Why a New Tax on AI Has Support From 3 Unlikely Places

A proposal to tax artificial intelligence processing power is gaining traction in policy circles, as the technology’s explosive growth shifts from abstract potential to a concrete, multi-billion-dollar physical footprint that is drawing public opposition. The debate, echoing a “robot tax” idea proposed by Bill Gates in 2017, has been re-energized by the massive resource demands of AI data centers, which is creating new political headwinds for the industry.

“We’re at a point now where we need to try and preserve jobs,” said Andrew Yang, a former presidential candidate and long-time proponent of universal basic income, in a recent interview. Yang argues that the value generated by AI companies is not matched by the taxes they pay, a gap that a so-called “compute tax” could close to fund social programs and offset AI-driven job displacement.

The abstract debate is being fueled by concrete local conflicts. In Kenilworth, New Jersey, a $1.8 billion AI data center from CoreWeave is proceeding despite a petition with over 4,000 signatures from residents who were largely unaware of the project. The facility could consume up to 250 megawatts, equivalent to the power used by 200,000 homes. Similarly, a proposed $9.8 billion Hut 8 data center campus in Nueces County, Texas, has raised concerns about water usage in a drought-stricken region.

These local skirmishes are crystallizing the costs of AI development, providing a new impetus for policymakers considering taxes to manage the industry’s growth. The conflicts highlight a growing disconnect between the tech industry’s expansion and the capacity of local infrastructure, turning abstract concerns about AI’s societal impact into tangible issues of power grid stability and water rights.

Data Center Pushback Fuels Tax Debate

The friction is most visible in New Jersey, where a recent Stockton University poll found 56 percent of voters support banning new data centers in their towns. Residents and local governments are increasingly pushing back against an industry they see as consuming vast local resources with few direct benefits. Several townships, including Pemberton and Monroe, have already enacted bans on new data center construction, citing concerns over energy and water demand.

This local opposition provides a powerful political backdrop for proponents of a compute tax. They argue that if the AI industry’s growth imposes significant external costs on communities—from strained power grids to increased water scarcity—then taxation is a valid tool to either slow that growth or generate revenue to mitigate its impact. Simon Johnson, a professor at the MIT Sloan School of Management, suggests a compute tax could make it less appealing for companies to replace thousands of workers with data centers.

A Tax on Progress or a Necessary Brake?

Opponents of a compute tax argue it could stifle innovation and push development to other countries. Pascual Restrepo, an economics professor at Yale University, notes that AI is already critical to advances in drug discovery and fraud detection. “Why do you want to increase the cost of all of that?” he asks. Critics also suggest that existing tax structures, like corporate income taxes, are sufficient to capture revenue from AI’s growth.

The alternative, favored by economists like Stanford’s Erik Brynjolfsson, is to realign existing tax incentives. Currently, the U.S. tax code often encourages companies to invest in machines over people. Brynjolfsson argues for shifting the tax burden away from labor and toward capital, making it more attractive for companies to use AI to augment their human workforce rather than replace it.

For investors, the compute tax debate introduces a new layer of regulatory risk to the AI sector, which has so far been valued primarily on technological potential. The stock prices of companies like Nvidia, Microsoft, and Amazon are predicated on continued, massive growth in AI compute. The emergence of a serious tax discussion, fueled by grassroots opposition to the very infrastructure that enables this growth, suggests that the industry’s path forward may not be as frictionless as once assumed.

This article is for informational purposes only and does not constitute investment advice.