Success for our clients: Federal Network Agency (BNetzA) discontinues work on the position paper on the allocation of withdrawal capacities from the power grid

On November 7, 2024, the 6th Decision-Making Chamber of the BNetzA (file number BK6-24-245) opened a consultation on a draft position paper on the allocation of take-back capacities at grid levels above the low-voltage level and invited comments. The paper provided for the publication of a recommendation for action following the consultation. On February 5, 2025, after reviewing the comments, the Decision-Making Chamber announced that it would not continue work on the position paper. We participated in the consultation and submitted comments on behalf of our clients.

Why the proposed approach was problematic:

• Rule‑of‑law concerns: Regulating capacity allocation via a non‑binding recommendation would have created de facto binding effects without a clear legal basis (Vorbehalt des Gesetzes) and risked encroaching on Article 12(1) GG (occupational freedom) for operators of large consumers (e.g., data centers, electrolysers). If the Chamber wished to regulate, it should do so by a binding, reviewable determination under § 17(4) EnWG, rather than through informal guidance. BNetzA’s own description of BK6’s toolkit includes Festlegungsprocedures, i.e., binding determinations.

• “Per‑head” repartition disadvantages large loads: The draft favored a per-head repartition model, which would systematically undervalue projects with high minimum capacities (e.g., hospitals, industrial sites, large data centers) and generate planning uncertainty that deters investment.

• One size doesn’t fit all: Participation requirements (e.g., uniform “project maturity” proofs) would unevenly burden different project types (from containerized storage to Tier‑3/4 data centers), resulting in indirect discrimination. Differentiated reservation criteria align better with project type and regulatory pathway.

• Network‑friendliness & locational transparency: Publishing available capacities by use case and location (e.g., data centers, industry, storage), and—where appropriate—allocating on a first‑come, first‑served basis within each use class, with robust, time‑limited reservations to prevent hoarding and ensure efficient, network‑friendly outcomes are more transparent.

• Transition rules: Grandfathering/transition provisions for existing reservations would be required to prevent ongoing projects to not strand by a mid‑stream change of approach.

  
  

Significance for the market

The Chamber’s decision is a clear win for due process and legal certainty. It confirms that informal instruments cannot replace measures with a proper legal basis where rights and competitive positions are at stake, and it preserves room for proportionate, reviewable regulation that can be tailored to different technologies and use cases.

What this means for stakeholders?

Until the agency pursues a future, properly grounded instrument (i.e., a binding, reviewable measure), allocation practices must continue to withstand close legal scrutiny and comply with non-discrimination principles.

In practice, stakeholders should keep working with the established processes of their local network operator. We observed that, in response to the Chamber’s draft, some market participants adopted a wait-and-see stance and, in a few instances, network operators paused or modified allocation procedures in anticipatory compliance. Those pauses and interim changes should now be revisited and, where appropriate, rolled back. If your project was affected, consider seeking reinstatement or reconsideration.

The market remains fluid. Real-time intelligence, sector expertise, and familiarity with local norms and decision-makers are essential to keep projects on track and to avoid inadvertent discrimination. Speak to us if you need support navigating local practice or unblocking a stalled allocation.

Sources & further reading:

• Announcement discontinuing the position paper (German) (5 Feb 2025).

• Consultation paper (German) (7 Nov 2024)