Product1 distinct publisher3 min readPublished
A meat plant scrapped tonnes of product on 28 April 2025, then bought 10 MWh of battery. The arithmetic behind that purchase shows how backup is now specified: against annual load, not runtime.
The Product Desk · Product desk
Compiled by The Product DeskSomething wrong?How this is made
Fribin's figures are unusually legible for this kind of purchase, and they show how the sizing was actually done. The plant reports more than 25 GWh a year, most of it refrigeration [4], which is an average continuous draw of roughly 2.85 MW [1]. Two 5 MWh modules make 10 MWh [6], or about three and a half hours at that average [2]. The April 2025 outage lasted several hours [1]. The pack was therefore never meant to carry the whole site through a repeat of the event that paid for it. It was bought to hold the part of the process that spoils, which is exactly what the plant did by hand on the day: refrigeration kept running, every line stopped [2].
That is a different specification exercise from the one diesel taught. Nobody in this story quotes hours of generator fuel; they quote annual consumption and installed megawatt-hours, and then decide which loads are allowed to survive. The national numbers are not in those units. Red Electrica registers power, 193 MW in April 2026 against about 28 MW before the blackout [7], which says nothing about how long anything holds. The mismatch also explains the arithmetic in the December award: 2,400 MW is about 12.4 times the registered figure, but the roughly 80% battery share works out at some 1,920 MW, or close to ten times, which is the comparison the BBC's account uses [5].
At the plant level the price is the interesting number. Around 1.5m euros for 10 MWh is about 150,000 euros per installed megawatt-hour [3]. That price had been available for years and was judged too expensive [5]; what changed was that one morning produced a measured loss in the hundreds of thousands of euros [3], and EU Next Generation funds covered part of the bill [6]. The first module was bought within about two months of the outage, the second a year later [6].
The requirement is also moving up the specification sheet. Sungrow's Alberto Bodegas says commercial and industrial buyers used to ask for partial or full backup and now ask for seamless transfer, with no noticeable interruption, which matters most for hospitals and data centres [10]. Miguel Matias of Self Energy puts the industrial case more bluntly: minutes of cover can be the difference between a stop and damaged machines [12]. Cover for hours and cover for milliseconds are separate products bought for separate reasons, and buyers are now naming both.
The exposure was manufactured deliberately. Firms across Europe have been converting industrial heat and process load from fossil fuels to electricity on EU subsidy, which cuts emissions and makes them more sensitive to a cut [14]. Portugal's Vista Alegre shows the pattern in order: it decided on storage in 2022 with Recovery and Resilience Plan money, and the blackout accelerated an installation that now runs to about 2 MWh at Ilhavo [15]. The subsidy created the dependency; the outage priced it.
Ranked by verification strength, evidence, and original report placement.
On 28 April 2025 Portugal and Spain went dark for several hours after a massive power failure that caused widespread chaos.
Fribin, a Spanish meat processing firm in Binefar, Aragon, was midway through its morning shift when the shutdown halted production lines; emergency systems lacked capacity for continued operation, so refrigeration was prioritised and all lines had to be stopped, according to technology and systems director Andres Altabas.
Many tonnes of meat in the processing stage had to be discarded, with losses of hundreds of thousands of euros, according to Altabas.
Fribin's energy needs amount to more than 25 gigawatt-hours a year, most of it for refrigeration.
Fribin had been considering a battery back-up system for years but had judged it too expensive; Altabas says the blackout catalysed the investment.
Fribin bought a first five megawatt-hour module in June 2025 and ordered a second module of the same storage capacity in early June 2026, amounting to a total investment of around 1.5m euros (1.3m pounds; 1.7m dollars), funded in part by the European Union's Next Generation funds.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Named operators and grid-operator figures, one outlet
The core facts are attributable and specific: on-the-record executives at Fribin, Vista Alegre and Primus Ceramics; module sizes, purchase dates and a total euro figure; Red Electrica capacity numbers; and an itemised IDAE award. That is materially better than anonymous vendor sourcing. It is capped by there being a single publisher and no primary documents, no independent audit of the vendor's demand claims, and an unstated year on the December award.
Real installations at named plants, small installed base
Adoption is observed, not projected: two 5 MWh modules bought at Fribin, about 2 MWh live at Vista Alegre, batteries in place since 2022 and a planned doubling at Primus Ceramics, and system-wide registered battery capacity rising from about 28 MW to 193 MW. The score is held down because 193 MW is a very small base for the Spanish grid and the 2,400 MW award is a pipeline, not deployed capacity.
Growth framing outruns the runtime and base it rests on
Mildly overstated. 'Almost sevenfold' and 'close to 10 times' are computed off a tiny 28 MW/193 MW base, and the 10x comparison only holds for the battery-only slice of the award, not the headline 2,400 MW. The purchases are also framed as backup without noting that 10 MWh against a 2.85 MW average draw buys roughly 3.5 hours - protection for equipment and an orderly stop rather than ride-through. Against that, the losses, spend and installations are concrete, so the gap is small rather than promotional.
Grant deadlines and vendor sourcing shape the picture
Incentives are unusually visible in the material itself. Both anchor purchases are part-funded by EU programmes (Next Generation at Fribin, Recovery and Resilience at Vista Alegre), the pipeline is a state grant award, and the vendor quoted says buyers are racing EU funding deadlines - so demand timing is subsidy-driven, not purely risk-driven. Two of the four expert voices, Sungrow and Self Energy, sell the products described, and buyers who have just committed capex have reason to justify it.
Consistent single-source reporting, no corroboration
Internally consistent and quantified, with named sources and figures that survive arithmetic checks, which supports moderate confidence in the direction of travel. But everything rests on one BBC article with no second publisher, no primary IDAE or Red Electrica document, an unstated year for the December award, and vendor assertions about demand that cannot be tested from the supplied material.
product
A four-hour battery retrofitted into an operating solar farm is the template load planners must price1 distinct publisher
product
A$1.2bn later, Supernode shows phased gigawatt-hour storage is bankable in tranches1 distinct publisher
product
368 new US power plants, and 302 of them are solar or batteries1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 25, 2026