Flexible connection agreements: boost or barrier for Poland's battery build?
Flexible connection agreements: boost or barrier for Poland's battery build?
For most new battery projects in Poland, the grid connection is now the first hurdle to clear. Grid operators refused 107 GW of connection requests in 2025, and the transmission grid has no spare capacity left for new generation until 2031.
Since 30 April 2026, an amended Energy Law has let operators offer a flexible connection agreement instead of a refusal. It trades the full access of a standard, firm connection for limits on how much power the battery can import or export. The limits last up to three years on one agreement, or indefinitely under a second type of agreement.
For a 2-hour battery starting in 2028, halving its grid access both ways for three years lowers its internal rate of return (IRR) by about 1 percentage point. However, it is up to the operators to decide whether to offer one, and by October 2026 none had published terms.
This article is for developers weighing a flexible offer, and for investors and lenders valuing those projects. A companion article on Polish grid reinforcement maps where and when grid capacity opens.
Key takeaways
- Flexible agreements are a boost on paper and a barrier in practice. They cost about 1 point of IRR for a connection that would otherwise be refused, but no operator has yet said where it would offer one.
- The cost is low because a battery signing now starts after ancillary revenues saturate. By 2028, arbitrage earns most of its income, and the limits last three years at most.
- A flexible agreement does not move a battery up the queue, but it turns a refusal into a connection. Turning one down to wait for a firm connection pays off if that connection arrives on time.
- A configurable agreement’s limits can last indefinitely, but in our forecast limits that follow real grid stress cost a 2028 battery about a quarter of a point of IRR.
A flexible agreement costs a battery revenue only while its limits last
A flexible agreement caps how much power a battery can import or export, for up to three years. For a developer facing a refusal, that means a three-year dent in revenue, not a permanent loss. For a 2-hour battery starting in 2028, halving grid access both ways cuts revenue by 14% across 2028 to 2030. Once the limits end in 2031, revenue returns to what a firm connection earns, which keeps the project's IRR cost small.
Poland's grid act added two agreements that limit import or export
Before 30 April 2026, a battery had one route to the grid. A standard connection agreement gives firm access up to the contracted capacity, in both directions and at all times.
The grid act of 13 March 2026 added two alternatives. The flexible agreement applies where a firm connection needs grid reinforcement, such as a new transformer or line. That work must already sit in the operator's development plan, agreed with the energy regulator, URE. It sets a limit in each direction and names the work the operator must do. The limits end when the operator finishes that work, or three years after the battery receives its final occupancy permit, whichever comes first.
The configurable agreement covers places where even planned reinforcement will not free enough capacity. Its limits can change with the time of day, grid conditions, the output of other plants or the reserves available to PSE. They can last indefinitely, and the operator has no duty to remove them.
| When available | Grid has capacity | Capacity needs planned reinforcement | No capacity even after planned reinforcement |
|---|---|---|---|
| How long limits last | No limits | Until reinforcement, at most 3 years after the battery's occupancy permit | Can be indefinite |
| Shape of limits | None | A limit in each direction | Can vary with time, grid conditions, other plants and system reserves |
The new agreements can also remove compensation for curtailment. Article 9c of the Energy Law withdraws it when an agreement drops the guarantee of reliable supply. Our reading is that a battery on either agreement would lose compensation for curtailment beyond its agreed limits.
A flexible agreement does not skip the queue, but turns a refusal into a connection
Operators publish available capacity net of what they have already promised to projects with valid connection conditions, so those projects keep their place. A flexible agreement instead uses capacity that is free for part of the time, or until the operator finishes reinforcement. A substation that fills with solar at midday, for example, can still take a battery's output in the evening.
The law says an operator may conclude either agreement, so developers cannot demand one. The usual advance payment, security and 24-month building permit rule still apply. A project signing in late 2026 therefore realistically starts in 2028. A refused battery keeps no place in line, so the cost of accepting a flexible agreement is the central question. Our companion article on Polish grid reinforcement maps where spare capacity opens up, and when.
Sign up to read this article for free
Unlimited access to our free articles
Monthly access to 2 Global Research articles
Benchmarks, Forecasts, Ko and more





