Wyden's Data-Center Tax Is the Gift That Keeps on Giving

Wyden's proposal turns data-center scrutiny into a recurring revenue concept

This is bigger than one legislative setback. Wyden is trying to turn the data-center boom into a standing federal revenue stream. His plan does it in straightforward terms: stop giving data centers the same tax breaks other projects rely on, then add a Data Center Public Investment excise tax. The second piece is the key one. The proposal is framed not as a one-off adjustment, but as a consistent revenue stream.

The market should care because the backdrop is broad, not niche. Hyperscaler spending is set to be roughly $700 billion on data centers this year, and the buildout shows no sign of ending. At the same time, states are starting to question whether the old bargain still works. At least 38 states offer dedicated tax incentives for data centers, but many are reassessing them as utility costs and local opposition rise.

Bulls will say this is a straightforward re-pricing of a boom industry that has already enjoyed extensive help. Bears will note that the proposal is still just a white paper with no specific details on rates or structure. That is fair. But the feedback process matters as a signal: the concept is already in the ring. If the fight stays centered on rates rather than on whether data centers should contribute at all, investors will have a clearer read on the policy trajectory.

Power and local benefit matter more than headline tax risk

The tax fight is the headline, but the business test is simpler: does the site still work once you account for power, grid work, and what the community gets? In 2026, site selection is a tax-, incentive-, and infrastructure-driven financial strategy. That is the real pressure point. Washington can change the headline risk, but local power constraints change the project math.

New Jersey shows the local recalibration

New Jersey offers a useful example. Lawmakers are pushing to end the tax credit program and save $250 million in taxpayer money, arguing that AI data centers are being run by some of the wealthiest companies around. The point is not just ideological. It shows that subsidies are increasingly treated as a negotiating starting point, not a free pass.

There is also a transparency filter. Across the country, only 11 publicly disclose recipients of data-center incentives. When disclosure is that thin, investors cannot assume every deal is being judged the same way. That raises the importance of deal quality: the stronger companies are the ones that can show solid leases, realistic power timelines, and terms that do not lean too heavily on opaque taxpayer support.

Grid impact is where margins get hit

The bigger issue is electricity. Starting in 2025, a sharp policy recalibration began as regulators moved from unconditional incentives to direct accountability for grid impact. In practical terms, the old playbook of winning a project with tax breaks is giving way to demands that large customers help bear the cost of their power load.

That matters more to returns than many investors realize. Subsidies affect the incentive package; power infrastructure can affect whether the project is viable at all. Transformers, substation upgrades, renewable-energy conditions, and longer utility timelines can all press on costs and timing. And this is not a blanket hit on every building. At least 4,000 data centers are spread across the country, so the heaviest load - and the toughest local negotiations - will likely fall first on the largest, most power-intensive AI-driven projects.

Investor checklist: power, cost sharing, and deal quality

For investors, the operating rule is shifting. The useful checklist is simple:

• Power path: Can the facility get the megawatts it needs without open-ended delays?
• Cost sharing: Who pays for grid upgrades, renewables, and related community costs?
• Deal quality: Is the business strong enough to work if the subsidy package is smaller or more conditional?

That points investors toward operators and landlords that understand power first, can hold hyperscaler demand, and are transparent enough to underwrite. If a company's story breaks unless the tax break stays intact, that is a warning sign.

For investors, the cleaner exposure is hard assets with real power access

With the tax-and-subsidy fight already under way, the practical move is simpler: own the landlords that have the hard assets tenants actually need today. The top data-center REITs already posted 39%-45% returns over the last year, so this is not an obscure idea. The question is whether investors are still buying the right version of the story.

What to look for

Favor operators with three basics you can verify:

• Power access: Clear path to the megawatts required for near-term capacity.
• Tenant quality: Leases backed by hyperscalers or other strong operators.
• Execution: Transparent reporting and fewer dependencies on opaque subsidy packages.

How to read the market reaction

Watch how the group responds to bad policy news.

• Bullish sign: landlords with power and hyperscaler tenants hold up better than incentive-dependent developers.
• Neutral sign: the whole group sells off on Washington headlines even though occupancy and leasing fundamentals remain intact.
• Bearish sign: tenants start pushing more cost onto landlords for power upgrades, community costs, or grid work.

What would weaken the thesis

The cleanest invalidation signals are straightforward:

• States move from scrutiny to broadly tougher terms, and tenants successfully shift more of that burden onto landlords.
• Power constraints ease enough that the sector's main bottleneck stops mattering.
• Leasing demand cools materially, removing the scarcity value that supports current narratives.

That last point matters because it would raise the bar on every new project economics model. In this tape, prefer tenants with cash, landlords with power, and leases you can trust. Skip the stories that only work if the tax break survives.