Google’s agreement with Constellation Energy marks one of the clearest signs yet that the expansion of artificial intelligence infrastructure is becoming an energy-market story as much as a technology story. Announced Tuesday, October 6, the 20-year agreement will support 890 megawatts of additional nuclear capacity in the PJM power market, as technology companies seek reliable electricity for rapidly expanding data-center operations.

AI Growth Is Increasingly Dependent on Reliable Power

The agreement comes as electricity availability becomes a strategic constraint for the next stage of AI expansion. Constellation will invest more than $4.3 billion to upgrade turbines, steam generators and digital control systems at 11 existing nuclear units in Illinois, Pennsylvania and New Jersey. The first uprate is expected to be delivered in 2028, providing additional firm generation without requiring an entirely new nuclear plant.

For Google, the significance extends beyond securing electricity. The company is effectively linking its AI data-center growth strategy with long-duration energy procurement, reducing exposure to potential power shortages as computing demand rises. The companies also agreed to a separate 15-year energy supply arrangement covering another 2,700 megawatts from Constellation’s existing PJM fleet.

Constellation Emerges as a Key AI Infrastructure Beneficiary

The financial-market reaction underscored the strategic importance of the agreement. Constellation shares rose sharply following the announcement, with Reuters reporting a gain of approximately 12.9% during Tuesday trading. The move came as the broader U.S. market reached fresh highs, with the S&P 500 and Nasdaq supported by easing Treasury yields and expectations for strong corporate earnings.

The development also reinforces a broader trend among hyperscalers. Amazon reached a separate 20-year agreement with Constellation only days earlier, supporting more than $3 billion of infrastructure investment at Maryland’s Calvert Cliffs nuclear facility. These transactions suggest that long-term corporate power contracts could increasingly provide the revenue visibility required to justify investment in existing nuclear assets.

Why the Nuclear-AI Connection Matters for Global Investors

For investors, the important development is not simply another corporate energy contract but the emergence of a wider AI power ecosystem. Nuclear operators, utilities, grid infrastructure companies and technology providers could increasingly become interconnected as data-center electricity consumption expands. Google and Constellation will also use Google Cloud and Gemini Enterprise in a five-year technology partnership focused on grid planning, asset monitoring and cybersecurity.

For Israeli investors, the trend is relevant because the economics of AI infrastructure are increasingly extending beyond semiconductor companies and cloud platforms into energy infrastructure and electricity reliability. However, the investment case remains subject to substantial uncertainties, including regulatory approvals, construction and operating costs, electricity-market pricing, interest rates and the pace at which AI-related demand ultimately materializes.

Going forward, markets are likely to monitor whether similar long-term agreements spread across the U.S. power sector and whether nuclear capacity can be expanded quickly enough to match data-center demand. The key balance will be between AI-driven electricity growth and the cost of supplying reliable power. Higher infrastructure spending could strengthen the long-term energy ecosystem, but elevated capital requirements, grid constraints, regulatory risks and changing technology demand could also pressure returns. For global asset allocators, the Google-Constellation agreement therefore represents an important early signal of how the AI investment cycle may increasingly intersect with the economics of power generation.