-qubit machine delivered | Slipped | Offered in November 2024 'with delivery in 2025'; the first such machine (Magne) is now slated for operation by early 2027. Source |
| 2025 | Pasqal | 2 logical qubits | Unclear | Pasqal's January 2026 blog says logical qubits were demonstrated on neutral-atom hardware in 2025 and shown live in early 2026, but gives no paper or detail confirming Pasqal's own 2-logical-qubit result. Source |
| 2025 | PsiQuantum | Commercially useful quantum computer (earlier target reported by press) | Slipped | A May 2024 analysis by UTS physicist Chris Ferrie says PsiQuantum had earlier planned for a 2025 deadline; no PsiQuantum machine was delivered in 2025. Source |
| 2025 | PsiQuantum | Break ground on Chicago Quantum Compute Center | Delivered | Construction began at the Illinois Quantum and Microelectronics Park on 30 September 2025, and phase-one construction was reported in November 2025. Source |
| 2025 | PsiQuantum | Break ground on Brisbane Quantum Compute Center | Slipped | Groundbreaking happened on 18 June 2026 at a new site, Moreton Bay Central, rather than in 2025. Source |
| 2025 | Quantinuum | Helios, about 100 physical qubits | Delivered | Helios launched commercially on 5 November 2025 with 98 physical qubits and named customers. Source |
| 2025 | QuEra Computing | Gen-2 system: 30 logical qubits, >3,000 physical, magic state distillation | Slipped | Logical magic state distillation was shown on Gemini (Nature, July 2025), but QuEra's current system list has no >3,000-physical-qubit product; its largest available machine is the 260-qubit Gemini. Source |
| 2026 | D-Wave | Initial gate-model system: 17-physical-qubit dual-rail system, logical error rate 2x below physical | Due this year | Restated as a 2026 delivery in the 6 August 2026 results; we found no announcement that it has been delivered or made generally available as of 27 October 2026. Source |
| 2026 | IBM | Kookaburra: first module with qLDPC memory plus a logical processing unit | Due this year | IBM's 2026 roadmap page still lists Kookaburra as planned and the August 2026 modular cryogenics release does not mention it; no delivery announcement found. Source |
| 2026 | IonQ | 256 physical qubits at 99.99% two-qubit fidelity | Slipped | Superion 256 was launched for order in September 2026 with customer deliveries in 2027, and IonQ's Q2 2026 results put 256-qubit commissioning in fiscal 2027. Source |
| 2026 | Pasqal | 10,000-qubit system with scalable logical-qubit architecture | Slipped | Pasqal's June 2025 roadmap moved the 10,000-qubit direction to around 2028, and its largest reported register in H1 2026 was 1,024 atoms. Source |
| 2026 | QuEra Computing | Gen-3 system: 100 logical qubits, >10,000 physical | Slipped | QuEra's current roadmap places its first >10,000-physical-qubit system (Libra) in 2028. Source |
| 2026 (implied: IBM said it was achieved a year ahead of schedule) | IBM | Real-time decoding of qLDPC codes | Delivered | IBM reported real-time error decoding in under 480 nanoseconds in November 2025 (company-reported). Source |
| 2026-2027 | Xanadu | Qubit Factory build | Unclear | Listed in the August 2026 roadmap; no progress report found. Source |
| 2027 | D-Wave | 49-physical-qubit dual-rail system, 20x error reduction | On track (company says) | Date unchanged in D-Wave's 6 August 2026 results release; future target, nothing delivered yet. Source |
| 2027 | IBM | Cockatoo: entanglement between modules via a universal adapter | Unclear | No recent IBM statement on Cockatoo's schedule was found. Source |
| 2027 | IonQ | Over 10,000 physical qubits, 800 logical qubits | Unclear | IonQ says Superion 10K is 'well underway' but its September 2026 release gives no delivery year; the roadmap page still shows 2027. Source |
| 2027 | Pasqal | 20 logical qubits | Unclear | Stated in the June 2025 roadmap; no later progress statement found. Source |
| 2027 | PsiQuantum | Chicago IQMP phase one (about 30 acres, incl. cryogenic plant) complete | Unclear | Reported completion target of 2027 for the park's first phase; no PsiQuantum date for a working machine there. Source |
| 2027 | Quantinuum | Sol: first 2D-grid trapped-ion system, up to 100 logical qubits | On track (company says) | In August 2026 Quantinuum said Sol remains on track for 2027 and its trap chip has returned from fabrication and is in validation. Source |
