Amazon’s 20-Year Nuclear Deal Shows How Large Corporate Buyers Can Help Finance Nuclear Investment

Amazon and Constellation have signed a long-term agreement supporting more than US$3 billion in investment at Maryland’s Calvert Cliffs nuclear plant, including a 190 MW uprate. For Africa, the deal raises an important financing question: could mines, data centres, industrial parks and other large electricity users eventually become anchor customers for nuclear projects?

By: NuclearAfrica
5 October 2026

Amazon and Constellation Energy have signed a 20-year power agreement that will support continued investment and additional generating capacity at the Calvert Cliffs nuclear power plant in Maryland, offering another example of how large corporate electricity consumers are beginning to influence the economics of nuclear energy.

The agreement, announced on 30 September, covers 690 MW of power and includes support for approximately 190 MW of additional nuclear generating capacity through upgrades to the existing plant. Constellation says the arrangement will enable more than US$3 billion in infrastructure investment, with the additional capacity expected to come online between 2030 and 2032.

The importance of the agreement extends beyond Amazon’s electricity procurement. Constellation says the long-term revenue certainty provided by the arrangement will also support its effort to seek another 20 years of operation for Calvert Cliffs and provide a foundation for considering additional nuclear development at the site.

Amazon separately says its commitment will help preserve the existing plant, enable the 190 MW expansion and support exploration of further nuclear generation at Calvert Cliffs. The electricity will continue to flow into the regional PJM grid rather than being physically dedicated exclusively to Amazon facilities.

This is an important distinction. The deal is not the construction of a private reactor feeding only Amazon data centres. It is a long-term commercial arrangement under which a major electricity consumer provides the kind of predictable demand and revenue that can justify substantial investment in an existing nuclear asset.

For African nuclear programmes, that financing logic may ultimately be more significant than the specific US project.

A long-term customer can change the economics of nuclear investment

One of the persistent challenges in nuclear finance is that nuclear plants require substantial capital before they generate revenue. The developer must invest in engineering, equipment, construction, licensing and infrastructure years before the project begins selling electricity.

Lenders and investors therefore want confidence that the electricity will have a reliable buyer once the facility begins operating.

A long-term corporate power agreement can help provide that certainty.

In the Calvert Cliffs case, Amazon is effectively giving Constellation greater confidence that a substantial portion of the plant’s output will have a commercially valuable customer over a long period. Constellation, in turn, can use that revenue visibility to justify investments that increase generation and support the plant’s continued operation.

This does not remove construction, regulatory or operating risk. But it addresses one of the central questions confronting any electricity project: who will buy the power, for how long and on what commercial terms?

For nuclear projects, where operating lives may extend for 60 years or more, that certainty can be especially valuable.

The agreement supports an existing plant rather than a completely new reactor

The context should nevertheless be kept clear.

Calvert Cliffs is an operating nuclear facility, not a greenfield reactor project. Its two units currently have a combined generating capacity of about 1,790 MW and supply enough electricity for the equivalent of more than 1.3 million homes.

The proposed 190 MW increase represents an uprate of roughly 10% rather than construction of an entirely new station. Constellation has separately discussed the possibility of adding as much as 2,000 MW of new advanced nuclear capacity at the site, but that remains an exploration rather than an approved construction project.

The present agreement therefore demonstrates most clearly how large customers can support investment in extending and expanding existing nuclear assets.

Financing a first-of-a-kind greenfield reactor remains considerably more difficult.

Still, the underlying principle is relevant: large electricity consumers can become part of the commercial architecture supporting nuclear investment.

Big technology companies are becoming nuclear market participants

Amazon’s agreement is part of a broader trend in which technology companies are becoming increasingly important participants in nuclear-energy markets.

The rapid expansion of artificial intelligence, cloud computing and data centres is creating demand for large volumes of reliable electricity. Data centres operate continuously and require high levels of power availability, making firm generation particularly valuable.

