A site search that starts with a region has already gone wrong. Dutch grid operators publish availability per service area — the catchment of one named substation — and two locations ten kilometres apart routinely sit in different states. “Is there power in the Randstad” is not a question the published data can answer, and any answer you are given to it was produced by averaging away the only distinction that matters.
This is the first edition of a quarterly read of that data. It contains no capacity percentages, no site counts, and no forecast. That is not modesty about the market; it is what this quarter’s evidence supports, and a report that pads the gap with plausible numbers is worth less than one that shows you where the gap is.
What the public lookup establishes
A six-digit postcode resolves to a named service area through a public lookup that needs no account, no key, and no commercial agreement. We tested ten postcodes across four regional operators in August; all ten resolved to a named area. The lookup is not a trade secret, and treating it as one is the first mistake in this market: the plumbing is public, the interpretation is where the work is.
Availability is published in four states, and separately for each direction of flow — drawing power from the grid is graded independently of feeding into it:
- available, no queue
- limited availability
- under investigation, with a queue
- shortage, with a queue
Two consequences follow immediately. A site graded for feed-in tells you nothing about its position for consumption, which is the direction a datacenter cares about. And “under investigation with a queue” is not a soft version of “available” — it is a state in which your connection request joins a line whose length is not part of the published grade.
What the public lookup is silent about
The lookup returns the area. It does not return the area’s status. There is no companion endpoint that hands you the grade for the area you just resolved — we checked. The status lives in the operators’ published capacity material, which means anyone building a per-postcode picture is joining two sources by hand and inherits every mismatch between them.
Three further silences matter more than the ones people worry about:
How old the figure is. The publication does not stamp the moment you are reading it. A grade that refreshed at the start of a quarter is a grade from the start of a quarter, and it will be presented to you in month three with exactly the same confidence.
What a specific building is contracted for. This is the weakest link in every supply claim in this market, including claims made about us. Contracted capacity per site is very probably not publicly derivable. Anyone quoting you a hard megawatt figure for a named facility, sourced from public data alone, is quoting you an inference. This is why our own screening returns a band with the source named — under 1 MW, 1 to 5, 5 to 15, above 15 — and why the exact figure only ever comes out of a conversation with the operator.
Whether a granted connection is still live. A heavy connection that was granted to a previous industrial use is worth a great deal if it is still in force and nothing if it lapsed. The public data does not distinguish the two.
The evidence ledger
Every claim above, graded. We publish this table each edition, and a claim only moves up a grade when there is something new to point at.
| Claim | Grade | Basis |
|---|---|---|
| Postcode resolves to a named substation service area via a public, unauthenticated lookup | Verified | Tested directly, August 2026: 10 of 10 postcodes across four regional operators |
| Availability is published in four states, separately per direction of flow | Verified | The published legend |
| The lookup returns the area but not its status; no companion endpoint exists | Verified | Tested directly, August 2026 |
| Capacity data refreshes at minimum quarterly; queue information monthly | Second source | Corroborated, not independently confirmed by us |
| A postcode-level dataset with a 0–3 congestion grade exists | Second source | Corroborated, not in our hands |
| Contracted capacity per individual site is publicly derivable | Not established — we assess it is probably false | No source found; this is why we publish bands, not figures |
| Shut-down horticulture systematically retains heavy live connections | Not established | Plausible on the economics, no source checked |
What to ask this quarter
If you are shortlisting sites in the Netherlands between now and November, three questions are worth more than another round of brochures.
- Which service area is this site in, and what is its grade for consumption? Not the region. The named area, and the direction you actually need.
- When was that grade last refreshed, and what has changed since? A source without a retrieval date is not a source.
- Is the connection granted, energised, or planned — and can you show that in writing? Three different things, routinely described with the same word.
A facility that answers all three in writing has told you more than any capacity map will.
Next edition
November 2026. The intent is to add two things this edition could not support: the refresh cadence confirmed first-hand rather than through a second source, and a read on whether the postcode-level grading can be reconciled with the published area grades. If either fails to hold up, the November edition will say so in this same table rather than quietly dropping the row.