| 2027 (2025 roadmap); launch in 2026 (January 2026 blog) | Pasqal | Vela QPU (200+ / 256+ qubits) | Due this year | The January 2026 blog says Vela launches in 2026 with over 256 qubits; no source found confirming it has shipped as of October 2026. Source |
| 2027 (manufacturable commercial systems in 2028) | IonQ | Full fault tolerance in a lab setting | Unclear | Stated as an expectation in the Superion launch release; no progress report against it yet. Source |
| 2028 | D-Wave | 181-physical-qubit dual-rail system, 2,000x error reduction | On track (company says) | Date unchanged in the August 2026 results; May 2026 coverage described it as ~175 physical qubits. Source |
| 2028 | Pasqal | Centaurus, early fault-tolerant system (500+ qubits) | Unclear | Pasqal's hardware page lists Centaurus with 'targeted availability for users: 2028'; no progress evidence. Source |
| 2028 | QuEra Computing | Libra: megaquop-class fault-tolerant system, >256 logical qubits, 1e-6 logical error rate, on Amazon Braket | On track (company says) | Announced June 2026 and repeated on QuEra's site in October 2026; nothing to judge yet. Source |
| 2028-2029 (initial use at QuEra) | QuEra Computing | Gigaquop-class system: >1,000 logical qubits, >20,000 physical, 1e-9 logical error rate | On track (company says) | Announced June 2026; long-range target. Source |
| 2029 | D-Wave | Advantage3: intermediate 20,000-qubit multi-chip annealer | On track (company says) | Stated in the August 2026 results; long-range target. Source |
| 2029 | Google Quantum AI | Useful, error-corrected quantum computer | On track (company says) | In March 2026 Google said it is 'increasingly confident' that commercially relevant superconducting quantum computers will arrive by the end of this decade. Source |
| 2029 | IBM | Starling: fault-tolerant system, 200 logical qubits running 100 million gates | On track (company says) | IBM's August 2026 release on connecting its first two modular cryogenic systems describes this as a step toward delivering Starling in 2029. Source |
| 2029 | Microsoft Quantum | Scalable quantum computer | Unclear | Announced with Majorana 2, which Microsoft says 'cut our timeline in half'; it rests on qubits whose topological nature is still disputed. Source |
| 2029 | Pasqal | Lyra fault-tolerant system with 100 high-fidelity logical qubits | Unclear | Stated in the June 2025 roadmap; the 2026 roadmap page now cites 200+ logical qubits by 2029 instead. Source |
| 2029 | Quantinuum | Apollo: universal, fully fault-tolerant quantum computer with hundreds of logical qubits | On track (company says) | Quantinuum's August 2026 results say Apollo is on schedule for 2029, with key subsystems in prototyping. Source |
| 2029 (2026 roadmap page); 2030 (June 2025 roadmap) | Pasqal | 200+ logical qubits | Unclear | The current roadmap page states the goal of 200+ logical qubits by 2029, a year earlier than the 2025 roadmap; no progress evidence. Source |
| 2029 (November 2025); 2029-2030 (August 2026) | Xanadu | Quantum data centre (in Canada) | Slipped | The November 2025 target of 2029 became a 2029-2030 window in the August 2026 roadmap. Source |
| 2029 / 2030 / 2031 | Xanadu | Up to 200 logical qubits (2029), up to 500 (2030), 1,000+ (2031) | Unclear | Company roadmap targets presented August 2026; not yet testable. Source |
| 2029-2030 | Xanadu | Up to 100,000 physical qubits and up to 500 logical qubits | Unclear | Stated in the annual report, which also says only twelve logical qubits have been shown so far, with photon loss not at required levels. Source |
| 2030 | D-Wave | 10-logical-qubit gate-model system | On track (company says) | Listed in the August 2026 results release; long-range target, nothing to judge yet. Source |
| 2030 | IonQ | 2 million physical qubits, 80,000 logical qubits | Unclear | Still shown on IonQ's roadmap page; no recent statement that it is on schedule. Source |
| 2031 | D-Wave | 100,000-qubit annealing system | On track (company says) | Stated in the August 2026 results alongside a prototype I/O interface; long-range target. Source |
| 2032 | D-Wave | 100-logical-qubit gate-model system (>1 million operations) | On track (company says) | Listed in the August 2026 results and the August 2026 Nature-paper release; long-range target. Source |
| 2033 | IBM | Blue Jay: 2,000 logical qubits running 1 billion gates | Unclear | IBM's June 2026 release still names Blue Jay as Starling's successor but gives no year. Source |
| after the 100+ qubit system (no date) | Rigetti Computing | 336-qubit Lyra system | Unclear | Named in the November 2024 roadmap but not mentioned in any 2025-2026 Rigetti release we opened. Source |
