Data center tax-free dividends change! More than a dozen states in the U.S. are re-evaluating tax incentives, as the expansion of AI computing power faces obstacles.

Tax incentives relevant to the construction of major data centers in the United States are being reexamined, with over ten states reassessing or scaling back subsidies, while more than 35 states continue to offer related incentives. Last year, Ohio's tax exemption for data centers exceeded $1.5 billion, reaching more than ten times the initial forecast, prompting some members of Congress to call for the elimination of these incentives and the renegotiation of long-term agreements. Amazon.com, Inc., a leader driving the unprecedented wave of AI data center construction in the U.S., emphasized that it has invested nearly $40 billion in the state since 2015, creating thousands of jobs. The core dispute between American tech giants and state governments revolves around how to further weigh investments, employment, and tax revenue against the public utility costs brought by data center expansion; for tech giants, the uncertainty surrounding long-term tax arrangements is becoming a new variable in capital expenditure returns. For Amazon.com, Inc., Meta, and Alphabet Inc. Class C, the direct impact of reducing tax incentives is higher construction costs and pressure on post-tax returns from projects. After losing sales tax exemptions on server clusters and related equipment for data centers, cash expenditures for purchasing and updating equipment have increased; if companies also have to bear more costs for power supply and grid construction, the payback period for data center investments may also be extended. However, Ohios suspension of new applications and the proposal from lawmakers to eliminate incentives and renegotiate existing contracts are measures at different stages, and it cannot be concluded that tech giants have lost all long-term tax-free status at this time. The launch of Astra has heightened market expectations for Artificial General Intelligence (AGI), resulting in a stronger demand trajectory for computing power, especially as the new growth model of "pay-for-results" is expected to drive even greater overall demand for computing power. More direct evidence of computing power demand has recently emerged from the AI development process itselfwhere AI begins "creating AI" via recursive self-improvement (RSI) research trajectories. As advanced cutting-edge large models led by Astra generate increasing demand for AI computing power, Morgan Stanley predicts that the combined capital expenditure for the four major supercloud and AI application firms in North America will rise from $917 billion in 2026 to $1.47 trillion in 2027 and $1.64 trillion in 2028, while the expected deployment capacity will expand from 35 GW in 2025 to 145 GW in 2028. As states reassess tax incentive agreements, Amazon.com, Inc., Meta, and Alphabet Inc. Class C face new risks. Over ten states are reevaluating or reducing tax subsidies for data centers. These facilities, once welcomed as crucial engines for the acceleration of the AI economy, are increasingly viewed as political risks by local leadership. As opposition to these facilities intensifies, major U.S. tech giants like Amazon.com, Inc. (AMZN.US), Facebook's parent Meta (META.US), and Alphabet Inc. Class C (GOOG.US) (GOOGL.US) face the risk of losing tax exemptions that could last for decades. More than a decade ago, Ohio legislators eliminated the sales tax on the purchase of computer servers and other equipment required by tech companies, betting that these multi-million dollar tax adjustments would attract much-needed investment to the state. This initiative worked, making Ohio one of the primary locations for data centers. However, reports indicate that the subsequent AI boom sharply expanded the scope of tax exemptions, surpassing $1.5 billion last year, exceeding the states initial estimates by more than tenfold. Now, some state legislators, including Democratic state representative Tristan Reed, are eager to eliminate the sales tax exemptions and renegotiate previous agreements with companies like Amazon.com, Inc., Meta Platforms, and Alphabet Inc. Class C. These companies secured tax-exempt status for decades through contracts with state governments. Reports indicate that in several states, the annual tax exemptions enjoyed by tech giants have exceeded $1 billion. They seem richer than God, and even without these incentives, they have the ability to build these facilities, Reed remarked. He is proposing a new tax on data centers and demanding that developers bear more costs for power and infrastructure. His district encompasses parts of Cleveland, where local residents have expressed opposition to these facilities. He hopes these new pressures will force companies to come to the negotiating table. In a statement, Amazon.com, Inc. noted that since 2015, it has invested nearly $40 billion in data centers in Ohio, creating thousands of jobs, and paid nearly $11 million in property taxes and