Published 2026-09-10 · Event 2020-10-01

West-coast hall and heat: GleSYS Falkenberg’s 3 MW and 5 MW rows

On 1 October 2020 GleSYS said heat delivery from its new Falkenberg hall to Falkenberg Energi had started, and listed 3 MW utility power at completion. A later operator page states 5 MW total capacity and continues to describe district-heat recovery. Those are two dated rows. Halland sits in SE4. This register does not invent a second Halland hall and does not convert 2020 potential heat wording into a measured megawatt-hour series.

West-coast data-center files are often thinner than Stockholm files and then over-read because of that thinness. One operating hall, one municipal utility, one dated heat start and two capacity sentences can look like a complete model of Swedish heat reuse. They are a complete model of one Falkenberg file. They are not a national template and not a residual-capacity map for Halland.

Read the 1 October 2020 release first. It is the dated event in this article. GleSYS said that on that day it started supplying district heating to Falkenberg Energi from a newly constructed data center on Sweden’s west coast. The release named Kanslistvägen 12, a total surface area of 4,000 square metres, a 2019–2020 timeline, and utility power of 3 MW. It described redundant power, renewable electricity labelled “Good Environmental Choice,” and a hall designed with heat recovery. Those are operator and counterpart sentences from a dated press page.

The same release used potential language that must stay potential. It said the venture had the potential to heat 1,700 residential buildings and to reduce carbon dioxide emissions by 1,570 tonnes every year. It said energy from the waste heat may amount to as much as 26 GWh annually, equivalent to 25 percent of Falkenberg’s total district-heating production. “May amount to” and “potential” are not a measured ledger. This register does not store those figures as delivered megawatt-hours. Readers who need a measured annual volume must find a later primary source. The pages used here do not publish one.

Read the later GleSYS Falkenberg operator page second. The page used for this register states a total capacity of 5 MW, describes 100 percent renewable energy and district-heat recovery, and discusses high-density racks as a design. Whether 5 MW superseded 3 MW through expansion is not explained on the cited pages. The two numbers use different wording — utility power in 2020, total capacity later — and may not even be the same metric. Keep them as two dated rows. Do not silently replace 3 with 5. Do not add them.

Falkenberg Energi is named in 2020 as the heat buyer and as a municipally owned company. Bo-Anders Antonsson, then chief executive in that release, called the incorporation of this type of third-party supplier in the district-heating network a leap for heat recovery in Falkenberg. That names a counterpart and a municipal system. It does not publish a multi-year invoice series. A delivery start is stronger than a prepared pipe in Stockholm operator copy. It is still not a 2026 measured volume.

Place the hall in the right geographies and only those. Falkenberg municipality, Halland County, bidding area SE4. Staffanstorp is also in SE4 and is a Microsoft location from November 2021; it is not this hall. GleSYS Stockholm in Västberga is a different municipality and has no megawatt on the cited Stockholm page. Stockholm Data Parks and Stockholm Exergi are a different county’s invitation pattern. Do not move Falkenberg’s dated start into Kista, and do not move Exergi wording into Halland.

Mechanisms around the hall remain ordinary Swedish mechanisms. A detaljplan and a building permit sit with Falkenberg municipality; Boverket explains the instrument. A connection answer sits with the network owner, not with SE4. Svenska kraftnät’s indicative map, if opened, does not show the underlying network and does not replace the formal application. Heat and power are different contracts. Renewable-electricity wording is a supply-attribute claim, not a reserved import.

Design sentences on the later operator page need the same hygiene as megawatts. Rack-density and GPU-density claims are design limits unless occupied load is stated. A PUE figure is only usable with a period and a boundary; operator pages in this set have stated a design PUE, which is not a planning score and not a heat volume. Carrier lists are not independently verified as-built diversity unless a current drawing says so. ISO certificates named by the company are company claims about management systems, not hall megawatts.

Common misreads follow a pattern. Replacing the 2020 3 MW utility row with the later 5 MW total without a source that explains the change. Treating 26 GWh potential as delivered heat. Adding Falkenberg into a Stockholm Exergi total. Treating SE4 as a connection refusal. Inventing a second Halland campus. Treating “West Coast’s largest” marketing language from 2020 as a ranking maintained by this register. Treating a 2020 completion sentence as a 2026 expansion certificate.

What to ask next is local and dated. What measured heat volume is public after 1 October 2020, and for which calendar year? How does the current 5 MW operator figure relate to the 2020 3 MW utility-power figure — expansion, remarketing, or a different metric? Which municipal documents cover Kanslistvägen 12? Which network owner has issued the current written position, and for how many megawatts of which definition? Which carriers are independently diverse at the site? Which GPU-density claims are design limits rather than occupied load?

The west-coast file is also a test of national heat language. EcoDataCenter describes recovery with Falu Energi & Vatten. atNorth and Conapto name Stockholm Exergi. GleSYS named Falkenberg Energi and published a start date. Those are three municipal utilities and three evidence states. None of them is a national heat market. A reader who needs a Swedish heat-reuse model should keep the three files on one page and refuse to add their adjectives.

This article proves that Falkenberg has an operating GleSYS hall, a dated heat-delivery start, a named municipal buyer, a 3 MW utility row from 2020 and a 5 MW total-capacity row from a later operator page. It does not prove the arithmetic relationship between 3 and 5. It does not prove a measured annual heat volume after 2020. It does not prove residual Halland grid capacity. Unknown measured heat stays unknown.

Why it matters

Falkenberg is one of the few Swedish halls in this set with a dated heat-delivery start. That makes it a reference file — and a temptation to over-generalise. The 3 MW and 5 MW rows are the capacity lesson. The unpublished measured heat after 2020 is the heat lesson.

What remains unknown

Whether 5 MW superseded 3 MW through expansion is unexplained on the cited pages. Measured annual heat delivery after 2020 is not in those pages.

Sources