| around 2023 or 2024 (as reported by TechCrunch; D-Wave's release gave no date) | D-Wave | Advantage2 annealer with 7,000+ qubits and 20-way connectivity | Slipped | Advantage2 reached general availability on 20 May 2025 with 4,400+ qubits (20-way connectivity), later and smaller than the 7,000+ first described. Source |
| around the turn of 2026/27 (July 2025); early 2027 (September 2026) | Microsoft Quantum | Magne: 50-logical-qubit Level 2 machine operational in Copenhagen | On track (company says) | Microsoft's 22 September 2026 blog still says Magne is slated for operation by early 2027; no evidence yet that it is running. Source |
| by or around end of 2026 | Rigetti Computing | 150+ qubit system at ~99.7% median two-qubit fidelity | Due this year | Not delivered as of October 2026, and not restated in the May or August 2026 results releases, which mention only the 108-qubit system and a ~3-year path to ~1,000 qubits. Source |
| by or around end of 2027 | Rigetti Computing | 1,000+ qubit system at ~99.8% median two-qubit fidelity | Slipped | The August 2026 results recast this as ~1,000 qubits at ~99.9% fidelity 'over roughly a three-year time horizon' (roughly 2029), later than end of 2027. Source |
| delivery in 2025 | Atom Computing | Commercial logical-qubit machine (Atom hardware plus Microsoft software) available to order, delivered to customers (joint with Microsoft) | Unclear | We found no report of a 2025 customer delivery; as of June 2026 the first named system, Magne, was still 'currently being installed'. Source |
| end of 2024 | Rigetti Computing | 84-qubit Ankaa-3 with 99+% median two-qubit fidelity | Delivered | Ankaa-3 launched on 23 Dec 2024 with a reported 99.0% median iSWAP and 99.5% median fSim two-qubit fidelity. Source |
| end of 2025 | IBM | Nighthawk (120 qubits, 5,000 two-qubit gates) delivered to users | Delivered | The first Nighthawk (ibm_miami) opened for early access to Premium and Flex customers in early January 2026, a few days after the stated end-of-2025 target. Source |
| end of 2025 | IonQ | #AQ 64 system | Delivered | IonQ reported #AQ 64 on a Tempo development system in September 2025 (IonQ's own benchmark; not independently verified). Source |
| end of 2025 | Pasqal | Orion Gamma, third Orion generation, 140+ physical qubits | Delivered | Orion Gamma (140+ qubits) is listed on Pasqal's hardware page, and a 140-qubit Orion system was inaugurated at CINECA in June 2026; reports do not confirm the exact Orion Gamma ship date, so it may have arrived after end-2025. Source |
| end of 2025 | Pasqal | 1,000 physical qubits (atoms) demonstrated | Delivered | Pasqal reported defect-free registers of 1,024 atoms in its H1 2026 results (company-reported, published September 2026), so this came in 2026, after the end-2025 target; it is an atom-array result, not a 1,000-qubit computing system. Source |
| end of 2025 | Rigetti Computing | 100+ qubit chiplet system at 99.5% median two-qubit fidelity | Slipped | In January 2026 Rigetti moved general availability to around end of Q1 2026; the 108-qubit system reached GA on 7 April 2026 at 99.1%, not 99.5%, median two-qubit fidelity. Source |
| end of 2026 | IBM | Nighthawk circuits with up to 7,500 gates | Delivered | IBM's August 2026 Nighthawk r2 post reports accurate observable estimation on circuits with more than 7,500 gates (company-reported). Source |
| end of 2026 | IBM | First verified quantum advantage confirmed by the wider community | Due this year | IBM and partners posted three advantage preprints in July 2026, but none is peer reviewed, the community tracker lists them as candidates, and an August 2026 preprint simulated the IBM-UChicago circuit classically in about 37 minutes; IBM's hardware page nonetheless says advantage was delivered. Source |
| end of 2027 | PsiQuantum | Brisbane utility-scale site operational | Slipped | Australian officials told Senate Estimates the project could be up to 12 months behind; the site moved from Brisbane Airport to Moreton Bay with groundbreaking only in June 2026, and the cryoplant is due in the second half of 2027 before site commissioning. PsiQuantum itself has not published a new date. Source |
| end of Q1 2026 (roadmap page); 'first part of 2026' (2025 roadmap); 'before mid-2026' (January 2026 blog) | Pasqal | Quantum advantage demonstration on an industry-relevant problem (250-qubit QPU) | Slipped | The target date passed without a published, independently checked advantage result; in September 2026 the CEO said Pasqal 'can deliver quantum advantage on meaningful applications today' but cited no benchmark. Source |
