Published capacity is conditional capacity

Three Zuid-Holland cases where the capacity map is correct and the reading is wrong.

A capacity map can be entirely correct and still lead to the wrong site decision.

That should not be possible. In Zuid-Holland it happens in at least three separate ways, each one documented in a grid operator's own report, none of them visible on the map.

The map is not wrong. The reading usually is.

Near Naaldwijk, a site can pass an off-take screen and be unbuildable on the export side, on the grid of an operator most people do not know is in charge there. In the Rotterdam harbour, relief on the transmission layer is not access on the regional layer, and the two move independently. In the north of the province, the published headroom exists partly because gas plants stay available, and no replacement has been contracted.

Three cases, one conclusion. The number on the map is not an input. It is a conclusion, and it holds only under conditions the map does not print.

A note on scope, because scope is the point. Zuid-Holland is a three-operator province. Stedin runs most of it, including Rotterdam and the harbour. Liander covers parts of the north. Westland Infra runs the grid in the Westland and Midden-Delfland, the glastuinbouw region around Naaldwijk. Every claim below applies to the named operator's grid area only. That precision is not pedantry. It is the first of the three errors.

Decision lens

Before a site is buildable, five things have to be true at once: the right operator, the right layer, the right direction, a number that describes now rather than a forecast, and a dependency that holds.

Most screenings check one.

What follows is the other four, in three real cases, each one an application of the Project Capacity Test.

Signal #1 Direction: Around Naaldwijk the grid is open, until you ask the second question

One open direction is half a site.

Near Naaldwijk, the operator is not Stedin. It is Westland Infra, the independent operator for the Westland and Midden-Delfland. On its grid, the two directions point opposite ways.

For off-take, Westland Infra has not declared congestion. In a notice dated 4 April 2024, it states there is no congestion for off-take in its region. Read that sentence again, because it says less than it appears to. It is a no-formal-congestion statement, more than a year old. It means no restriction has been announced. It does not mean a large new connection is available on request.

For feed-in, the same operator tells a different story. Westland Infra pre-announced structural congestion for feed-in in late 2022, and the status still varies by station. Inkoopstation De Lier has since been released. Inkoopstation Westerlee remains under the pre-announcement. As of that same 4 April 2024 notice, the feed-in waitlist held 150 customers requesting 90 MW. That figure is more than a year old, and it is quoted with its date precisely because it is not fresh.

Why this matters commercially. For a bidirectional asset such as a battery, one open direction is half a site. A location near Naaldwijk can clear an off-take screen and still be unbuildable on the export side, on the same operator's grid, on the same day. An aggregate capacity map shows neither the split nor the operator. The safe conclusion here is narrow and worth stating plainly: both directions have to be verified separately, with the correct operator. Anything beyond that is inference, not information.

Source: Westland Infra, Congestie: landelijk probleem met regionale verschillen, 4 April 2024 (no off-take congestion; 150 customers and 90 MW on the feed-in waitlist), and Westland Infra, Congestie (De Lier released, Westerlee still under the feed-in pre-announcement). The "omgeving Naaldwijk" exception phrasing is Stedin's own, on its province page.

Different operator. Same mistake.

Signal #2 Layer: The Rotterdam harbour does not have one congestion problem. It has two.

Transmission relief is not project access.

Stedin has a word for what is happening in the harbour, and it uses it itself: gestapelde congestie. Stacked congestion.

On 1 July 2026, Stedin announced off-take congestion at three of its own stations, Hellevoetsluis, Europoort and Wellebrug, all at 25/10 kV, and described each as sitting inside a larger congested TenneT area. Underneath, TenneT's high-voltage grid in the harbour has been at its limit for years, and Stedin's regional grid in Europoort and parts of Botlek has been full for off-take since 31 October 2025.

Two layers. They do not move together, and they can move in opposite directions.

The structural fix is a new TenneT 380 kV station at Europoort, which the operators expect in the period 2032 to 2035. Until it arrives, the waitlist cannot be served in full. Stedin's update of 30 March 2026 puts the forecast off-take shortfall for the harbour region at 9 MW from 2027, rising to 64 MW in 2030, with the next official recalculation due in December 2026.

Those two figures are Stedin's. The underlying TenneT study is not publicly reachable, so they are reported as the operator states them, not as independently verified magnitudes, and they carry medium confidence below. The stacked structure itself carries high confidence.

Why this matters commercially. A headline that TenneT is expanding the harbour grid, or that a recalculation has shifted the forecast, does not mean a specific 25/10 kV connection point has opened. For anything in the Europoort or Botlek footprint, the residual shortfall is a congestion-management and flexibility question, not a connect-and-go question, and its timeline is set by the slower layer above. Relief upstairs is not access downstairs.

Source: Stedin, Update netcapaciteit (the 1 July 2026 station announcements and the term "gestapelde congestie"), Stedin, Transportschaarste afname, Rotterdams havengebied (regional grid full for off-take since 31 October 2025; the 380 kV Europoort station expected 2032 to 2035), and Stedin press release, 30 March 2026 (9 MW from 2027 to 64 MW in 2030; next recalculation December 2026). The underlying TenneT congestion study returns HTTP 403 and was not read directly.

Different layer. Same mistake.

Signal #3 Dependency: Some published capacity in Zuid-Holland exists because gas plants stay switched on

Headroom that depends on a plant staying available is borrowed headroom.

This is the least visible of the three, and the hardest to screen for, because nothing on the map hints at it.

