National Grid invests $1.75bn in US data center power platform

National Grid will invest $1.75 billion for a 35% stake in Joulent to help finance its first 2.67 GW gas-fired West Texas power project (Kilby). The facility is developed via a 50/50 partnership with Chevron and will supply power to a Microsoft-operated data center campus under a 20-year power purchase agreement, targeting first power delivery by 2028. The project has already secured critical equipment (including GE Vernova turbines) and reserved engineering/construction capacity.

Analysis

This is less a single-project story than a signal that AI/data-center power demand is forcing capital toward firm, dispatchable generation. The clearest second-order beneficiary is GE Vernova: every additional large gas-fired build locks in turbine demand, but more importantly it extends the aftermarket/service annuity if this becomes a repeatable platform rather than a one-off. The market may still underestimate how capacity-constrained the supply chain is — turbines, transformers, EPC crews, and interconnects are the real bottlenecks, which should keep pricing power elevated for the best-positioned equipment vendors and contractors.

For Chevron, the value is in option-like exposure to a new power-development vertical with limited balance-sheet intensity relative to upstream cash flow. That said, the earnings impact is not immediate; 2028 delivery means near-term reaction is mostly sentiment, not P&L. The more tradable read-through is that large tech load growth is increasingly being met with gas and utility partnerships, which is structurally supportive for gas midstream names and selective power infrastructure suppliers, while pressuring the thesis that renewables alone can absorb hyperscaler demand at scale.

The contrarian view is that the market will overprice the "AI power" narrative while underpricing execution risk: permitting, interconnection, gas transport, and financing can all delay cash generation by years. If natural gas prices spike or Microsoft re-prioritizes capex, the economics deteriorate quickly. The thesis is falsified if we start seeing cancellations, slower turbine order growth, or evidence that data-center demand is shifting toward nuclear/behind-the-meter alternatives instead of new gas-fired capacity.

Key Decisions for Investors

• Long GEV on pullbacks; use a 6-18 month horizon and treat this as a backlog/service-cycle extension trade. Falsify if order intake or turbine lead-time commentary weakens in the next 1-2 quarters.
• Small long CVX vs. short a renewables-heavy utility/clean-energy basket over 1-3 months. The edge is not commodity beta; it is capital allocation into dispatchable power where firm-load economics are tightening.
• Initiate a watchlist long in U.S. gas infrastructure/producers (e.g., KMI, WMB, EQT) if more hyperscaler power projects are announced. Best entry is on confirmation that additional PPAs or pipeline commitments follow; otherwise keep it as an alert, not a conviction trade.
• Do not chase NGG/NG at current levels solely on this headline; the cash-flow impact is too delayed. Revisit only after financing close, permits, and interconnect approvals are visible.
• Set a catalyst alert on Henry Hub above ~$4.50 and on any permitting setback for the West Texas site; either would compress project economics and reverse the near-term bullish read-through.