Google Data Center Deal Will Save Xcel Energy Customers Up To $1.5 Billion Over 15 Years

Xcel Energy struck a long-term electric service agreement with Google that could save Minnesota residential and small-business customers about $1.1 billion over the deal's life, while Google pays the infrastructure and generation costs for a 750-megawatt campus. Xcel also raised its five-year capital plan 33% to $60 billion and reaffirmed 2026 EPS guidance of $4.04 to $4.16, while Alphabet is advancing a massive AI buildout with Q1 2026 capex of $35.67 billion and full-year capex guidance of $175 billion to $185 billion. The main near-term risk is regulatory approval of the Clean Energy Accelerator Charge by the Minnesota Public Utilities Commission.

Analysis

The market is starting to price a structural change in the utility growth model: hyperscaler load is no longer just a demand shock, it is becoming a contracted infrastructure finance engine. That matters because it lowers the political discount rate on utility capex — if regulators can show incumbent customers are protected, they are more willing to approve transmission, generation, and storage spend that would otherwise be delayed for years. The second-order winner is not just XEL’s earnings growth, but its ability to de-risk a much larger pipeline of incremental load requests from other data-center operators.

For Alphabet, the strategic value is less about electrons and more about schedule certainty. In AI, the marginal advantage increasingly goes to the firm that can convert capex into usable compute fastest; every permitting delay creates an option value transfer to competitors with capacity already online. That should support a premium multiple even after the stock’s strong run, because the market is underestimating how much infrastructure execution is now part of cloud economics rather than a sunk cost.

The contrarian issue is that the “clean energy at scale” narrative may compress margins for the broader utility ecosystem. Developers, EPCs, transformer makers, grid equipment suppliers, and long-duration storage names should see faster order flow, but procurement inflation can offset some of the rate-base upside if the market gets too enthusiastic about buildouts. The biggest near-term risk is regulatory friction: one adverse commission ruling would not kill the thesis, but it could slow replication and reintroduce the old ratepayer-subsidy debate.

NVDA is the cleanest way to express that AI capex is still intact, but this article is not incrementally bullish enough to justify chasing it versus the infrastructure enablers. The better read-through is that AI demand is moving downstream into grid hardware, storage, and regulated assets, where cash flows are more visible and valuation risk is lower. That makes the setup more attractive on a 6-18 month horizon in utilities and power infrastructure than in the most crowded AI beneficiaries.

Key Decisions for Investors

• Long XEL vs short a utility basket (or XLU) for 6-12 months: the setup favors XEL’s idiosyncratic growth re-rating while broader utilities lack a comparable monetization path for data-center load. Risk: one adverse regulatory decision could compress the spread, so keep the short leg sized smaller than the long.
• Add to GOOGL on any 3-5% pullback over the next 1-3 months: the market is likely underpricing the value of execution certainty in AI infrastructure. Target a 12-18 month hold; stop if capex intensity rises without a corresponding cloud backlog inflection.
• Initiate a long on grid-enablement beneficiaries rather than pure AI semis over the next quarter: prefer companies tied to transmission, transformers, switchgear, or storage buildout. These names should see a 2-3 quarter order backlog acceleration if the template spreads beyond Minnesota.
• Do not chase NVDA here; use it as a hedge only. If hyperscaler capex stalls on regulatory bottlenecks, semis can de-rate faster than regulated assets re-rate, so a modest NVDA short against a long utility/infrastructure basket can improve risk-adjusted returns.
• Option expression: buy 6-12 month call spreads on XEL to capture a multi-state template rollout while capping premium outlay. The risk/reward improves if additional utility agreements are announced before the Minnesota approval process fully resolves.