Saturday, October 24, 2026
AI Infrastructure · News & Analysis
Home › Power & Energy › Report
Power & Energy · Report

Data Center Frontier reports Big Tech is turning to power from nuclear reactors that are already running.

Buying output from existing reactors gives hyperscalers faster nuclear-backed supply than waiting for new builds or small modular reactors.
Trade pressSlicast · October 23, 2026 at 16:36 UTC · US · Source: Data Center Frontier
importance 50

The next wave of nuclear power for AI data centers may come not from new reactors, but from generating stations that have been operating for decades.

In a pair of agreements announced just a week apart, Google and Amazon committed to long-term electricity purchases supporting approximately 1,080 MW of additional nuclear generating capacity from Constellation Energy's existing fleet. The transactions follow earlier commitments by Meta to support reactor upgrades at operating nuclear facilities in Ohio, Pennsylvania and Illinois, establishing a broader commercial model in which hyperscale technology companies help finance increases in nuclear output.

Under its October 6 agreement with Google, Constellation plans to invest more than $4.3 billion in equipment and technology upgrades at 11 operating reactor units across Illinois, Pennsylvania and New Jersey. The 20-year power purchase agreement supports 890 MW of new generating capacity, with the first upgrades expected to begin delivering additional electricity in 2028. A separate 15-year agreement covers 2,700 MW of existing nuclear generation, underscoring the distinction between securing today's electricity supply and financing additional capacity. Amazon's September 30 agreement similarly supports 190 MW of new generation at Maryland's Calvert Cliffs nuclear station, alongside purchases of existing output. Together, the two arrangements underpin more than $7 billion in planned infrastructure investment.

Unlike proposed advanced reactors, these facilities already have operating licenses, grid connections, established workforces and decades of operating experience. Increasing their output avoids some of the cost, permitting and construction challenges associated with building entirely new nuclear plants, although significant upgrades still require extensive engineering, equipment modifications and regulatory approval.

The potential scale is substantial. According to the Nuclear Regulatory Commission's current application schedule, the agency anticipates 31 applications for reactor power uprates through 2032, representing approximately 2,421 MW of potential additional electrical generating capacity. That is roughly equivalent to the output of two large nuclear reactors without constructing new reactor units. However, these figures represent anticipated applications, not approved projects or guaranteed additions to the grid.

For data center developers facing immediate power shortages, the strategy offers no quick fix. Many of the largest upgrades will not be completed until the early 2030s, and additional generation entering a regional transmission system does not necessarily resolve local power-delivery constraints. Nevertheless, the emergence of long-term, hyperscaler-backed financing for nuclear uprates marks an important shift: major technology companies are investing not only in securing existing electricity supplies, but also in expanding the generating capacity available to the grid.

**Google Brings Agentic AI to Nuclear Operations**

Beyond financing additional generating capacity, Google's agreement with Constellation includes an expanded five-year technology partnership to deploy Google Cloud and Gemini Enterprise across the utility's operations.

Constellation plans to use agentic AI workflows for engineering and permitting, power-flow modeling, interconnection planning, equipment monitoring and outage management. The companies also envision using Google Cloud technologies to strengthen security across operational technology networks supporting critical infrastructure.

The arrangement illustrates how AI is beginning to influence both sides of the electricity equation: driving demand for additional generation while providing tools intended to accelerate infrastructure development and improve operating efficiency. Those benefits remain prospective, however. Constellation has not demonstrated measurable improvements from the new deployment, and AI-assisted planning does not eliminate the engineering, safety and regulatory requirements governing nuclear uprates.

**Inside a Nuclear Uprate: What Actually Changes?**

Increasing the output of an operating nuclear plant is not simply a matter of turning up the reactor. Depending on the scope of the project, a power uprate can require changes to instrumentation, control systems, turbines, generators, cooling equipment and other infrastructure, along with engineering analysis demonstrating that the plant can operate safely at higher power levels.

The approach is well established. According to the Nuclear Regulatory Commission, uprates approved over the past 20 years have added approximately 6 GW of electrical generating capacity to the U.S. grid, equivalent to the output of roughly six large reactors.

The NRC recognizes three principal categories of reactor power uprates:

- **Measurement uncertainty recapture uprates** generally increase licensed reactor power by less than 2%, using improved instrumentation and more accurate measurement of operating conditions, particularly feedwater flow.

- **Stretch uprates** typically increase licensed power by approximately 2% to 7%, taking advantage of operating margins already incorporated into a plant's design. These projects generally require limited equipment modifications.

- **Extended power uprates** involve larger increases, sometimes reaching 20%, and can require extensive modifications to high-pressure turbines, condensate pumps, motors, generators, transformers and other plant systems.

The distinction matters because nuclear reactors are licensed according to their thermal power output, measured in megawatts thermal (MWt), rather than simply the electricity ultimately delivered to the grid. Increasing reactor thermal power can affect steam production, heat removal, cooling requirements and the operating limits of equipment throughout the facility.

Improvements to turbine efficiency and other conventional generating equipment can also increase electrical output without necessarily raising reactor thermal power. Those gains should be distinguished from formal reactor power uprates, which require NRC approval to increase the licensed thermal operating limit.

The engineering implications become particularly significant for larger uprates. Equipment replacements must be coordinated with scheduled refueling outages, while safety analyses must demonstrate that critical plant systems can accommodate the proposed operating conditions. For utilities, that creates a complex project-management challenge involving component procurement, outage scheduling, engineering and regulatory review.

The advantage is that much of the most expensive infrastructure already exists. Reactors, turbine buildings, cooling systems, transmission connections and supporting facilities can often be modernized rather than constructed from scratch. But the work is neither automatic nor risk-free, and the size of the proposed capacity increase does not necessarily indicate how quickly or easily it can be delivered.

**Amazon's Calvert Cliffs Pivot: From On-Site Development to Grid Supply**

Amazon's September 30 agreement with Constellation illustrates a different approach to securing nuclear power for AI infrastructure. Following the withdrawal of a proposed data center development at Maryland's Calvert Cliffs Clean Energy Center amid local opposition, the companies reached a 20-year electricity agreement supporting more than $3 billion in upgrades to the existing nuclear facility.

The agreement covers 690 MW of nuclear power, including approximately 190 MW of new generating capacity expected to come online between 2030 and 2032. That represents an increase of roughly 10.6% over Calvert Cliffs' current 1,790 MW capacity. The companies also signed a related retail electricity supply agreement supporting Amazon's operations across the 13-state PJM market.

Beyond the additional generation, Amazon's long-term commitment provides revenue certainty for Constellation to pursue another 20 years of licensed operation at Calvert Cliffs. The plant's two reactors currently have licenses expiring in 2034 and 2036, respectively, making the agreement significant for preserving existing nuclear capacity as well as expanding output.

Under the revised arrangement, all electricity generated at Calvert Cliffs will continue flowing into the PJM regional grid rather than being dedicated to an on-site data center. The model avoids some of the regulatory and community challenges associated with directly colocating large computing facilities at nuclear plants, while allowing Amazon to support incremental generation through a long-term commercial commitment. It does not, however, eliminate the transmission and local distribution constraints affecting where data center capacity can actually be deployed.

**Meta Backs 433 MW of Nuclear Uprates Across PJM**

Meta's broader nuclear power strategy includes 20-year agreements with Vistra to support

Read the original
Data Center Frontier reports Big Tech is… · Slicast