Google’s Constellation nuclear deal is not a plan to build one new 890 MW reactor. It is a long-term agreement designed to unlock about 890 MW of additional nuclear capacity by upgrading existing Constellation reactors already connected to the PJM grid.
That distinction matters. Nuclear uprates can add firm power without waiting for a completely new plant, but 890 MW is a capacity number—not a guarantee that exactly 890 MW will flow every hour of every year.
Industry / Technology: What Is a Nuclear Uprate?
A nuclear uprate increases the maximum electrical output of an operating plant. The work can include turbine improvements, steam-generator upgrades, digital controls, plant optimization, and other equipment changes that improve thermal or electrical efficiency.
Because the reactors, transmission connections, operating staff, and most plant infrastructure already exist, uprates can add capacity faster than a greenfield nuclear project. They still require engineering, outages, equipment installation, regulatory approvals, and plant-specific work before the extra megawatts are available.
Project / Contract: What the Google–Constellation Agreement Adds
The agreement announced in October 2026 targets about 890 MW of incremental nuclear capacity across Constellation units in Illinois, Pennsylvania, and New Jersey. Constellation says the first uprate is expected to come online in 2028, with additional projects following over time.
The broader commercial package includes long-term power commitments that support the economics of the existing fleet. For Google, the strategic value is not only clean-energy accounting. The company is helping finance physical capacity that can serve a grid facing rapid load growth from data centers, manufacturing, and electrification.
This also differs from a standard power purchase agreement tied only to current generation. Part of the economic logic is that contracted demand helps justify capital spending that can create additional output from assets that already exist.
How Much Energy Can 890 MW Produce?
Capacity and energy are different. If 890 MW were available at full output for every hour of a 365-day year, the theoretical energy would be:
890 MW × 8,760 hours = 7,796,400 MWh per year
That is about 7.8 TWh per year. But actual annual output will be lower because plants have refueling outages, maintenance periods, operating constraints, and other downtime.
| Capacity factor assumption | Annual energy from 890 MW |
|---|---|
| 100% | 7.80 TWh |
| 95% | 7.41 TWh |
| 90% | 7.02 TWh |
The table is a simple energy conversion, not a forecast of what these specific uprates will generate.
Why PJM Is an Important Part of the Story
PJM is one of the largest U.S. power markets and is facing unusually strong load growth. Data-center development is a major reason. That makes incremental firm capacity especially valuable because new demand is arriving faster than many new generation projects can be built and connected.
An uprate does not remove the need for transmission expansion or new generation elsewhere in PJM. It simply adds another tool: extract more dependable output from existing nuclear sites while other grid investments continue.
Companies & Markets: Why This Matters to Google and Constellation
For Google, growing data-center demand creates a problem that renewable-energy certificates alone cannot solve: the grid needs enough physical generation capacity at the right times and locations. Firm nuclear output can complement wind and solar because it is not weather-dependent.
For Constellation, a long-term corporate customer can improve revenue visibility for capital-intensive plant upgrades. That does not mean every dollar of the agreement becomes profit. Uprates require engineering work, outage coordination, equipment procurement, licensing, and ongoing operating costs.
→ Related Guide: Why Are Nuclear Plants Restarting for AI Data Centers?
Market Reaction
The most important market question is whether the agreement creates durable incremental earnings rather than simply a large headline megawatt figure. Investors need to separate three things: existing nuclear output under contract, new output created by uprates, and the capital spending required to deliver that new output.
The 890 MW figure proves the potential scale of the capacity addition. It does not by itself tell us the project margin, return on invested capital, or how much of the economics belongs to each plant.
What the Price Chart Shows—and What It Cannot Tell Us
A stock-price move after a nuclear agreement can show that the market changed its expectations for Constellation, nuclear demand, or long-term contracting. It cannot prove that the deal alone caused the move.
Constellation shares are also influenced by power prices, capacity-market expectations, nuclear operating performance, policy, interest rates, and broader market conditions. A chart can show reaction and volatility; it cannot isolate the financial value of 890 MW without the contract economics.
Business Reality / Limitations
- The 890 MW is incremental capacity, not annual energy.
- Actual MWh depend on plant availability and the timing of each uprate.
- New capacity arrives over time rather than on the announcement date.
- Uprates still require capital investment and plant outages.
- Contract terms do not automatically reveal project-level profitability.
There is also a grid-location issue. The capacity is being added inside PJM, which is valuable because PJM is dealing with rapidly rising load forecasts. But the benefit depends on where generation, transmission constraints, and new load are located inside the region.
→ Related Guide: AI Data Center Power Consumption in the U.S.
What to Watch Next
- Which reactor uprates receive approvals and on what schedule
- Whether the first 2028 delivery target remains on time
- How much capital Constellation assigns to the upgrade program
- Whether other hyperscalers sign similar nuclear-uprate agreements
- How PJM capacity prices and load forecasts change as new data centers connect
Bottom Line
The Google–Constellation agreement is important because it turns existing nuclear infrastructure into a source of new grid capacity. If the full 890 MW is delivered, the theoretical maximum energy equivalent is about 7.8 TWh per year before accounting for outages and operating conditions.
For the market, the deeper question is not whether 890 MW sounds large. It is how quickly those megawatts arrive, what they cost to create, and whether long-term contracted revenue produces attractive returns for the plants that deliver them.
About the author
I have practical experience in electrical construction and electrical project coordination in South Korea. For U.S.-focused guides, I research official codes, government data, utility documents, and manufacturer specifications to explain electrical and energy topics as accurately and practically as possible.
Sources & Official References
- Constellation Energy — Google and Constellation announce 890 MW nuclear capacity agreement
- Google — Why we’re backing America’s existing nuclear plants
Is there something you’d like to see?
Suggest an electrical topic, practical guide, or calculator you would like to see on JHOTAS-LAB. Your suggestion may help shape future content.
