A fault-tolerant commercial quantum computing service is offered to enterprise customers before 2030.
Hardware developers are racing to demonstrate error-corrected logical qubits that outperform classical supercomputers on commercially viable algorithms.
Across 6 frontier models answering independently, the panel thinks this is unlikely: the median is 20%.
They do not agree: GPT puts it at 42% and Qwen at 12%, 30 points apart. A spread that wide usually means the question turns on something none of them can see yet.
Nothing is settled until 5 Oct 2031, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
It turns on whether error correction can scale from demonstrations to millions of reliable physical qubits at viable cost.
- Google, IBM, Quantinuum and PsiQuantum must turn logical-qubit milestones into customer-accessible systems by 2029, not merely lab results.
- No general-purpose fault-tolerant enterprise service yet exists; past hardware roadmaps have commonly slipped, and algorithms need far more logical qubits than demos.
- A credible 2027–28 system showing sustained logical-error suppression plus a funded million-qubit build would raise this materially.
42% -
GeminiGoogle
Turns on scaling physical qubits beyond 10,000 with low error rates to yield enough logical qubits for commercial advantage.
- IBM, Google, and Quantinuum roadmaps target 2029-2030 for early fault tolerance, but complex hardware milestones historically slip 2-3 years.
- True enterprise commercial utility requires hundreds of logical qubits with error rates under 10^-6, far exceeding current lab demos.
- A breakthrough in neutral atom or topological scaling providing 1,000 reliable logical qubits before 2029 would push odds above 60%.
28% -
ClaudeAnthropic
Whether IBM, Google, or IonQ can scale error-corrected logical qubits to commercial-grade reliability by 2030.
- Current roadmaps (IBM 2029, Google, PsiQuantum) target early fault-tolerant systems around 2029-2030 but with narrow commercial utility.
- Error correction overhead and qubit scaling remain unsolved at the scale needed for enterprise-grade reliable service, not just lab demos.
- A surprising breakthrough in error-corrected qubit count or logical error rates before 2029 would most change this, but current trajectory suggests early pilots not full…
22% -
GrokxAI
Any vendor must ship a cloud FT logical-qubit machine (not NISQ) by end-2029.
- IBM, Google, Quantinuum or PsiQuantum first need ~50-100 logical qubits below threshold; none exceed a handful of noisy logical qubits today.
- Post-first-logical-qubit, scaling to a commercial FT service has historically taken 5+ years and QC roadmaps routinely slip 3-5 yrs.
- A 2028 public 10-logical-qubit FT demo would most raise the odds; otherwise enterprise offering stays post-2030.
18% -
DeepSeekDeepSeek
This turns on whether error-corrected logical qubits reach commercial scale before 2030, not on qubit-count headlines.
- IBM's roadmap targets 200 logical qubits by 2029 and 2,000 by 2033; Google and Quantinuum are on similar timelines, so 2030 is the earliest plausible window.
- A commercial service needs fault tolerance plus a customer-ready algorithm, and no error-corrected logical qubit has yet beaten a classical supercomputer on a business…
- Precedents like IBM Quantum Heron and Google's Willow show rapid progress but still leave logical error rates orders of magnitude above what enterprise workloads require.
18% -
QwenAlibaba
Fault-tolerant QC requires millions of physical qubits with error rates below threshold; no one is close to that scale by 2029.
- IBM's 2029 roadmap targets ~100K physical qubits, far short of the millions needed for full fault tolerance on useful problems.
- Google's Willow (Dec 2024) showed below-threshold error correction but on a single logical qubit pair, not a commercially useful system.
- Even if a breakthrough in qubit quality occurred today, engineering a deployable enterprise service within 4 years has no precedent in the field.
12%
6 of 6 models answered · 30 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 6 Oct 2026. Forecast by google/gemini-3.8-flash, anthropic/claude-sonnet-5, openai/gpt-5.6-terra, x-ai/grok-4.6, deepseek/deepseek-v4.1-flash, qwen/qwen3.8-max-0902 via OpenRouter.