| later in 2026 | Rigetti Computing | Raise Cepheus-1-108Q median two-qubit fidelity to 99.5% | Due this year | The August 2026 results still report about 99.1% median two-qubit fidelity on the 108-qubit system and no longer mention the 99.5% target. Source |
| mid-2025 | Rigetti Computing | 36-qubit multi-chip system at 99.5% median two-qubit fidelity | Delivered | Cepheus-1-36Q reached general availability on Rigetti QCS in August 2025 with a reported 99.5% median two-qubit fidelity (company-reported). Source |
| no date given | Google Quantum AI | Milestone 3: long-lived logical qubit | In progress | Google's October 2025 post names the long-lived logical qubit as the next milestone; July 2026 Nature work improved logical-qubit stability but no completion has been announced. Source |
| no date given | Google Quantum AI | Neutral-atom programme: demonstrate deep circuits with many cycles | In progress | Announced March 2026 with a hardware team led by Adam Kaufman in Boulder; no results or dates published yet. Source |
| No date given | Atom Computing | Utility-scale fault-tolerant system design (DARPA Quantum Benchmarking Initiative, Stage C verification) | In progress | DARPA selected Atom Computing for QBI Stage C on 7 October 2026, in which government teams check whether its proposed design can be built and run as intended; no dates given. Source |
| No date given | Microsoft Quantum | Milestone 1: create and control Majoranas | Unclear | Microsoft's roadmap page marks it achieved (research published May 2023 in an APS journal), but the topological gap protocol behind it was challenged in a peer-reviewed Nature comment in June 2026. Source |
| No date given | Microsoft Quantum | Milestone 2: hardware-protected (topological) qubit | Unclear | Microsoft says it is achieving this with Majorana 1 and 2, but Nature's editors stated the February 2025 paper is not evidence of Majorana zero modes, and Majorana 2 data cover parity lifetime in one wire only. Source |
| No date given | PsiQuantum | Utility-scale, fault-tolerant (million-qubit) quantum computer | In progress | Current company materials (2026) describe the goal and the two sites but give no completion date. Source |
| No date given | Xanadu | Modular, networked photonic computer (Aurora) | Delivered | Aurora was described in a peer-reviewed Nature paper published 22 January 2025, which itself calls it a scale model and says optical loss remains the main hurdle to fault tolerance. Source |
| ready for first tasks by the new year 2026/27 | Atom Computing | Magne: Level 2 system with 50 logical qubits for QuNorth, Copenhagen (joint with Microsoft) | Due this year | Atom Computing said on 3 June 2026 that Magne is 'currently being installed'; we found no report that it is operating as of 27 October 2026. Source |
| second half of 2026 | Rigetti Computing | Publish an updated technology roadmap, including a path to quantum advantage in about three years | Due this year | Promised in the April 2026 GA release; we found no updated roadmap published as of 27 October 2026. Source |
| second half of 2027 | PsiQuantum | Brisbane cryoplant (Linde) delivered | Unclear | Stated in the June 2026 groundbreaking release; no later update found. Source |
| through fiscal 2026 | Xanadu | Device performance improvement and early fault-tolerance building blocks | Due this year | Q2 2026 results report an edge-coupling loss of 0.085 dB per facet, but the August 2026 roadmap still puts photon loss at 24.1x above the fault-tolerance threshold in 2026. Source |
| unclear | Google Quantum AI | Milestone 1: beyond-classical computation | Delivered | Google lists Milestone 1 as completed in 2019. Source |
| unclear | Google Quantum AI | Milestone 2: logical qubit prototype (more physical qubits reduce errors) | Delivered | Google lists Milestone 2 as completed in 2023. Source |
| unclear | Google Quantum AI | Error correction below the surface-code threshold (Willow) | Delivered | Published in Nature: the logical error rate roughly halved each time the code grew from 3x3 to 5x5 to 7x7 on Willow (announced December 2024). Source |
| within the next three to four years (stated May 2026) | Rigetti Computing | 1,000+ qubit system in the UK | In progress | Announced with up to $100M of UK investment; no specific year given. Source |
