Manifold MarketsNo KYCVerified resolution dataFairly calibratedCurrent community forecastYes 62.2%
Forecasters
20
Question type
binary
Methodology
Play-money forecasting platform
Source type
Forecast
Market data
Updated 13 hours ago
StaleSep 23, 24, 8:41 AM UTCJan 2, 29, 7:59 AM UTC
Trends
Related Markets
Related questions for context, not a comparison of equivalent contracts.
Will we actually see someone bomb an AI datacenter before 2027?
Manifold MarketsWhich company will operate the largest AI datacenter campus at the end of 2027?
Manifold MarketsWill an AI datacenter be hit with an EMP before 2031?
Manifold MarketsWhich company has the best AI model end of October?
PolymarketAnthropic IPO Closing Market Cap
PolymarketBest AI at the end of 2026?
KalshiRules
This question resolves yes if there is credible reporting of an incident of sabotage, between September 2024 and December 2028 inclusive, of a datacenter used for AI training or inference in the US (or dedicated power infrastructure for such a datacenter) resulting in either $100M+ in physical damages to the datacenter (and/or its dedicated power infrastructure), or a 2+ month delay in the facility's effective use of at least 50,000 H100 (or better) GPUs.
- In particular, for an incident to count as sabotage, the following must be widely believed:
- The incident was the result of a premeditated, malicious action by some party, rather than an accident.
- Damaging AI-relevant infrastructure was a primary motivation of the incident, such that the damage to AI-relevant infrastructure appeared to be targeted rather than incidental.
- For example, large-scale attacks on critical infrastructure wouldn't count if there is scant evidence that the responsible party was not primarily interested in damaging AI-relevant infrastructure.
- To determine whether this condition is met, I'll gauge the extent to which reporting characterizes AI-relevant infrastructure as the target of such an attack.