Nuclear power offers characteristics that are attractive to such consumers: high capacity factors, long operating lives, low operational carbon emissions and the ability to produce electricity around the clock.

The Calvert Cliffs agreement is therefore not simply a sustainability commitment. It reflects the growing commercial importance of dependable power for the digital economy.

Amazon says the arrangement will help manage energy costs across facilities in the PJM market while supporting additional carbon-free generation for the wider grid.

This relationship between nuclear energy and large industrial electricity demand deserves particular attention in Africa.

Africa has its own potential anchor customers

African nuclear-newcomer countries generally focus on the national utility as the expected purchaser of electricity from a future reactor.

That is understandable. Electricity systems are usually organised around national utilities or regulated wholesale markets.

But the emerging international model suggests that governments should also consider whether major industrial consumers could become long-term anchor customers.

Africa has electricity-intensive sectors that could potentially play such a role.

Mining is an obvious example. Mining companies require large quantities of dependable electricity, often over long project lives. Mineral processing, aluminium production, steel, fertiliser manufacture and other heavy industries also depend on reliable power.

Data centres are another increasingly relevant category as Africa’s digital economy expands.

Future green hydrogen, desalination, synthetic-fuel and industrial-processing projects could create additional large loads.

If a nuclear project can secure a combination of utility demand and long-term industrial customers, its revenue structure may become more robust than one relying entirely on a financially weak national electricity distributor.

That does not mean corporations should replace the national electricity system as the primary customer.

It means they could potentially share part of the demand risk.

The off-taker problem is particularly important in Africa

This possibility matters because the financial condition of electricity utilities is one of the major barriers confronting infrastructure investment across parts of Africa.

A power-purchase agreement is valuable only if the institution signing it can reliably pay for the electricity.

Where utilities face high distribution losses, accumulated debt, politically constrained tariffs or weak collection rates, lenders may demand sovereign guarantees before they are willing to finance large projects.

That transfers risk to the government.

If a utility fails to pay, taxpayers may ultimately carry the obligation.

A financially strong industrial customer can therefore add another layer of payment security to a project.

For example, a mine or data centre with substantial revenues and a long-term electricity requirement could contract for part of the output while the national utility purchases the remainder.

Such an arrangement would need careful design because nuclear plants do not operate like small captive generators. Their output is normally integrated into the broader power system.

But the Calvert Cliffs arrangement demonstrates that physical electricity does not necessarily have to flow directly from the reactor to the corporate customer for a long-term commercial relationship to support investment.

This may matter greatly for Ghana

The model deserves particular attention in Ghana.

The country is advancing its nuclear power programme while simultaneously seeking greater industrialisation, expanded mineral processing and development of new energy-intensive industries.

A future Ghanaian nuclear project would almost certainly need a credible long-term electricity-purchase structure before international lenders committed significant construction finance.

Traditionally, that would imply a long-term arrangement involving the national electricity system and government support.

But Ghana could also examine whether a future nuclear project could be linked commercially to new industrial demand.

Mining companies, large manufacturing facilities, industrial parks, aluminium processing, data centres and other high-load customers could potentially enter long-term electricity arrangements that strengthen the project’s revenue base.

This would have to be based on actual electricity demand, not hypothetical projects created simply to justify a reactor.

The sequencing is important.

Industrial demand should strengthen an already credible nuclear business case, not be invented to rescue an oversized project.

Nuclear power could also support industrialisation if the demand comes first

There is a broader development-policy question here.

African nuclear programmes are often justified partly on the basis that reliable electricity will support industrialisation.

That argument is reasonable, but it can become circular.

A government may say it needs a large reactor because industry will require the electricity. At the same time, industrial investors may say they will not build factories until dependable electricity exists.

One way to break that cycle is to develop energy and industrial policy together.

Instead of building generation first and hoping industrial demand arrives later, governments could secure investment commitments from electricity-intensive industries alongside development of the power project.

