Summary

  • Uruguay and Solaria have signed a strategic agreement for up to 500 MW of data centres, developed in 200 MW and 300 MW phases. Solaria is to provide land, infrastructure and services and develop the connection to the country’s 500 kV transmission grid.
  • The government’s US$5.7 billion estimate covers the whole project, not Solaria’s investment or revenue. A proposed long-term solar purchase arrangement with UTE is another part of the plan, not proof of a fully contracted campus.

Solaria is offering data-centre developers something less glamorous than computing power: somewhere to plug it in. Its agreement with Uruguay makes that proposition unusually tangible. Alongside the land comes a planned transmission connection; alongside the campus, solar generation; alongside both, a free zone that Solaria would develop and run.

That is a substantial role for a solar developer. It is not the same as being the company that fills the buildings with servers. Nor does it mean the US$5.7 billion attached to the announcement belongs on Solaria’s order book.

The distinction matters because several businesses sit inside this one project. Preparing a site, selling electricity and operating a data centre have different customers and different payment dates. A large headline can obscure those differences just when readers most need to see them.

There is already money in preparing the ground

Powered Land, Solaria’s name for supplying developers with land and a secured grid connection, is not simply a label invented for Uruguay. In its first-half results, the company reported €99.6 million of data-centre infrastructure sales for 2026. It said revenue came mainly from reaching development milestones under that model.

That helps explain the attraction of the new agreement. The company can have a commercial role before the complete campus is operating: it prepares something another business needs. But the financial result belongs to the existing business. The half-year filing is a corporate financial record, not a delivery certificate for this new campus. It does not establish a tenant, a construction contract or financing for the project announced in October.

Solaria’s 15 October filing with the CNMV describes its proposed responsibilities in Uruguay. It would develop and operate the free zone, manage land and provide infrastructure and services to data-centre operators. It would also develop the connection to the national 500 kV transmission network. Those are commitments to develop the site’s operating conditions, not a report of completed facilities.

One total, several bills

The government’s announcement puts the overall investment at US$5.7 billion: US$5.5 billion for data centres and US$200 million for 350 MW of solar generation. It does not divide the bill among Solaria, operators, lenders and other investors. Without that allocation, the total reveals the intended scale, not one company’s capital exposure or future sales.

The initial phase envisages 200 MW of data centres and 350 MW of solar. A later phase would add another 300 MW of data centres, with possible further solar development. The government places the solar installations in Florida, the first data-centre phase in Canelones and the later phase in San José, all in Uruguay.

It forecasts completion of the first phase in 2029 or 2030, 3,100 construction jobs and 1,500 permanent jobs when the whole project is operating. These are official expectations, not present employment. They give the announcement political significance, but they do not settle who must deliver each component or who pays if it arrives late.

Forty years does not tell you the delivery hour

UTE, the state electricity company, occupies a different position in the deal. Solaria says solar plants would be built on land provided by UTE, which would purchase their output. The strategic agreement envisages a power purchase agreement lasting up to 40 years.

A prospective buyer over that length of time is important to a generation project. Yet a duration is not a price, a delivery start or a settlement arrangement. The published summaries do not supply those final terms. They also do not explain how UTE’s purchase would fit the arrival of demand from successive data-centre phases.

The capacity figures cannot fill that gap. A reader might see 350 MW of solar beside 200 MW of initial data centres and assume a comfortable surplus. But solar nameplate capacity does not promise continuous output, and the data-centre figure is not an hourly consumption schedule. The relevant supply and balancing arrangements would have to explain how the two work together. The summaries do not establish either that supply is sufficient or that it is inadequate.

The next useful announcement would name a customer

The government and company summaries examined here do not name a data-centre operator, construction contractor or project financier. That is a limit of these documents, not proof that no counterparties have been chosen elsewhere. It is also a guide to what would make the next disclosure more useful.

A first operator’s commitment, with reserved capacity and a start date, would connect the site plan to a user. A delivery milestone would show what Solaria must finish to earn a payment. An allocation of connection costs would show who is funding the route to the grid. Conditions for the extra 300 MW would help distinguish a later ambition from the next committed stage.

None of this requires dismissing the strategic agreement. It requires reading it for what it establishes. Solaria has an existing infrastructure business and an announced role stretching from land to grid connection and free-zone operation. The open question is how that role becomes a set of payable, deliverable obligations in Uruguay.

The first server need not be running for Solaria to have built something valuable. But a prepared site needs more than a large project number: it needs an identifiable user and a clear handover. That is where this announcement will become a business readers can measure.