Advantage2
Advantage2
Register profile — the machine's variant of its architecture
qubit type: rf-SQUID flux qubit, sixth-generation annealing processor
gate mechanism: coherent quenches 3.6-27 ns on 1,222-5,627 qubits (not in the cited releases); Multicolor Annealing and Fast-Reverse Anneal added Jan 2026
connectivity: Zephyr, 20-way
control: on-chip flux DAC multiplexing - ~300 bias lines in total, ~200 multiplexed lines addressing the DAC network, for ~1e4 qubits (0.02-0.03 control lines per qubit); ~12.5 kW system power; multi-year cryogenic uptimes
control placement: on-chip flux DAC
readout: QFP / SQUID flux readout through on-chip shift registers
gate evidence: no published figure
codes: qac
flags: annealer reference only, cryogenic details missing, claim rebutted
Technologies by layer — the machine's cell per layer
| 1 Qubit carrier | rf-SQUID flux qubit (annealer) · primary ▤ §Experimental testbed (ADV2 prototype: 1,222 qubits, lower-noise process, Zephyr, 12 mK) ✅ rf-SQUID flux qubit, sixth-generation annealing processor; 4,400+ qubits, 40,000+ couplers; GA 20 May 2025; 2× coherence, 4× lower noise vs Advantage |
| 2 Encoding | — none — nothing in this layer ◌ 🔎 analog annealer has no encoding layer; the anneal architecture declares an empty L2 |
| 3 Gate mechanism | Quantum annealing (analogue evolution) · primary ▣ §Quantum annealing (Methods); Fig. 1 (sampling post-quench states); Fig. 5 (ta = 7 ns and 20 ns) ✅ analog anneal; Multicolor Annealing (mid-anneal projection) and Fast-Reverse Anneal added 27 Jan 2026; coherent quenches 3.6–27 ns on 1,222–5,627 q (Atlas g_anneal) |
| 4 Connectivity / transport | Long-range on-chip couplers (c-couplers, mm-scale) · primary ▣ §Zephyr Graph; figures 'Couplers in the Zephyr topology', 'Unit cell in the Zephyr topology' ✅ Zephyr topology, degree 20; 40,000+ couplers |
| 5 Control | On-chip flux-DAC multiplexing · primary ▣ Fig. 3–5 (Φ-DAC schematic, SFQ sources); Fig. 8 (XYZ addressing); §III.E XYZ-addressing line count ✅ on-chip flux-DAC multiplexing: ~300 bias lines, ~200 multiplexed DAC lines for ~10⁴ qubits (0.02–0.03 lines/qubit); ~12.5 kW; multi-year cryo uptimes (Jan 2026) |
| 6 Readout | Flux readout via QFP / SQUID shift registers (annealers) · primary ▦ §Advantage2_system1 (readout time range; readout error rate) 🔎 Advantage2_system1 readout time 17–101 µs (one qubit to full QPU), readout error ≤ 0.001; mechanism not named on the page (QFP shift registers shown for Advantage, arXiv:2108.02322) Dispersive microwave readout (+TWPA, Purcell) · alternate ◌ 🔎 path-assigned ro_disp describes transmon readout, not annealer readout; actual is QFP/SQUID flux readout |
| 7 Code | — (Atlas gap: ∅G-qac) · primary ▤ abstract; device section ✅ penalty-spin and stacked quantum annealing correction on Advantage2_system1.10 (sweep 7 Oct 2026, source read) |
| 8 Decoder | |
| 9 Interconnect | — none — nothing in this layer ◌ 🔎 monolithic single QPU, no interconnect; no machine-specific source |
| 10 Manufacturing | Superconducting-qubit lithography (Al/AlOx/Al junctions, Nb wiring, 300 mm) · primary ▤ §Experimental testbed ('ADV2 was fabricated using a lower-noise process') ✅ superconducting flux-qubit lithography (in-house D-Wave process not published); no machine-specific fab source |
Standard records
connectivity degree → 20 neighbours · best (Zephyr)
May 2025 · source
control lines per qubit → 0.03 lines/qubit · best (~300 bias lines for ~1e4 qubits)
the most transferable piece of D-Wave engineering · Jan 2026 · source
evidence: ✅ 6 / 11
Pictures
Every picture carries the story of what it shows and its credit line.

