Cycle F candidate — Quantum Compute
conceptAI & Compute
The AI boom is hitting the physical limits of regular computer chips — and quantum computers are lining up as the next foundation, arriving in waves that could crack encryption years apart.
Who they are
The engine's candidate for the next computing foundation: quantum computing[1].[2][3][4][5]
What they do
It reads the current 'compute crisis' as a sign that classical silicon chips are hitting physical limits, requiring quantum architecture to handle massive prediction and capital-allocation problems.
How it works
The $40B Alphabet/Anthropic effort and its 5GW power commitment signal those limits; vendors are converging on 2028-2030 (IBM's Starling in 2029, Quantinuum by 2030, PsiQuantum's ~1M qubits in Brisbane by end-2027), and 'Q-Day' (when encryption breaks) is expected to arrive in waves, with some asset types falling years before banking systems. Reality check now in the record: the biggest number a quantum computer has honestly factored is 21, and only with shortcuts — while the paper estimates of what breaking real encryption would take have collapsed roughly twenty-fold since 2019. Both facts are held together: today's machines are toys, and the goalposts are racing closer.
Why it matters
The engine keeps all its readings open on this one, timing it toward a roughly 2040 window for mass commercial viability.
The engine's record — word for word
Substrate-handoff candidate (Sixth Prompt 2.4-iii). 'Compute crisis' driving the Alphabet/Anthropic $40B + 5GW grid commitment indicates physical limits of classical silicon; quantum architecture required to resolve complex capital-allocation + predictive algorithms. Phoenix metronome timing: aligns with 2040 transit window for mass commercial viability. Apex Superposition: admits all four. [Q-Day report] Vendor convergence on 2028-2030: IBM Starling (200 logical qubits, 2029, qLDPC); Quantinuum (universal FTQC by 2030, 'skinny logic' 48 logical from 98 physical); PsiQuantum (~1M photonic qubits, Brisbane end-2027). Neutral-atom architectures threaten P-256 with only ~26,000 physical qubits (all-to-all connectivity, multi-hour coherence) — Q-Day arrives in WAVES (ECC-curve assets fall years before RSA banking mainframes). [Report #186] THE HONEST DEMONSTRATED RECORD (vs roadmapped vs marketed, held apart): the celebrated Shor factorizations (15, 21) used circuits compiled with prior knowledge of the answer — no demonstration to date fulfills uncompiled Shor beyond trivial scale — and the 2019 IBM attempt at 35 succeeded on only ~14% of runs (Amico et al., PRA 100:012305). The 2022 SQIF claim (Yan et al., arXiv 2212.12372: 48-bit integer, 10 superconducting qubits, '372 physical qubits for RSA-2048') rests on a classical lattice-reduction step the cryptanalytic community judges non-scaling (Aaronson's 'cargo cult quantum factoring'; the technical replies Schneier circulated). Regev's 2023 multidimensional speedup (arXiv 2308.06572, ~n^1.5 gates vs Shor's n^2) trades gate count for qubit count plus classical lattice post-processing. THE TREND POINTS THE OTHER WAY AT THE RESOURCE LAYER — see gidney_rsa_2025 (the 2019 Gidney-Ekerå ~20M-noisy-qubit / ~8-hour RSA-2048 baseline, published Quantum 2021, collapsed to <1M qubits / under a week in Gidney's May 2025 revision) and google_quantum_paper (Mar 2026, secp256k1 under 1,200 logical qubits). Both directions held: the demonstrated record is microscopic AND the resource estimates are collapsing ~20x per half-decade. Q-Day timing stays parked (SYNTHESIS-CANDIDATE; no prediction rewrite).
Follow the trail
Walk this on the live map →