ENMG Analysis
Google and Constellation Energy have announced a 20-year power purchase agreement supporting Google nuclear uprates at 11 existing reactors in Illinois, Pennsylvania and New Jersey. Constellation expects to invest more than $4.3 billion in the equipment and technology required to add 890 megawatts of generating capacity. The improvements are expected to add 890 megawatts of incremental capacity to the PJM Interconnection grid, with the first uprate targeted for delivery in 2028.
The agreement combines a long-term commercial commitment with physical improvements at operating reactors. That distinction matters. The project is not a proposal to build 890 megawatts of new nuclear generation from the ground up. Instead, Constellation plans to increase the authorized output of existing units by modernizing equipment and improving thermal and electrical efficiency.
The companies said the collaboration is expected to sustain approximately 4,400 existing jobs, create about 7,200 construction jobs and expand the amount of electricity available across the PJM footprint. The companies also described the 890 MW as comparable in capacity to a large conventional reactor or three or more small modular reactors, although that comparison is about output capability rather than a like-for-like technology substitution.
What the 890 MW represents
A megawatt is a measure of power capacity, or the rate at which a facility can produce electricity. It is not the same as a megawatt-hour, which measures the amount of electricity generated over time.
The 890 MW figure therefore does not mean that the plants will produce only 890 MWh, or that the grid will receive a fixed quantity of electricity. The actual energy generated will depend on operating hours, maintenance outages, market dispatch and the capacity factor of the upgraded units. The agreement’s central claim is that the projects will increase the reactors’ licensed power levels and their electrical generating capability. Each plant will require the appropriate NRC authorization, while PJM must evaluate how the additional electrical capacity will be integrated and recognized within the regional grid.
A nuclear uprate is a licensed increase in a reactor’s authorized power output. Depending on the type and size of the uprate, the increase may be achieved through more precise reactor-power measurements, changes to instrumentation and controls, or major equipment modifications involving turbines, generators, condensers, pumps, transformers and other plant systems. The operator must also demonstrate that safety margins, cooling systems, electrical equipment and other plant systems can support the higher output.
That work differs from greenfield construction. An existing nuclear plant already has an operating site, trained workforce, fuel-handling systems, transmission connections and an established relationship with regulators and grid authorities. Those advantages can shorten the path to additional capacity, although they do not eliminate technical reviews, outage planning or licensing requirements.
Why existing plants can avoid some interconnection delays
New generators entering the U.S. electricity market often face lengthy interconnection queues. A proposed facility may need transmission studies, system-impact analysis, cost allocation and construction of new network infrastructure before it can deliver power.
An uprate at an operating nuclear plant generally starts from a different position. The plant already holds an interconnection agreement and has existing transmission service. Existing interconnection rights do not necessarily extend automatically to the additional megawatts, and transmission upgrades may still be required if PJM identifies reliability or equipment limitations. Incremental output may therefore face a shorter path than a new generator seeking a first-time grid connection.
That does not mean the 890 MW will automatically be accepted by PJM without further analysis. The regional grid operator must still assess transmission impacts, equipment limits, reliability requirements and the capacity value assigned to the upgraded units. Constellation must also complete plant-specific engineering work and obtain any required license amendments from the Nuclear Regulatory Commission.
The commercial schedule remains partly uncertain. The first uprate is expected in 2028, but the companies have not publicly provided a complete unit-by-unit schedule, the allocation of the 890 MW among the 11 reactors or the precise dollar allocation for each project. Those details will be important for assessing execution risk and the timing of the grid benefit.
What the Google agreement does financially
A power purchase agreement, or PPA, is a contract that provides revenue certainty by establishing terms under which a buyer pays for electricity, capacity or associated attributes over a specified period. In this case, Google’s 20-year agreement is intended to support Constellation’s investment in equipment and technology at its existing nuclear fleet.
A PPA does not necessarily mean that electrons from a particular reactor will travel directly to a specific Google facility. Unless a contract includes physical delivery arrangements or identifies a particular resource for delivery, the electricity still enters the interconnected power system and is managed through market and grid operations. The financial commitment and the physical flow of electricity are related but not identical.
The arrangement also includes a separate 15-year energy supply agreement covering an additional 2,700 MW from Constellation’s existing PJM fleet. That agreement is not the same as the 890 MW uprate program. Constellation says the additional contract will provide revenue certainty and help keep existing nuclear resources economically viable without removing them from the grid.
The distinction is important for market analysis. The 890 MW program is intended to add physical generating capability. The 2,700 MW agreement primarily supports the economics of existing capacity. Both may help preserve nuclear generation in PJM, but only the uprates are expected to increase the fleet’s authorized output.
AI demand is changing the contracting model
The Google-Constellation arrangement reflects the growing role of large technology companies in power procurement and infrastructure finance. Google and Constellation also established a strategic framework to evaluate new clean generation, storage and demand-response opportunities across the United States, with near-term attention directed toward PJM.
Constellation selected Google Cloud and Gemini Enterprise under an expanded five-year technology alliance. The stated objective is to use artificial intelligence to reduce infrastructure costs, improve plant and project performance and help safeguard the grid.
The broader issue is the speed at which data center demand is developing. How AI data center power demand is reshaping the grid has become a central question for utilities, regional transmission organizations and regulators because load growth can arrive faster than new generation and transmission projects can be built.