The result could be a more integrated strategy in which major electricity consumers help provide the demand certainty required for financing while the nuclear project gives those industries confidence that long-term power will be available.

This kind of coordination could be especially valuable for nuclear-newcomer states because of the size and capital intensity of the investment.

Data centres deserve attention, but expectations should remain realistic

The rapid expansion of data centres has attracted particular interest in global nuclear discussions.

Technology companies have the financial capacity to sign long-term agreements and place high value on dependable electricity.

That makes them attractive potential nuclear customers.

However, African governments should avoid assuming that data centres alone will justify large reactor programmes.

The electricity demand of the African data-centre market remains much smaller than in the United States, and facilities are concentrated in a relatively limited number of markets.

A national nuclear strategy should therefore not be built around speculative projections of artificial-intelligence demand.

Data centres may become one component of the business case.

They should not automatically become the business case itself.

The same caution applies to green hydrogen and other emerging industries.

Firm contracts are more valuable than optimistic forecasts.

Corporate power contracts can shift risk, but they do not eliminate it

Long-term corporate agreements are attractive partly because they create revenue certainty.

But the details matter.

A 20-year contract raises questions about electricity pricing, inflation, creditworthiness, termination rights, market changes and what happens if the corporate customer’s electricity requirements decline.

A nuclear project may operate for several decades beyond the duration of the initial commercial agreement.

The developer therefore still needs a strategy for the remainder of the plant’s operating life.

Government must also consider whether large corporate contracts could affect electricity availability or pricing for other consumers.

In countries where electricity access remains constrained, public concern would be legitimate if a large publicly supported generation project appeared to prioritise multinational corporations over households and domestic businesses.

The structure should therefore ensure that corporate participation strengthens the wider electricity system.

The Calvert Cliffs arrangement is relevant in this respect because Amazon emphasises that the electricity continues to flow through the regional grid rather than being physically separated for its exclusive use.

Grid integration remains essential

This point is especially important for African nuclear newcomers.

A long-term industrial customer does not remove the need for a strong national grid.

Even where a mine or data centre contracts for part of the output, the nuclear plant still needs transmission infrastructure, system reserves, frequency control and arrangements for planned and unplanned outages.

A large reactor cannot simply be treated as a dedicated generator connected to one customer.

The project must fit within the wider electricity system.

For Ghana, Kenya and other newcomer states, grid readiness therefore remains one of the most important infrastructure requirements.

Industrial power contracts can improve bankability.

They cannot compensate for an electricity system that is technically unable to absorb the reactor.

Long-term contracts may reduce financing costs

The strongest attraction of a corporate PPA is ultimately financial.

Nuclear projects become more expensive when investors perceive high risk.

If lenders are uncertain about future electricity demand or whether the customer will pay, they demand a higher return.

That increases the cost of capital.

A credible long-term buyer can reduce part of that uncertainty.

The result may be lower financing costs, which can materially affect the eventual price of nuclear electricity.

This matters greatly for Africa because borrowing costs are already one of the continent’s major disadvantages in financing capital-intensive infrastructure.

Reducing revenue risk cannot eliminate sovereign risk, currency risk or construction risk.

But every category of risk that can be reduced potentially improves overall bankability.

Amazon is not taking ownership of the nuclear plant

Another distinction is important.

Corporate involvement in nuclear finance does not necessarily require the company to own or operate a reactor.

Amazon is not becoming the nuclear operator at Calvert Cliffs.

Constellation remains responsible for operating the facility and for meeting regulatory requirements.

The corporate customer is helping provide the commercial support that makes investment more attractive.

That model may be particularly appropriate in Africa.

Mining companies, data-centre operators or manufacturers do not need to become nuclear operators.

They could simply become long-term customers.

Nuclear ownership, regulatory responsibility and operational competence would remain with the properly licensed owner-operator.

