Sweden versus Denmark siting — a Danish grid-full headline is not a Swedish spare megawatt

Reading order

Seat the country and the paper class before you compare slogans. A Swedish application cell on the 2 September 2026 area table is not a Danish requested-gigawatt sentence in a May news report. Energinet’s March “grid is full” title is not a Swedish reservation table. DEN01’s Ballerup heat paper is not SWE04. Write the issuer, the date and the object on one line before you argue that one country is “open” and the other is “closed”.

Then keep the Danish process window on its own row. The 24 March 2026 Energinet notice says the grid is full and prioritisation is needed. The 27 May 2026 notice introduces a new model for larger connections and says new agreements will first be signed in the autumn. CNBC records the pause as temporary: three months, or until Energinet can complete an overview. A temporary pause is not a cancelled Ølgod campus and not a spare Swedish megawatt.

Then keep the 60 GW and 14 GW sentences labelled as requested volumes in a news report of an Energinet briefing. Around 60 GW waiting against about 7 GW Danish peak load is a comparison printed by CNBC on 4 May 2026. Data centres as 14 GW of those possible new connection projects is the same news attribution. Those figures are not a hall list, not energised IT load and not a Swedish area-table cell.

Then keep named Danish land and heat papers on their own geography. Varde’s Krageris page is a 20 April 2026 approval of plan proposals, an eight-week hearing that ended 25 June 2026, and an expected later political treatment. DEN01 Ballerup is a 22.5 MW metro announcement with heat from 2028 for more than 8 000 homes. DEN02’s Selected Group greenhouse paper is a different Ølgod instrument. None of those papers is a Swedish reserved offtake.

• Country and paper class before a Nordic siting slogan.
• Energinet’s autumn signing window is a process fact, not a hall.
• 60 GW and 14 GW stay requested, news-attributed volumes.
• Ballerup heat is Danish offtake, not SWE04.

What a Swedish queue clock and a Danish pause clock actually measure

The 2 September 2026 Swedish area table remains useful here because it refuses a reserved-power reading. The printed data-centre consumption cells stay applications until reservation and förhandsbesked. A blank SE4 cell is not proof that Sweden should “take” Danish demand. Adding 495 MW, 1 320 MW and 5 169 MW does not create a Nordic spare-power score and does not become 14 GW of Danish requests.

Energinet’s March and May 2026 notices measure a Danish large-connection pause and a later signing window. The March title says the grid is full and prioritisation is needed. The May title says a new model arrives and new agreements wait until autumn. CNBC’s three-month temporary reading stays attached to that news report. Political agreements and an adjusted regulatory framework are described there as still needed. Those sentences do not cancel Varde’s Krageris proposals and do not open a Swedish plot.

The 60 GW and 14 GW figures measure requested connection volume as reported by CNBC from an Energinet briefing. Around and almost a quarter are the news wording. Fourteen gigawatts is not a list of named halls. This page does not convert 14 GW into a campus census and does not move any of that total onto SE3 or SE4.

Named Danish project papers stay project papers. Varde’s 20 April 2026 council step approved proposals for a local plan, a municipal-plan addendum and a wastewater-plan addendum at Krageris. The municipal page gives about 174 hectares, room for up to 225 000 square metres of data center and up to 475 000 square metres of greenhouses. Up to and expected are the municipality’s words. Square metres are not megawatts. This page does not import a 250 MW sentence from another DEN02 announcement.

Heat papers stay heat papers. DEN01’s 26 November 2025 Vestforbrænding agreement is Ballerup municipal heat, 22.5 MW announced metro site power, first-quarter 2026 as an operating aim, and more than 8 000 homes from 2028 as an offtake illustration. DEN02’s 11 August 2026 Selected Group greenhouse paper is a different Ølgod instrument with a third-quarter 2028 first-phase target. Konti-Skan Connect’s Swedish pause stays an interconnect file. None of those objects is a reserved Swedish hall megawatt.

What named primary sources show

Svenska kraftnät’s 2 September 2026 area table remains the dated Swedish queue-cell source. Energinet’s 24 March and 27 May 2026 notices remain the Danish process titles. CNBC’s 4 May 2026 report remains the dated attribution for 60 GW, about 7 GW peak and 14 GW data-centre requests.

atNorth’s 26 November 2025 DEN01 notice and Vestforbrænding’s Ballerup page remain the Danish municipal-heat sources. Varde’s Krageris page remains the dated proposal-step source. The Selected Group papers remain a separate Ølgod heat-intention file.

Konti-Skan Connect’s pause stays a Swedish interconnect paper. No cited source here publishes a reserved megawatt that converts a Danish requested gigawatt into a Swedish plot.

A matrix instead of an open-versus-closed slogan

Draw one page with country rows and evidence columns. Swedish application megawatts, Danish requested gigawatts, pause titles, autumn signing windows, municipal plan proposals and named heat papers are different columns. Empty cells stay empty. A 14 GW news-attributed request cell is a Danish footnote, not a Swedish spare-power column.

Do not invent a Sweden-versus-Denmark league table from the 2 September 2026 cells or from the May news totals. Those remain different objects from different issuers and methods.

Common misreads

The first misread is to treat 14 GW of Danish data-centre requests as energised halls, or as a volume that Sweden should absorb. The figure is a news-attributed requested connection volume. The second is to treat Energinet’s autumn signing window as proof that Denmark is closed forever, or that Sweden is therefore open. The third is to treat a blank SE4 cell as southern spare power waiting for Danish demand.

The fourth is to paste DEN01’s 22.5 MW and 8 000-home sentences onto SWE04 or Ølgod. The fifth is to treat Varde’s April proposal approval as a building permit. The sixth is to treat Konti-Skan Connect’s pause as a Danish hall reservation. The seventh is to collapse this page into the Sweden-versus-Norway or Sweden-versus-Finland page.

What to ask next

Ask which country the municipality sits in, which paper class — Swedish application, Danish requested gigawatt, pause title, plan proposal or named heat agreement — anyone is quoting, and which dated letterhead you actually hold. Ask whether a heat sentence is Ballerup, Ølgod or a Swedish offtake. Ask which network-owner letter you already hold.

Then open the Sweden-versus-Norway page, the Sweden-versus-Finland page, the electricity-areas page and the heat-reuse pages. If the operator will not name the country and the paper class, record the gap. Do not fill it from an open-versus-closed slide.

1. Is the live paper a Swedish queue cell or a Danish requested gigawatt?
2. Has 14 GW been treated as energised Danish halls?
3. Has the autumn signing window been treated as a permanent close?
4. Has DEN01 heat been pasted onto SWE04 or Ølgod?
5. Has a Krageris proposal been treated as a building permit?
6. Has a Sweden-versus-Denmark league been invented?

What this does and does not prove

This page proves that a dated Svenska kraftnät area table already prints Swedish data-centre application cells, that Energinet already published a 2026 large-connection pause and an autumn signing window, and that named Danish heat and plan papers already sit on Ballerup and Krageris — and that those objects are not a Nordic winner, not reserved Swedish megawatts and not energised Danish IT load.

It does not prove that a named Swedish or Danish hall can connect next year. It does not invent a hall list inside 14 GW. It does not treat Konti-Skan Connect as a campus offtake.

Empty cells stay empty. Until a later primary source names a reserved megawatt that actually converts a Danish requested gigawatt into a Swedish plot, the siting clock stays a method, not a census.