A History of Quantum ComputingInformation Is Physical, 1871–1982Quanta, 1900–1916The New Mechanics, 1922–1927Simulating Nature, 1926–1985Spooky Action, 1932–1964Secrets in Light, 1968–2017The Test, 1969–2022The Algorithm, 1977–1996Superconducting Circuits, 1984–2016Fixing Errors, 1994–2000First Qubits, 1994–2015Supremacy, 2011–2023Below Threshold, 2023–2025Where It Stands, 2026Clerk with pencil and paper (additions/s) (1880): 0.07Millionaire mechanical calculator (additions/s) (1895): 0.5Adding tabulator, punched cards (additions/s) (1907): 2.5Zuse Z3: 0.8 s per addition (1941): 1.25Harvard Mark I: three additions a second (1944): 3ENIAC: 5,000 additions a second (1946): 5,000CDC 6600: about 3 megaflops (1964): 3×10⁶CDC 7600: about 10 megaflops (1969): 10⁷Cray-1: 160 megaflops peak (1976): 1.6×10⁸Cray-2: 1.9 gigaflops peak (1985): 1.9×10⁹CM-5/1024, Los Alamos (1993): 5.97×10¹⁰Numerical Wind Tunnel, Tokyo (1995): 1.7×10¹¹ASCI Red, Sandia (1997): 1.338×10¹²ASCI White, Livermore (2001): 7.226×10¹²Earth Simulator, Yokohama (2002): 3.586×10¹³Blue Gene/L (2004): 7.072×10¹³Roadrunner, Los Alamos (2008): 1.026×10¹⁵K computer, Kobe (2011): 1.051×10¹⁶Sunway TaihuLight, Wuxi (2016): 9.301×10¹⁶Summit, Oak Ridge (2019): 1.486×10¹⁷Frontier, Oak Ridge (2022): 1.102×10¹⁸Frontier, Oak Ridge (2023): 1.194×10¹⁸El Capitan, Livermore (2024): 1.742×10¹⁸LineShine, Shenzhen (2026): 2.1984×10¹⁸Geneva: 23 km under Lake Geneva (1995): 23Geneva–Lausanne: 67 km (2002): 67Toshiba Cambridge: 122 km (2004): 122Los Alamos/NIST: 148.7 km (2007): 148.7Geneva/Corning: 307 km (2015): 307China/Corning: 404 km (too slow to be practical) (2016): 404Geneva/ID Quantique/Corning: 421 km (2018): 421USTC/Tsinghua: 509 km (twin-field) (2020): 509China: 833.8 km (twin-field) (2022): 833.8USTC: 1,002 km (twin-field) (2023): 1,002IBM: 32.5 cm across a lab bench (1989): 3.25×10⁻⁴Munich/QinetiQ: 23.4 km between two mountains (2002): 23.4La Palma–Tenerife: 144 km (2007): 144Micius satellite to ground: up to 1,200 km (2017): 1,200China–Austria via Micius (trusted relay) (2018): 7,600Beijing–Stellenbosch via Jinan-1 (trusted relay) (2025): 12,900NIST CNOT: two qubits in one beryllium ion (1995): 2NIST: four entangled ions (2000): 4Innsbruck: 14 entangled calcium ions (2011): 14Maryland/NIST: 53-ion quantum simulator (2017): 53Quantinuum H2: 56 qubits (2024): 56Quantinuum Helios: 98 qubits (2025): 98NEC: two coupled charge qubits (2003): 2Google/UCSB: nine-qubit error-detection chip (2015): 9IBM: five-qubit processor on the cloud (2016): 5IBM Q 20 Tokyo (2017): 20Google Bristlecone (announced; no published results) (2018): 72Google Sycamore (53 of 54 qubits used) (2019): 53IBM Eagle (2021): 127IBM Osprey (2022): 433IBM Condor (2023): 1,121Google Willow (2024): 105USTC Zuchongzhi 3.0 (2025): 105IBM Nighthawk r2 (2026): 120Harvard/MIT: 51-atom simulator (2017): 51Harvard/QuEra: 256-atom simulator (2021): 256Atom Computing: 1,180 atoms (no gate data published) (2023): 1,180Harvard/QuEra logical processor: up to 280 atoms (2023): 280Harvard/MIT: 3,000+ atoms run continuously for 2+ hours (2025): 3,000Caltech: 6,100 trapped atoms, not entangled (2025): 6,100D-Wave Orion demo (2007): 16D-Wave One, bought by Lockheed Martin (2011): 128D-Wave Two (2013): 512D-Wave 2X (1,000+ qubits) (2015): 1,000D-Wave 2000Q (2017): 2,000D-Wave Advantage (5,000+ qubits) (2020): 5,000D-Wave Advantage2 (4,400+ qubits) (2025): 4,400NIST: two-ion entangled state, 83% (Bell state) (2000): 0.17NIST: geometric phase gate, 97% (Bell state) (2003): 0.03Innsbruck: 99.3% gate (2008): 0.007Oxford: 99.9% gate (NIST matched it) (2016): 0.001Quantinuum H-Series: 99.914% across all pairs (2024): 8.6×10⁻⁴Oxford Ionics: 99.97% (2024): 3×10⁻⁴IonQ: 99.99% (company claim, preprint) (2025): 10⁻⁴Quantinuum Helios: 99.921% across all pairs (2025): 7.9×10⁻⁴Google/UCSB: 99.4% CZ gate (2014): 0.006MIT fluxonium: above 99.9% (2023): 0.001Google Willow: 0.14% mean error across chip (2024): 0.0014USTC Zuchongzhi 3.0: 99.62% (2025): 0.0038Harvard/MIT: 97.4% CZ gate (2019): 0.026Harvard/MIT/QuEra: 99.5% on 60 atoms in parallel (2023): 0.005Harvard/MIT: 99.854% CZ gate (preprint) (2026): 0.00146QuTech: above 99.5% (2022): 0.005