This separation allows industrial companies to support project economics without creating unnecessary complications in nuclear governance.

There is also a lesson for existing African nuclear assets

The agreement is relevant not only to newcomer programmes.

South Africa operates Africa’s only commercial nuclear power station, Koeberg, and is considering the future role of nuclear energy in its electricity system.

Large industrial buyers could potentially become more significant participants in supporting life extension, uprates or future nuclear projects across established nuclear markets as electricity-sector structures evolve.

The wider lesson is that nuclear finance need not depend exclusively on government appropriations and traditional utility revenue.

There may be more room for large electricity users to support investment where the commercial structure is transparent and the broader electricity system benefits.

The Microsoft precedent shows this is becoming a pattern

Amazon’s Calvert Cliffs agreement is not an isolated case for Constellation.

The company is also pursuing the restart of the former Three Mile Island Unit 1 in Pennsylvania, now known as the Crane Clean Energy Center, under a 20-year power-purchase agreement with Microsoft. Constellation says that project is supported by the Microsoft agreement and separate federal loan financing, although restart remains subject to regulatory approvals.

Together, these transactions point to an emerging model in which major technology companies provide long-term commercial support for nuclear assets while utilities remain responsible for operating them.

That is worth watching closely.

If the approach proves successful, it may become another mechanism through which nuclear projects secure the predictable revenues required for investment.

Africa should study the model without copying it blindly

The economic circumstances of the United States and African nuclear-newcomer countries are very different.

The US has deep capital markets, established nuclear operators, sophisticated electricity markets, large technology companies and mature regulatory institutions.

Most African newcomer programmes do not yet have those advantages.

Simply copying an American corporate PPA structure would therefore be inappropriate.

But the underlying question is highly relevant.

Who are the credible long-term customers for a future African nuclear plant?

That analysis should go beyond the national utility.

Governments should map future electricity demand across mining, manufacturing, digital infrastructure, transportation, desalination and other energy-intensive sectors.

They should determine which loads are sufficiently large, stable and creditworthy to support long-term contracts.

Those conversations should take place while the nuclear programme is still being developed, not after construction has begun.

Corporate participation should produce wider national benefits

If large industrial customers eventually play a role in supporting African nuclear projects, governments should also ensure that the arrangement contributes to broader national development.

Reliable electricity for a mine or data centre has economic value, but the public benefits should extend further.

Projects should strengthen transmission infrastructure, support local supply chains, develop technical skills, increase tax revenues and improve the resilience of the national power system.

Where substantial government guarantees are involved, the public should be able to see clearly what national benefits justify that exposure.

The principle should be simple.

Private demand can help finance public infrastructure.

But public risk should not be assumed solely to subsidise private electricity consumption.

Conclusion

Amazon’s 20-year agreement with Constellation is significant because it demonstrates how large corporate electricity customers are beginning to become active participants in nuclear investment.

The agreement supports more than US$3 billion in infrastructure investment at Calvert Cliffs, including approximately 190 MW of additional nuclear capacity expected between 2030 and 2032. It may also help provide the revenue certainty needed to pursue another 20 years of operation for the existing plant.

It is not a new reactor construction agreement, and it should not be presented as one.

But the financing principle deserves attention.

Long-term electricity customers can help reduce revenue uncertainty and provide developers with confidence to commit capital.

For African nuclear-newcomer countries, this opens an important area for policy thinking.

Future reactor projects may not have to depend entirely on national utilities and sovereign guarantees for their revenue base. Mines, industrial facilities, data centres and other large users could potentially become anchor customers where their demand is real, long term and financially credible.

The challenge will be integrating those customers without distorting national electricity markets or transferring excessive private risk to taxpayers.

Africa should therefore watch these corporate nuclear agreements carefully.

The important question is not whether Amazon’s exact model can be copied.

It is whether African countries can develop their own arrangements in which industrial demand, reliable electricity and nuclear investment reinforce one another rather than being planned separately.