Uruguay’s 90%+ renewable grid: what a company actually checks before it sets up

The phrase «Uruguay, 99% renewable» fits on a slide, sounds like certainty, and skips the awkward part: where the number comes from and how far it moves from year to year. For a company weighing whether to site a plant, a logistics hub or a data center, the useful reading is the opposite of the slogan. The round figure describes a year of rain; the siting decision lasts a decade. What matters is not whether the grid hit 99% once, but the floor it drops to when water is scarce, how diversified the system is to hold that floor, and what sits behind it when diversification is not enough.

The 99% is real, and it measures a year of rain

The number has a primary source. According to the National Energy Balance (BEN) of the Ministry of Industry, Energy and Mining (MIEM), in its final version presented in August 2025, the 2024 electricity generation matrix reached 99% renewable, split as hydro 43%, wind 28%, biomass 24% and solar 4% (as-of 2024).

Two details are worth carrying forward. First, cite the final balance, not the April 2025 preliminary estimate, which gave hydro 42%, biomass 26% and a different emission factor, and was later revised. Second, 99% is a snapshot of good hydrology, not a constant. The defensible claim for a ten-year plan is not «I am guaranteed 99%» but «I have a grid that stays in a 90–99% renewable band and absorbs hydrological variability with a bounded fossil margin». That is a more modest promise, and a far more solid one. To see it, look at the dry years.

The floor of the range

That 91–92% is the real floor, and it is the number a prudent company sizes its decision against. The 2025 figure is provisional and should be treated as such until the final BEN is published; the 2024 row is the only recent one that is settled.

What is telling is how the system held the floor when the water ran short. In 2023 hydroelectric generation fell to roughly 3,429 GWh (MIEM basis; UTE-derived estimates circulate at ≈3,431 GWh — the small gap is a difference of source, not of fact), its lowest point in two decades. The gap was not filled by fossil fuel, whose share kept falling for a third consecutive year, but by a combination of wind — which that year became the leading generation source — biomass, and imports from neighbours, which hit a fourteen-year high of 1,398 GWh, around 12% of annual demand (ADME, as-of 2023). A dry year sank the hydro and the system answered with other levers without dropping below 90%.

What holds the floor up: the mix, and the insurance policy

A common misconception is that a «green» country is green thanks to state-of-the-art sun. In Uruguay solar is still marginal — 3–4% of generation — and what holds the grid up is a less glamorous trio: hydro, wind and biomass. The robustness comes from that diversity, not from a star technology.

According to the MIEM/DNE installed-capacity series published in the National Statistics Institute’s Yearbook, the system totalled 5,298.5 MW in 2024: hydro 1,538.0 MW, wind 1,516.5 MW, biomass 731.2 MW, photovoltaic 336.2 MW (up from 300.6 MW in 2023) and fossil 1,176.7 MW (as-of 2024). Those 1,500-plus MW of wind are what set Uruguay apart from regional grids that rest almost entirely on large dams.

Two caveats on that table, both of which matter if you are modelling supply rather than quoting a brochure:

• Hydro is partly binational. Salto Grande is shared with Argentina, and its output is split. The MIEM/DNE series appears to count Uruguay’s share rather than the full plant — confirm which convention a given figure uses before you build a supply model on it.
• The 1,176.7 MW of fossil capacity is the reason the floor holds. It contributes almost nothing to annual generation, but it is firm capacity: dispatchable, weather-independent, and the thing that stands between a bad hydrological year and unserved demand. Alongside it sit the interconnections — notably the 500 MW converter station at Melo linking to Brazil, plus the Argentine links — which in 2023 did more work than the thermal fleet. A company reading only the renewable percentage will miss both, and they are what make the percentage survivable.

Wind and solar are also complementary in time: solar concentrates around midday and in summer, while wind’s seasonal and hourly profile runs differently — stronger in winter, and present through the night. Wind is not a nighttime-only resource, but its shape offsets solar’s, which strengthens system resilience.

Two ways of counting the same grid

Different percentages circulate for the same year, and the reason is methodological. The MIEM counts all national generation, including self-producers — in particular the biomass that pulp plants burn and consume on site without injecting it into the grid. The Electricity Market Administration (ADME) counts what is actually delivered to the National Interconnected System (SIN). The same year therefore reads differently:

Neither is wrong; they answer different questions. If you are sizing a grid connection or a supply contract, ADME’s basis is the relevant one. If you are looking at the country’s energy footprint, MIEM’s is.

The practical consequence is a trap worth naming. Take the preliminary 2024 biomass figure of 26% on the national basis, set it against 14% for 2025 on the SIN basis, and it looks like a collapse. It is a mirage of denominators, not a shutting of boilers: measured on the same SIN basis, biomass in fact rose slightly, from 13% in 2024 to 14% in 2025 (2025 SIN mix, preliminary: hydro 46%, wind 34%, biomass 14%, solar 4%). Never subtract one basis from the other.

One clarification on the headline numbers themselves, since they are sometimes misdescribed: the 99% for 2024 and the 98% for 2025 both come from DNE balances on the same national basis and are directly comparable. The difference between them is hydrology and the final-versus-preliminary status of the 2025 figure — not a change of methodology.

