Texas is giving data centers more than $1 billion in tax breaks each year

Texas is projected to lose $3.2 billion in sales tax revenue over the next two years from its data center exemption, with annual foregone revenue expected to reach nearly $1.8 billion by fiscal 2030. Lawmakers are considering repeal, reduction, or tighter eligibility rules amid rapid growth in AI-driven data center construction and rising concerns over electricity use and local opposition. The debate could materially affect Texas’ position as the top U.S. data center hub and may influence similar tax incentive reviews in other states.

Analysis

The market is underappreciating how fast this can migrate from a narrow Texas budget debate into a broader repricing of AI infrastructure economics. The first-order risk is not that one state trims incentives; it is that the industry’s capital stack has been implicitly subsidized by a tax model that compresses upfront project costs and raises IRR by more than marginal basis points. If Texas meaningfully narrows the exemption, the immediate losers are the landowners, power developers, and local contractors with the most speculative pipeline exposure, because many projects were likely underwritten assuming permanence of the current regime.

The second-order effect is on power demand planning. Data centers have been a cornerstone of utility load-growth narratives, but if incentive generosity starts to roll back across Texas, Virginia, and Illinois in parallel, the growth path shifts from “faster and local” to “selective and utility-constrained.” That argues for a dispersion trade: winners are utilities and grid equipment suppliers tied to actual interconnection and transmission buildout, while losers are pure-play speculative developers that need tax relief to justify subscale locations or weak power economics.

The contrarian view is that repeal risk may be overstated near term because policymakers still want the jobs, transmission upgrades, and prestige of being an AI hub. A clean repeal is politically noisy but a phased cap, narrower eligible spend list, or sunset extension is more likely, which would dampen headline risk while preserving most projects already financed. That means the best short is not the sector broadly, but the most levered “incubator” names and adjacent local beneficiaries priced for uninterrupted growth.

If the debate worsens, the tail risk is that higher effective project costs push some capacity to cheaper power jurisdictions or to self-generation models, which could ironically accelerate behind-the-meter gas and turbine demand. That would be bearish for speculative colocation economics but constructive for distributed power and electrical equipment demand over a 12-24 month horizon.

Key Decisions for Investors

• Short speculative data-center REIT exposure on rally days; prefer a basket short versus EQIX/DLR only if valuation already discounts limited policy risk, and focus on names with the highest embedded Texas growth assumptions over the next 6-12 months.
• Go long grid capex beneficiaries (ETN, HUBB, PWR) into the legislative review window; risk/reward is favorable because even a partial rollback likely redirects spend toward transmission, switchgear, and EPC work over 12-24 months.
• Pair trade: long CEG or NEE / short a basket of local Texas land-bank or private-data-center proxy names if accessible; the thesis is that actual load-serving assets survive policy tightening better than greenfield speculation.
• Buy medium-dated put spreads on the most levered data-center development story ahead of interim hearings; the catalyst is months away, while implied volatility should be cheap relative to policy uncertainty.
• Stay constructive on natural gas turbine and distributed generation suppliers if repeal rhetoric intensifies; a shift toward behind-the-meter resilience can support orders even if some hyperscale projects delay.