In the delimited congestion area TenneT calls Noordelijk Zuid-Holland, the grid operators concluded jointly, on 4 December 2025, that keeping existing gas plants and greenhouse CHP available is necessary to prevent grid overload. Without that deployment, the forecast shows overloads arising from 2028. Extra high-voltage capacity does not arrive until the period 2032 to 2035.

The Gemeente Rotterdam summary of the same study is blunt about where that leaves things. As yet, TenneT and Stedin have contracted no flexible power. Without controllable power, a grid-security problem runs from 2028 through 2034.

The scope is exact, and worth keeping exact. This is the Noordelijk Zuid-Holland congestion area, for off-take, in the years 2028 to 2034. Rotterdam sits at its southern edge. "Necessary" means available for grid security. It does not mean running as baseload, and it does not mean contracted. This is not a province-wide statement, and it is not a claim that Zuid-Holland runs on gas.

Why this matters commercially. The capacity a developer reads off a map in this area rests partly on existing controllable generation that is not permanent and not yet replaced. So the thing to track is not only grid expansion. It is whether flexible power gets contracted, whether the plants currently assumed available stay available, and what the ministry and the regulator decide about them. A number that depends on a decision nobody has taken yet is a forecast wearing the clothes of a fact.

Source: TenneT with Stedin and Liander, press release of 4 December 2025, read via the Alphen aan den Rijn municipal record because tennet.eu returns HTTP 403 (gas plants and CHP necessary; overloads from 2028; high-voltage capacity 2032 to 2035), and Gemeente Rotterdam memo M2512-505, 4 December 2025 (no flexible power contracted yet; grid-security problem 2028 to 2034). Confirmed independently by Liander (too little peak-shaving, so no room yet on the waitlist).

Different dependency. Same mistake.

Three operators, three layers, three directions, one error repeated: the number was read as an input when it was a conclusion.

What we are watching

These are the four events that would invalidate the readings above.

The December 2026 harbour recalculation. Stedin says the next official recalculation of the harbour region is due in December, folding in the new priority framework. That is the next real test of the harbour numbers. No earlier figure should be read as a trend until it lands.
The three new Stedin areas. Hellevoetsluis, Europoort and Wellebrug were announced only on 1 July 2026, each opening a congestion-management study. Whether those studies find room, and where, is the regional-layer signal.
Contracting of controllable power in the north. In Noordelijk Zuid-Holland, no flexible power has been contracted. When it is, or when a gas plant's availability changes, the published headroom in that area changes with it.
Westerlee. Westland Infra's capacity study for inkoopstation Westerlee is still running. If Westerlee is released for feed-in, as De Lier already was, the direction picture near Naaldwijk shifts.

None of these are predictions. They are the specific things that would move the three readings above, and each is tracked against its primary source.

Sources and confidence

Every claim is checked against the Project Capacity Test: operator, layer, direction, time, dependency, confidence. Each signal above is a case where one of those breaks. Each was re-verified against a reachable primary source before publication.

Direction near Naaldwijk, off-take not in formal congestion while feed-in is restricted, on Westland Infra's grid: high confidence, Westland Infra primary source. The off-take statement is dated 4 April 2024 and is a no-formal-congestion statement, not proof of contractible capacity; the 150-customer, 90 MW feed-in waitlist carries the same date and is quoted with it.
Rotterdam harbour off-take congestion stacked across TenneT and Stedin layers, with 25/10 kV stations announced 1 July 2026, and the term "gestapelde congestie": high confidence, Stedin's own pages. The word is Stedin's.
Harbour off-take shortfall of 9 MW from 2027 rising to 64 MW in 2030, next recalculation December 2026: medium confidence. These figures are reported by Stedin in its press release of 30 March 2026. The underlying TenneT congestion study is not publicly reachable (HTTP 403) and was not read directly, so the figures are stated as the operator states them rather than as independently verified magnitudes. No comparison is drawn to any earlier figure, because the earlier October 2025 figure was defined and dated differently and is not comparable.
Structural fix is TenneT's 380 kV Europoort station, expected 2032 to 2035: high confidence, Stedin harbour page. The range is the operators' own indicative expectation, not a committed date.
Noordelijk Zuid-Holland dependency on existing gas plants and CHP staying available, overloads from 2028 without it, reinforcement 2032 to 2035, no flexible power contracted yet: high confidence, triple-sourced across the TenneT press release (read via the Alphen municipal record, tennet.eu being HTTP 403), Gemeente Rotterdam memo M2512-505, and Liander. Scoped to the Noordelijk Zuid-Holland area for off-take, 2028 to 2034, not province-wide.
Feed-in congestion exists in parts of the province beyond the Westland: high confidence that it exists. No provincial figure is published, so no numeric claim is made.

Source list

“In the north of Zuid-Holland, keeping existing gas plants and greenhouse CHP available is, in the grid operators' own words, necessary to prevent grid overload until high-voltage reinforcement arrives between 2032 and 2035, and no flexible power has yet been contracted.”

Source

Forward internally

Forward this to your team if anyone is asking:

Which operator actually runs the grid at our site, and have we verified both directions with them?
Does our harbour location connect on the layer that gets relieved, or the one that does not?
Is the headroom we are counting on in the north dependent on plants that may not be there?
Which Zuid-Holland locations survive all five screening questions, not just the one printed on the map?

The cost of misreading a capacity map is not paid at screening. It is paid later, when the connection request comes back and the assumption underneath it turns out to have been the wrong operator, the wrong layer, or a dependency nobody checked.

Capacity is a conclusion, not a starting point.

Wattlas tracks four operators, three layers and both directions across the Netherlands, against primary sources, and publishes when something material changes. To respond, or to request the underlying source table: [email protected]

Request founding access

Grid Intelligence