Scope 2: the emission factor and how it is credited

For a corporation, a 90%+ renewable grid is an accounting input, not a reputational ornament. The number that matters for reporting is not the «% renewable» but the grid’s carbon intensity. According to the MIEM, the SIN emission factor was 6 tCO₂/GWh in 2024 — a 90% reduction versus 2023 and the lowest value in twenty years — and 12 tCO₂/GWh in the preliminary 2025 balance (as-of 2024 final and 2025 preliminary). That is the metric that goes into an emissions balance.

How the advantage is credited depends on the Scope 2 method:

• Location-based, using the grid’s average emission factor: connecting to the Uruguayan grid lowers reported purchased-energy emissions automatically, with no panel of your own and no additional instrument.
• Market-based, the method initiatives such as RE100 require: a clean grid alone is not enough. Consumption must be backed by contractual energy attributes — renewable energy certificates or guarantees of origin, or power purchase agreements. Whether and how those instruments are available in Uruguay should be verified case by case rather than assumed to replicate European schemes.

One further caveat that follows directly from the counting logic above: the SIN factor is low partly because biomass combustion is treated as carbon-neutral under standard biogenic-carbon convention, with those emissions reported separately rather than in the grid factor. Given that biomass supplies roughly an eighth of SIN generation and around a quarter of national generation, a company with strict internal reporting rules should know that this convention is doing real work in the number.

Solar, the indicative plan, and how to read announcements

If you are looking for where the opportunity and the risk of the coming years sit, look at that 3–4% solar: the smallest component of the matrix, and the one the country set out to grow. Total installed photovoltaic capacity reached 336.2 MW in 2024 — a figure that covers parks plus self-consumption, and should not be confused with distributed generation alone, which is a much smaller fraction.

There is a medium-term plan, and its status matters as much as its contents. The MIEM’s Indicative Plan for the Expansion of Electricity Generation 2024–2043 (published February 2025) is a scenario study, not a procurement commitment: it identifies what would be economically convenient under different demand paths, finding solar attractive from 2025 and wind from 2030. Figures circulating from it — on the order of 1,130–1,375 MW of solar and 2,100–2,420 MW of wind over twenty years, depending on the scenario — should be read as cumulative new additions on top of the existing base, not as a total-capacity target, and should be checked against the document itself before being used in a model. Reading «reach 1,100 MW by 2040» and «add 1,130–1,375 MW by 2043» as contradictory targets is a mistake; it is the same trajectory seen at two different years, in a study that commits no one.

The same discipline applies to promises. The investment-promotion agency Uruguay XXI positions the country as a benchmark destination for renewable-energy investment and repeats the 99% figure (promotional material, as-of 2025–2026); take that for what it is and contrast it with the MIEM data above, which are what do or do not sustain it.

The most-cited corporate case makes the point without rhetoric. Google is building — as of mid-2026, still in mechanical and electrical installation, not operating — a data center at the Parque de las Ciencias in Canelones, an investment of more than USD 850 million. At the construction launch, a Google Cloud representative described Uruguay as using renewable energy in more than 90% of its energy matrix, as reported in the Canelones government’s account of the event. Two points of precision: the formulation was «more than 90%», not a higher figure sometimes attributed to it in the press; and it was said by a company representative at a ceremony, which is not the same as a figure published in Google’s own corporate communication. Announcements of grid-scale storage and of corporate contracting schemes deserve the same treatment: flag them as announced, not accomplished. The gap between «announced» and «operational» is exactly what separates an informed decision from a bet on a headline.

What the percentages do not tell you

Three things a siting decision needs that no renewable share will answer, and that should be requested directly from UTE, ADME and the MIEM:

1. Reliability metrics. A 99% renewable grid is not the same as a 99.99% available one. Ask for SAIDI/SAIFI at the relevant voltage level and for the local network’s constraints.
2. Price and contracting. The large-consumer tariff structure, the terms on which energy can be bought under term contracts rather than at the spot price, and how those interact with any market-based Scope 2 claim.
3. Connection. Available capacity at the intended point of connection, the queue, and the works the connection would require. Diversity of generation does not help if the substation cannot take the load.

Conclusions

Uruguay offers something uncommon: a grid that stays between 90% and 99% renewable year after year — with a floor of 91–92% even in severe drought — held up not by a star technology but by a diversified mix of hydro, wind and biomass, with more than 1,500 MW of wind as its regional differential and roughly 1,177 MW of thermal capacity plus regional interconnection as the insurance behind it. Its carbon intensity is very low — 6 tCO₂/GWh in 2024 — which lowers an operation’s location-based Scope 2 simply by connecting, and it already functions as a location criterion for intensive consumers.

The prudent approach is to size the decision against the floor rather than the headline, to know which basis each number is counted on, and to verify the volatile ones — annual percentages, emission factor, capacity plans — against the official MIEM and ADME publications before signing.

Verification checklist

Sources

MIEM/DNE — National Energy Balance (2022, 2023, final 2024) and Preliminary Balance 2025: generation matrix by source, SIN emission factor. ADME — Annual Report (2023, 2024): generation delivered to the SIN. INE — National Statistics Yearbook, with MIEM/DNE data: installed capacity. MIEM — Indicative Plan for the Expansion of Electricity Generation 2024–2043. Presidency of Uruguay — presentation of the BEN 2024. As context: Google’s corporate blog and the Canelones government’s account of the construction launch; Uruguay XXI; the GHG Protocol Scope 2 Guidance.

Generation figures are volatile and depend on the year’s hydrology; verify current data with the MIEM and the ADME before making investment decisions.