A quantum computer successfully breaks standard RSA-2048 encryption before 2035.
A prominent security concern driving quantum search queries is whether Shor's algorithm will be realized at sufficient physical qubit scale to compromise modern public-key cryptography.
Across 6 frontier models answering independently, the panel has all but ruled it out: the median is 7%.
They broadly agree, 16 points between GPT at 18% and Qwen at 2%.
Nothing is settled until 15 Nov 2036, when the outcome is checked against sources and every number above gets a permanent score.
What each one said
-
GPTOpenAI
It turns on whether fault-tolerant machines can scale from early logical qubits to millions of reliable physical qubits by 2035.
- Shor factoring RSA-2048 needs error-corrected logical qubits and enormous gate depth, far beyond 2026 demonstrations.
- Hardware roadmaps target fault tolerance around 2030, but credible RSA-2048-scale resource estimates generally require millions of physical qubits.
- A public 1,000+ logical-qubit, low-error system by 2031 would materially raise the odds; repeated roadmap slips would lower them.
18% -
GeminiGoogle
Factoring RSA-2048 requires millions of noisy physical qubits with fault tolerance that hardware roadmaps cannot deliver by 2035.
- State-of-the-art systems have only reached ~1,000 noisy qubits, while Gidney-Ekerå estimates need ~20M physical qubits for RSA-2048.
- Roadmaps from IBM and Google target only 100k to 1M qubits by 2033, meaning fault-tolerant Shor execution before 2035 requires a leap.
- A massive algorithmic reduction slashing required logical qubits or unprecedented optical-modular scaling would sharply raise odds.
12% -
GrokxAI
Error-corrected logical qubits must hit ~4k+ with low enough error by ~2033.
- IBM/Google/IonQ roadmaps reach ~100k noisy physical qubits mid-2030s, still orders of magnitude short of the millions needed for RSA-2048 Shor.
- No quantum computer has factored anything larger than 21; nine years from NISQ is well below historical scaling of useful compute.
- A topological-qubit breakthrough or crash Chinese program would be the only realistic mover.
7% -
DeepSeekDeepSeek
Whether error-corrected logical qubits scale from today's ~100 to the ~4,000 needed for RSA-2048 within nine years.
- RSA-2048 needs roughly 4,000 logical qubits and millions of physical qubits with error correction; best current devices are under 1,500 physical qubits with no logical…
- Google's 2024 Willow result and Quantinuum's 2025 logical-qubit demos show progress, but the error-correction overhead gap is still ~3 orders of magnitude.
- A cryptographically relevant demonstration by 2035 would require sustained ~10x annual qubit-and-fidelity scaling, which no roadmap currently projects.
6% -
ClaudeAnthropic
Breaking RSA-2048 needs ~1M+ noisy physical qubits with error correction, far beyond current ~1000-qubit devices.
- Current best quantum computers (IBM, Google) have ~1000 physical qubits; RSA-2048 break estimates require millions of error-corrected qubits.
- Error correction overhead and coherence times are improving slowly; no credible roadmap puts cryptographically relevant quantum computing before early-to-mid 2030s at…
- A surprise breakthrough in qubit scaling or error correction (e.g., topological qubits working at scale) is the main thing that could shift this upward.
4% -
QwenAlibaba
Requires millions of physical qubits with fault tolerance; no one has demonstrated even 10 logical qubits as of late 2026.
- RSA-2048 needs ~4,000 logical qubits (~20M physical at current error rates); IBM's roadmap targets ~100K physical qubits by 2033, still 100x short.
- Google's Willow (105 qubits) and IBM's Condor (~1,121) are orders of magnitude below the threshold; logical qubit demonstrations remain at single-digit scale.
- A breakthrough reducing physical-to-logical overhead by 100x or a novel algorithm bypassing Shor's qubit requirements would be needed to shift this materially.
2%
6 of 6 models answered · 16 points between the highest and lowest. None was shown the market price.
Question sourced from a news sweep on 18 Nov 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.