National Grid Ventures to invest $1.75bn to accelerate power solutions for U.S. data centers and AI

National Grid Ventures will invest $1.75 billion for a 35% stake in Joulent LLC, targeting contracted power and electrical infrastructure for the AI-driven large-load economy. The deal supports Joulent’s 2.67 GW Project Kilby in West Texas, providing dedicated power to a Microsoft-operated data center under a 20-year PPA with first power targeted for 2028. With NGV expecting to connect more than 10 GW of demand across the U.K. and U.S. over the next five years, the news is modestly positive for related infrastructure and data-center power exposure.

Analysis

The economic value here is not the headline capital check; it is the validation that the scarce asset in AI power is speed-to-power, not cheap electrons. That makes the beneficiaries the firms with permitting, interconnection, turbine allocation, EPC capacity, and operating know-how — i.e., GEV and the broader gas-turbine / grid-equipment complex — while the losers are slow-moving regulated utilities and land-constrained data-center developers that cannot self-source load quickly.

For MSFT, the strategic benefit is real but mostly defensive: this reduces the probability that power becomes the binding constraint on Azure buildouts, but it also implies a higher capital-intensity model for AI infrastructure than the market may be underwriting. The second-order effect is that hyperscalers increasingly externalize power-development risk to specialist partners, which should preserve cloud demand but modestly pressure long-run FCF conversion if bespoke power arrangements become the norm.

The contrarian point is that investors may overrate near-term earnings impact and underappreciate the 2028+ option value. This is not a 1-3 month revenue event; it is a 6-18 month signal that merchant-style behind-the-meter generation is becoming a financing template for large loads. The key falsifiers are project slippage, permitting pushback in West Texas, or evidence that utility interconnection queues clear faster than expected, which would compress the premium on private power solutions.

Key Decisions for Investors

• Long GEV on pullbacks; use a 6-12 month horizon. This is a multi-cycle backlog validation story rather than a one-day catalyst, with upside if management uses the deal to re-rate the growth duration of turbine/service demand.
• Small tactical long NGG ADR only if you want exposure to the optionality embedded in data-center power origination. Size modestly: the investment is strategically useful but too small to move near-term earnings, so the trade is about multiple support, not EPS.
• Do not chase MSFT on this announcement. Treat it as a watch item for future capex disclosure; if Azure capex or lease commitments re-accelerate without a corresponding revenue inflection, the market may eventually penalize FCF conversion.
• Relative-value idea: long GEV / short XLU for 3-6 months. The thesis is that power equipment and execution capacity outperform regulated utility beta as 'speed-to-power' scarcity gets priced in.
• Alert trigger: if the project misses 2028 first power or if turbine/EPC lead times extend, fade the whole theme. That would indicate the market is paying up for a story that cannot actually scale on time.