related fees in the state last year. Meta and Alphabet Inc. Class C declined to comment, and Ohio's governor's office also did not respond. The attack on tax incentive provisions is the latest example of rising opposition to the AI wave across the U.S. Reports have indicated that John Perkins, a city council member in Independence, Missouri, had previously voted to approve billions of dollars in tax incentives for a data center but later lost in a recent election. Despite recent polls showing that the vast majority of Americans do not want data centers built near them, President Trump continues to urge voters to welcome these facilities. If you want your state or town to get rich, want to pay less in taxes, want to create tremendous wealth, want your houses and other assets to be worth more, then you want a big AI data center, he stated on Friday. If you want to live in poverty, crime, and filth, then I would say dont approve the construction of a data center. Media reports indicate that Ian Bocaccio, a senior researcher at the tax service company Ryan, said that after states like Ohio, Arizona, and Illinois expressed intentions to pause or completely eliminate some tax exemptions, the investment attractiveness has declined. He indicated that he is urging clients of data centers to consider utilizing other investment tax incentives offered by various states. Industry observers generally believe that, ultimately, the actual benefits and future growth advantages brought by these facilities are too significant for the government to exclude. How does this affect the global theme of AI computing power investment? Resistance to data centers is broadening the constraints on AI investment, extending beyond chip and power supply to construction permits and community acceptance. Annenberg reported in August that a survey conducted from June to July showed that the percentage of adults opposed to new local data centers rose from 49% to 61%, although public sentiment regarding the overall impact of AI has not changed significantly. At the same time, Data Center Watch confirmed that at least 75 projects, worth approximately $130 billion, were either obstructed or postponed in the first quarter. For the AI supply chain, the direct impact is that the timelines for new computing power coming online have become more uncertain, potentially delaying equipment delivery, cloud service expansion, and the recognition of AI-related revenues that the market is focusing on, thereby causing a pullback due to compressed valuations in the AI computing supply chain. The contentious economic substance revolves around who bears the new costs of electricity generation, transmission, and water supply. After tech companies signed the "Electricity Ratepayer Protection Pledge," disputes over cost-sharing still exist. Reports in early September indicated that Microsoft Corporation is appealing new regulations in Virginia requiring data center developers to prepay for transmission infrastructure construction costs. This case illustrates that companies bear their own electricity costs involves issues of payment timing, infrastructure scope, and risk allocation. If more expenses must be borne by developers in advance, the initial capital investment and financing requirements for projects will increase; if costs are passed on to residents, this may continue to provoke political resistance. How these commitments are incorporated into electricity pricing agreements and construction conditions will directly impact the investment returns for data centers. For the theme of AI computing power investment, policy changes primarily affect where to build, at what costs, and when to realize investment returns. For example, the tax incentives in Ohio cover computing equipment, cooling systems, power equipment, and some construction materials, so the adjustment scope may span the entire data center campus. According to this cost transmission mechanism, Amazon.com, Inc. AWS and Alphabet Inc. Class C Cloud need to reassess the expansion costs of marketable computing power, while Meta faces changes in the infrastructure costs needed to support ad recommendations and AI products. However, capital spending may not decrease but instead be redistributed among regions, and it is important to note that rising tax burdens do not in themselves indicate a decline in final AI demand. The potential for mid- to long-term positive momentum depends on whether the new costs can lead to clearer construction and access conditionsif cost-sharing is clarified, and the project can secure a stable power supply, companies will find it easier to plan production and establish robust timelines for generating AI-related revenues. AEP Ohio has already implemented specific rates for data centers and post-grid systems, advancing new projects through batch studies, signed contracts, and clearly defined anticipated power delivery dates. This somewhat illustrates that tech companies taking on more infrastructure responsibilities and continued project expansion can happen simultaneously.