Quantum Computer Roadmaps

Ian Reppel

Quantum computer roadmaps for various qubit modalities and vendors, including references to official sources. Qubit modalities include superconductors, trapped ions, neutral atoms, quantum dots, photonics, spin (i.e. nitrogen vacancy centres in diamond), NMR, and quantum annealers. For each quantum computer generation, the number of physical qubits, and, where available, quantum volume, logical qubits, relaxation times, and gate fidelity statistics are listed.

A visual representation of this data is available in Roads to Quantum Advantage (2022) and Quantum Computing Roadmaps (2024). For more details on the different technologies, please check What is Quantum Computing?

Data

Organization Modality Designation Year Qubits (physical)The actual number of physical qubits on the quantum device. This is the maximally addressable qubits in quantum algorithms, although in practice that number is lower due ancillary qubits needed in error correction schemes. Qubits (logical)The number of logical (a.k.a. algorithmic) qubits that can be addressed in quantum algorithms. Ideally, this is the base-2 logarithm of the quantum volume. Quantum volumeA NISQ-era metric that assumes square circuits of qubits and measures the effective number of qubits multiplied by the circuit depth (i.e. number of layers) in which quantum algorithms can be run within acceptable tolerances. QumodesNumber of quantized beams of (squeezed) light in continuous-variable photonic quantum computers. T1 (μs)The T1 time is the relaxation time, which measures the characteristic time to go from the |1⟩ to the |0⟩ state. T2 (μs)T2 measures characteristic time of decoherence due to dephasing, the randomization of the qubit phase, and phase breaking, the loss of superposition. Fidelity
  • 1Q stands for one-qubit gates (e.g. phase rotation).
  • 2Q stands for two-qubit gates (e.g. CNOT or Hadamard).
  • SPAM is an acronym of State Preparation and Measurement.
Notes
Alibaba Superconductors 2018/Q1 11
2022/Q2 2 99.97% (1Q)
99.72% (2Q)
Fluxonium qubits
Alice & Bob Photonics Boson 4 2024/Q4 4 0 Cat qubits
Helium 2026 10–16 1 10-2 logical error rate Cat qubits at 1.5 μs clock speed
Lithium 2027 48 4 10-3 logical error rate Cat qubits at 0.8 μs clock speed
Beryllium 2028 250 5 10-4 logical error rate Cat qubits at 0.8 μs clock speed
Graphene 2030 2000 100 10-6 logical error rate Cat qubits at 1 μs clock speed
Alpine Quantum Technologies Trapped ions Pine 2022/Q1 20
Amazon Superconductors Ocelot 2025/Q1 9 5 cat (data) qubits with 4 ancilla qubits
0.0165 logical error rate
Atom Computing Neutral atoms Phoenix 2021/Q3 100
2024 1,180
Baidu Superconductors Qian Shi 2022/Q3 10 31.0 8.7 99.8% (1Q)
96.60% (2Q)
Median instead of mean T1/T2 values
ColdQuanta Neutral atoms 2021/Q3 100
2022 300
Hilbert 2024 1,000
