What Prediction Markets Are Forecasting

Prediction markets currently give about a 1 in 3 chance that the United States will pass a law halting new AI data center construction before 2027. Traders collectively see this kind of moratorium as unlikely, but not impossible. This suggests a significant minority believes political or public pressure could build enough to force a temporary stop on building these large, energy-intensive facilities.

Why the Market Sees It This Way

The low probability reflects two main realities. First, there is a strong national and economic incentive to build AI infrastructure. Both major political parties and the Biden administration have framed leading in artificial intelligence as a priority for economic and national security, making a broad federal halt counter to stated goals.

Second, while local moratoriums have happened, they address local grid or water concerns. For example, some counties in Virginia and Arizona have paused data center approvals due to strain on power supplies. These are specific responses to infrastructure limits, not a national policy against AI development. A federal law would need to overcome the powerful lobbying efforts of tech companies and the current political consensus that the U.S. must compete aggressively in AI.

Key Dates and Events to Watch

Watch for two types of events. The first is any major, widespread power grid failure or water shortage that the public and media directly blame on data center growth. A crisis framed this way could rapidly increase political pressure for federal action.

The second is the outcome of the 2024 election and any significant shift in the political climate in 2025. A new administration or changed congressional priorities could alter the calculus, though a full moratorium would still be a major reversal from current policy. Also monitor whether several states simultaneously enact their own moratoriums, which could create momentum for a federal response.

How Reliable Are These Predictions?

Markets are generally decent at aggregating views on political outcomes, but this is a very specific and novel question. There is little direct historical precedent for a federal moratorium on a specific type of industrial infrastructure tied to a priority technology. The prediction may be more sensitive to sudden, high-impact news events than a typical election market. The low trading volume also means the current odds could shift more easily if new information emerges.