UpTrajectory Review

Congress is now treating AI infrastructure as a taxable event in its own right, not merely as collateral damage of broader tech expansion. Two distinct proposals have emerged that would specifically target data centers—the power-hungry facilities that train and run large AI models—with new federal levies. This represents a notable shift from the past decade of tax policy, where data centers benefited from accelerated depreciation, state and local incentives, and relatively light federal scrutiny. The Tax Foundation authors frame this as policymakers grappling with three externalities they had previously ignored: land consumption, energy grid strain, and workforce displacement. What makes this moment significant is that the taxation is arriving before the infrastructure is fully built, suggesting Congress wants to shape the buildout rather than merely clean up after it.

For small-business operators, the immediate risk is not that you own a data center. It is that you will pay for this tax regime through every vendor contract that touches cloud computing, AI APIs, or managed IT services. The large operators—Amazon, Microsoft, Google, and the emerging AI labs—will pass through any new cost structure, and they have the market power to do so without losing share. A small manufacturer using Azure for inventory forecasting, or a local marketing firm paying for OpenAI's API, will see those costs rise. The more subtle danger is competitive: if these taxes are structured as user fees or excise taxes tied to compute capacity, they may entrench the largest players who can absorb compliance costs that would crush smaller infrastructure competitors. The policy could inadvertently lock in the oligopoly it purports to regulate.

What is genuinely new here is the specificity. Congress is not talking about a generic 'tech tax' or digital services fee aimed at advertising revenue, as European counterparts have done. These proposals zero in on the physical footprint of AI—megawatts of power, acres of land, gallons of cooling water. That specificity is both more honest and more complicated than prior efforts. It acknowledges that AI is an industrial activity, not merely a software layer. Where the Tax Foundation analysis likely merits skepticism—and where small operators should press for detail—is whether the revenue is earmarked for grid expansion and workforce retraining, or simply disappears into general funds. The authors note the three stated concerns but do not, in the excerpt provided, indicate whether the bills themselves make those linkages. A tax on energy consumption that does not fund energy infrastructure is just a cost shock with no compensating benefit.

The downstream effects will split unevenly across geographies and sectors. Rural communities currently courting data center investments with tax abatements may find those deals renegotiated or abandoned if federal taxes erase the margin. Electric cooperatives and municipal utilities that planned revenue from massive new loads will face stranded capacity. Meanwhile, the carbon accounting implications are unresolved: if a Virginia data center pays a federal tax but still draws from a coal-heavy grid, has anything meaningful changed? For small operators, the second-order effect to watch is vendor consolidation. If taxes raise the cost of bare-metal infrastructure, the incentive to vertically integrate—own the chip design, the facility, the power purchase agreement—grows stronger. That is a game only the largest can play.

Watch whether these proposals move as standalone bills or get folded into larger packages where their specifics get diluted or weaponized. The Tax Foundation piece signals this is early-stage, which means small-business trade groups still have room to shape the outcome. Operators should ask their representatives two questions: first, whether any tax is calibrated to avoid burdening small-scale cloud users; second, whether revenue flows to the actual problems cited—grid reliability and worker transition—or merely to deficit reduction. The takeaway for action is to audit your current AI and cloud spend now, before pricing changes, and to diversify across providers where contract terms permit. Dependency on a single platform was already risky; if that platform faces a targeted tax regime with pass-through certainty, it becomes expensive too.

Takeaway: Audit your AI and cloud vendor contracts now—targeted infrastructure taxes will flow through to small-business pricing faster than regulation catches the largest providers.