D-Wave Annealer 2007 28
One 2011/Q2 128
Two 2013/Q2 512
2X 2015/Q3 1,152
2000Q 2017/Q1 2,048
Advantage 2020/Q3 5,760
Advantage 2 2025/Q2 4,400 Originally expected in 2024 with 7,000 qubits
Fujitsu Superconductors 2023 64
2026 1,000
Google Superconductors Foxtail 2016 9
2017 20 99.5% (2Q)
Bristlecone 2018/Q1 72 99.9% (1Q)
99.4% (2Q)
99.0% (readout)
Sycamore 2019/Q4 53 99.85% (1Q)
99.35% (2Q)
1 qubit out of 54 defective, so 53 'effective' qubits
Sycamore 2023/Q1 72 >99.85% (1Q)
99.64% (2Q)
96.9% (readout)
1000:1 error correction overhead estimate unaltered due to extrapolation to distance-25 surface code
Willow 2024/Q4 105 >99.965% (1Q)
99.67% (2Q)
99.23% (readout)
Milestone 3 2025 100 1 10-6 logical error rate
Milestone 4 2026 1,000 1–10 10-6 logical error rate
Milestone 5 2027 10,000 10 10-8 logical error rate
Milestone 6 2028 100,000 100 10-13 logical error rate
2029 1,000,000 1,000 1000:1 error correction
IBM Superconductors (Tenerife) 2017/Q1 5 2 4 51.1 25.9 99.90% (gate)
98.64% (readout)
(Tokyo) 2018/Q1 20 3 8 84.3 49.6 99.81% (gate)
93.21% (readout)
Falcon (Johannesburg) 2019/Q4 27 4 16
Hummingbird (Raleigh) 2020/Q1 65 5 32
Falcon R4 (Montreal) 2020/Q3 27 7 128 109.09 97.17 99.96% (1Q)
98.58% (2Q)
97.69% (SPAM)
Eagle 2021/Q4 127 6 64
Falcon R10 (Prague) 2022/Q2 27 9 512
Osprey 2022/Q4 433 10 1,024 Estimated QV based on doubling every 6–12 months
Condor 2023/Q4 1,121 12 4,096
Flamingo 2024/Q4 1,386 14 16,384
Kookaburra 2025 4,158 16 65,536
Starling 2029 200 100m operations
Blue Jay 2033 2,000 1b operations
Infleqtion Neutral atoms 2024 1,600 2 99.90% (1Q)
99.50% (2Q)
2Q fidelity for CZ only
2026 8,000 >10 99.95% (1Q)
99.90% (2Q)
Sqorpius 2028 40,000 >100 99.95% (1Q)
99.95% (2Q)
Intel Quantum dots 2018/Q2 2
2022/Q1 3
Tunnel Falls 2022/Q4 12
Superconductors 2017/Q4 17
Tangle Lake 2018/Q3 49
IonQ Trapped ions Harmony 2019 11 8 99.72% (1Q)
96.54% (2Q)
99.71% (SPAM)
Aria 2020/Q4 32 4,000,000 99.98% (1Q)
99.3% (2Q)
13:1 error correction
Forte 2022/Q2 32
2023 29
2024 36
Tempo 2025/Q3 100 64 algorithmic qubits
2026 256 12 10-7 logical error rate
2027 10,000 800
2028 20,000 1,600
2029 200,000 8,000 10-12 logical error rate
2030 2,000,000 80,000
IQM Superconductors Spark 2021/Q4 5 99.8% (2Q)
2023/Q4 20 99.8% (2Q)
Radiance 2025/Q1 54 1–2 99.9% (2Q) Originally expected 2024/Q3
Radiance 2026 150 1–2 99.92% (2Q) Originally expected 2025/Q1
Radiance 2027 300 1–2 99.94% (2Q) Originally expected in 2026
Radiance 2027 1,000 4–36 10-5 logical error rate
Radiance 2028 5,000 60–180 10-6 logical error rate
Radiance 2030 40,000 240–720 10-7 logical error rate
Radiance 2031 100,000 600–1,800 10-8 logical error rate
Radiance 2033+ 1,000,000 2,400–7,200 10-9 logical error rate
Microsoft Topological Majorana 1 2025/Q1 8 1% measurement error on single-shot fermion parity
Majorana zero modes not yet confirmed