Nuclear uprates are one response, but they are not a complete solution. PJM will still need a combination of existing gas, coal, hydroelectric, wind, solar, battery storage, demand response, transmission investment and potentially new nuclear construction to manage changing load patterns. The value of each resource will depend on reliability requirements, fuel availability, weather conditions, market rules and the timing of new demand.
The agreement is also aligned with the goals of the U.S. government’s UPRISE program, which seeks to maximize output from existing nuclear infrastructure through public-private collaboration. ENMG’s coverage of the nuclear capacity expansion strategy and the UPRISE framework examines that policy direction in greater detail.
Oracle’s Point Beach subscription is a different mechanism
Separately, Oracle announced on October 2 a commitment to subscribe to a portion of the electricity generated by the existing Point Beach nuclear plant in Wisconsin. Point Beach is a twin-unit pressurized water reactor facility operated by NextEra Energy on the shore of Lake Michigan near Two Rivers. Its first unit began commercial operation in 1970 and the second in 1973.
Oracle said the subscription is intended to help absorb rising energy costs associated with the plant, reduce pressure on household electricity bills and support reliable, carbon-free power for the Wisconsin grid. The arrangement remains subject to approval by the Public Service Commission of Wisconsin.
Under the announced structure, Oracle would subscribe to a portion of the electricity costs that We Energies incurs through its existing Point Beach purchase-power arrangement. The exact financial treatment, electricity allocation and customer protections will depend on the regulatory filing and the Wisconsin commission’s decision. It would support the economics of an existing plant; it would not add new generating capacity in the way the Constellation uprates are intended to do.
Mahesh Thiagarajan, executive vice president of Oracle Cloud Infrastructure, said the commitment would directly benefit more than one million Wisconsin utility customers. He added that Oracle expects to absorb approximately $300 million in rising energy costs while supporting the reliable power on which the state depends.
The commitment is connected to Oracle’s Project Lighthouse data center development in Port Washington. According to Oracle, the project is expected to generate more than $11 billion for the local economy, create more than 4,000 skilled construction jobs over three years and support approximately 1,000 ongoing operations positions. Oracle has also said it will fully fund Project Lighthouse’s energy costs.
The potential effect on ratepayers will depend on the final subscription structure, state approval and how costs are allocated among Oracle, the utility and other customers. The announced goal is to reduce pressure on household bills, but the outcome should not be treated as guaranteed before the Wisconsin commission completes its review.
A growing series of technology-sector nuclear agreements
The Google-Constellation announcement follows several related transactions involving large electricity users and nuclear operators. Constellation’s 20-year agreement with Amazon supports an additional 190 MW and a life-extension effort at the Calvert Cliffs plant in Maryland.
Google has also agreed to support uprates at Georgia Power’s share of the Vogtle and Hatch plants, adding approximately 96 MW. Its agreement with Finland’s Fortum covers up to half of the Loviisa plant’s capacity and provides financial support for a lifetime extension through 2050. Google-supported plans also call for the restart of NextEra Energy’s Duane Arnold plant in Iowa.
NucNet reported that the Constellation investment represents one of the larger recent private-sector commitments to increasing output from operating nuclear assets. The series of deals indicates that technology companies are increasingly using long-term power contracts to manage electricity exposure while helping preserve or expand firm generation.
Those contracts also shift risk. Technology companies assume longer-term financial obligations, plant owners gain more predictable revenue, utilities and regulators must evaluate cost allocation, and grid operators must determine how additional capacity fits within reliability and market rules.
What to watch next
The next milestones will be regulatory filings, equipment procurement, outage schedules and PJM’s treatment of the incremental capacity. Analysts and market participants should also watch whether other nuclear operators can replicate the model at plants with suitable equipment margins, available transmission capacity and customers willing to provide long-term revenue support.
At Point Beach, the key issue is the Wisconsin commission’s review of Oracle’s proposed subscription and the final treatment of plant costs. At the federal level, the DOE’s push for large-load interconnection reform at FERC could influence how future data center projects secure power without transferring excessive costs to other customers.
Flexible operating arrangements may also become more important. Flexible interconnection for data center load could allow some large loads to connect under conditions that account for system constraints, though such arrangements would need clear reliability rules and enforceable operating requirements.
The Google-Constellation deal is significant because it pairs private demand with physical improvements at existing nuclear plants. Its success will depend less on the announcement itself than on whether the uprates receive the necessary approvals, stay within budget, arrive on schedule and deliver capacity that PJM can count on. The Oracle-Point Beach proposal addresses a different problem: preserving the economics of an existing plant and assigning more of its costs to a large new customer. Together, the agreements show how the next phase of nuclear development may involve not only reactor construction, but also contracts, equipment modernization, regulatory oversight and direct participation by the companies driving electricity demand.
Keep In Touch with Shale Magazine
As the new era of energy unfolds, you can bet we’ll be the boots on the ground to keep you informed. Subscribe to Shale Magazine for sharp insight into the arenas that matter most to your life. And don’t forget to listen to our riveting podcast, The Energy Mixx Radio Show, where our very own Kym Bolado interviews the most extraordinary thought leaders, business innovators, and industry experts of our time.
