This market resolves YES if, before January 1, 2030, the United States federal government enacts a law or binding federal regulation that creates a tax, fee, excise, levy, surcharge, or mandatory payment whose base is substantially tied to AI/data-center compute.

For purposes of this market, “compute tax” is interpreted broadly and includes any federal tax/fee based on one or more of the following:

• data-center electricity consumption;

• data-center energy capacity or grid interconnection capacity;

• AI accelerator/GPU ownership, import, deployment, or usage;

• FLOPs, accelerator-hours, GPU-hours, or similar compute-capacity metrics;

• large AI model training runs above a defined compute threshold;

• AI inference/API usage or AI compute revenue;

• AI lab or hyperscaler revenue/profits specifically attributed to AI compute;

• mandatory federal payments imposed on large AI/data-center operators to offset grid, energy, labour-displacement, safety, or regulatory costs.

It does not need to be officially called a “compute tax.” A law or regulation described as an AI infrastructure fee, data-center surcharge, accelerator excise tax, AI deployment levy, frontier-model training fee, hyperscaler AI levy, or similar would count if compute/data-center/AI infrastructure usage is a major basis for liability.

Edge cases:

• If a broad federal energy, carbon, or grid fee applies to all large electricity users but includes a special higher rate, special threshold, or special category for data centers/AI compute, it counts as YES.

• If a law taxes “digital services” generally but does not substantially target AI compute, data centers, model training, inference, GPUs, or hyperscalers’ AI infrastructure, it resolves NO.

• If a law is enacted before the deadline but implementation begins after 2030, it still resolves YES.

• If a law is passed by Congress but vetoed and not overridden before the deadline, it resolves NO.