OQC Superconductors Lucy 2022/Q1 8 1 2 99.91% (1Q)
94.16% (2Q)
90.44% (SPAM)
Origin Quantum Superconductors Wuyuan I 2020/Q3 6
Wuyuan II 2023/Q1 24
Pasqal Neutral atoms 2022/Q3 324 Lab hardware, not commercial yet
2024/Q2 1,110
2027 10,000
PsiQuantum Photonics 2025 1,000,000 300
QpiAI Superconductors Kaveri 2025/Q4 64
Quandela Photonics Ascella 2022/Q4 5
2023 12
Quantinuum Trapped ions H1-1 2020/Q3 20 7 128
H1-1 2021/Q1 20 9 512 99.97% (1Q)
99.5% (2Q)
H1-1 2021/Q3 10 10 1,024
H1-2 2021/Q4 20 11 2,048
H1-2 2022/Q2 12 12 4,096
H1-1 2022/Q3 20 13 8,192 99.9959% (1Q)
99.71% (2Q)
99.72% (SPAM)
Arbitrary-angle gates
H1-1 2023/Q1 20 15 32,768 99.9955% (1Q)
99.795% (2Q)
99.69% (SPAM)
Arbitrary-angle gates
H2 2023/Q2 32 16 65,536 99.997% (1Q)
99.8% (2Q)
99.8% (SPAM)
All-to-all connectivity
H2 2025/Q2 56 23 8,388,608 99.98% (1Q)
99.91% (2Q)
99.85% (SPAM)
All-to-all connectivity
Helios 2025/Q4 98 50 99.997% (1Q)
99.92% (2Q)
99.95% (SPAM)
All-to-all connectivity
H3 2025 17 100,000 Estimated QV based on 5× each year
H4 2027 21 500,000
H5 2029 24 2,500,000
Quantum Brilliance Spin Gen1 2021/Q2 5
50
QuantWare Superconductors Tenor 2023/Q1 64
VIO-40K 2025/Q4 10,000
QuEra Neutral atoms Aquila 2023/Q4 256 99.9% (initialization)
62% (readout)
Post-computation QEC on non-universal gate set with [[8,3,2]] colour code only
2024 256 10 Built-in QEC
2025 3,000 30 Built-in QEC
2026 10,000 100 Built-in QEC
2025 1,024
QuiX Photonics 2020/Q4 12
2022/Q1 20
2023 50
Rigetti Superconductors Agave 2017/Q2 8 13.38 15.05 95% (1Q)
87% (2Q)
Aspen-1 2018/Q4 16 8 12 93.23% (1Q)
90.84% (2Q)
Aspen-4 2019/Q1 13 3 8 30.47 20.13 99.88% (1Q)
94.42% (2Q)
Aspen-9 2021/Q2 32 3 8 33 16 99.39% (1Q)
94.28% (2Q"
Median instead of mean values
Aspen-11 2021/Q4 40 3 8 30 14 99.55% (1Q)
92.15% (2Q)
96.78% (SPAM)
Aspen-M 2022/Q1 80 3 8 29 22 98.94% (1Q)
91.13% (2Q)
96.95% (SPAM)
Ankaa-3 2024/Q4 84 99.5% (2Q)
Cepheus-1-36Q 2025/Q3 36 99% (2Q)
Lyra 2025 336 Originally planned for 2023
2025 1,000 Originally planned for 2025
2027 4,000 Originally planned for 2027
SpinQ NMR Gemini 2021 2
Triangulum 2022 3
SQC Quantum dots 2023 10
2030 100
USTC Photonics Jiuzhang 2020/Q4 76
Jiuzhang 2.0 2021/Q4 113
Jiuzhang 3.0 2023/Q4 255
Superconductors Zuchongzhi 2021/Q2 62
Zuchongzhi 2 2023/Q2 66 176 qubits after upgrade
Zuchongzhi 3 2025/Q1 105 99.90% (1Q)
99.62% (2Q)
99.13% (readout)
Xanadu Photonics X2 2018 2
X12 2019 12
X24 2020 24
Borealis 2022/Q2 216

References

General

Alibaba

Alice & Bob

Alpine Quantum Technologies

Amazon

Atom Computing

Baidu

ColdQuanta

D-Wave

Fujitsu

Google

IBM

Infleqtion

Intel

IonQ

IQM

Origin Quantum

Microsoft

Pasqal

PsiQuantum

QpiAI

Quandela

Quantinuum

Quantum Brilliance

QuantWare

QuEra

Quix

Rigetti

SpinQ

Silicon Quantum Computing (SQC)

University of Science and Technology (USTC)

Xanadu