DT-004 · Data Centre Grid Connection Terms · Finland

Finland is experiencing an unprecedented data centre investment wave — from 285 MW in 2024 to a projected 1.5 GW by 2030, with 3.4 GW in the planning pipeline. These facilities operate 24/7 with near-zero short-term price elasticity. WP-015 §2 identifies this as Shift I: demand rigidification. The structural question this log tracks is whether connection terms require data centres to contribute to system resilience, or whether they simply add inelastic load to an already tightening grid.

Central finding: Finland has no binding additionality requirement for data centre grid connections. Data centres can connect to the grid by signing PPAs with existing generation — without creating new capacity. The TEM flexibility mechanism proposal (800 MW by 2030) is voluntary and market-based. No irruptibility obligation, no energy reserve requirement, no priority inversion framework exists. The structural gap is widening as load commitments grow faster than the policy response.

1 · The Additionality Problem

The distinction that determines whether data centre growth strengthens or weakens the system is additionality: does a data centre's power purchase agreement finance new generation capacity, or does it merely commit existing capacity away from the spot market?

PPA type	System effect	Spot market effect	Adequacy effect
Additionality PPA — contract finances new generation (wind, solar, nuclear)	Increases total capacity	Neutral — new capacity enters market	Positive — system grows with load
Non-additionality PPA — contract commits existing generation away from spot	No new capacity created	Reduces available market volume → residualisation	Negative — Shift III materialises

According to CPC Finland CEO Erik Trast (April 2026), most data centre operators currently seek fixed-price contracts with existing production — not new. This is the mechanism WP-015 §4 describes as market residualisation: when pre-committed capacity grows, the spot market operates only on the residual volume and loses its ability to signal and allocate scarcity.

2 · Actor Map — Current Positions

PM Orpo (Government)

"Data centres are investments Finland should promote. Not once has another investment been blocked because of them."

Decision horizon mismatch — describes past, not 2027–2030 risk window. No additionality position.

Risto Siilasmaa (Investor)

"Data centres consume so much energy that more economically useful investments may not materialise because green electricity runs short."

Shift I validation — implicit additionality concern from allocation perspective.

Erik Trast (CPC Finland)

"The electricity system will hit a wall within a couple of years. Data centres should always be required to sign contracts guaranteeing new production."

Direct additionality requirement. Shift III diagnosis. Timeline: "2 years" = 2027–2028 = WP-015 risk window.

Risto Kuusi / Asta Sihvonen-Punkka (Fingrid)

"This scale of demand growth is not a problem for the transmission grid." / "Individual data centre projects do not directly affect electricity prices."

Operationally correct for transmission engineering. Does not address adequacy, additionality, or 72h+ endurance.

Anni Mikkonen (Suomen Uusiutuvat)

"If data centres sign long-term contracts with new production, that production can be built."

Implicit additionality requirement — describes the solution without making it binding policy.

Peter Lund (Aalto University)

"System flexibility will be crucial. There is a danger of moving forward without thinking things through."

Flexibility = WEM §12 target. "Without thinking through" = absence of irruptibility and priority inversion framework.

DIESL (Dec 2025)

"Evidence suggests most data centre operators do not significantly adapt electricity consumption based on price."

Empirical confirmation of inelastic load — Shift I operationalised. Low short-run price elasticity measured directly.

Jairi Palonen / Maine comparison (Apr 2026)

"Maine froze large data centre construction until 2027 because capacity was insufficient and costs threatened to fall on households."

Priority Inversion Risk made explicit: inelastic commercial load vs. vulnerable household load. 300,000 energy-poor households in Finland.

3 · Timeline

4 · The Priority Inversion Risk

A concept absent from current Finnish policy discussion: Priority Inversion occurs when economically optimised but inelastic commercial load displaces socially critical load during scarcity. In a stress event, data centres continue operating at full load while households, hospitals, and small industry face rationing — not because of explicit priority rules, but because data centres have pre-committed capacity through PPAs and cannot be interrupted.

Finland currently has 300,000 energy-poor households. Grid upgrade costs from data centre connections are socialised across all electricity users. The Maine precedent (construction freeze until 2027) reflects a political system that recognised this dynamic before the stress event — not after.

5 · What Finland Currently Requires — and Does Not Require

6 · WEM Snapshot — 17.4.2026

Date	Event	Source
2014	Finland reduces data centre electricity tax to near zero (facilities >5 MW). Effective demand subsidy with no additionality condition.	Finnish Government
2022	Tax threshold lowered to 0.5 MW — broader eligibility, still no additionality requirement.	Finnish Government
2023	Orpo government commissions Veli-Matti Mattila to review data centre policy. Mandate: economic benefits, not system adequacy.	TEM
2024	Mattila report published. Conclusion: Finland should attract data centres. No additionality requirement recommended. System adequacy not in scope.	TEM / Mattila 2024
2024	Data centre capacity: 285 MW operational. 3.4 GW in planning pipeline. Investment pipeline: €12 billion committed.	EK / Energiateollisuus
2025-Q1	Government proposes data centre flexibility mechanism — 800 MW target by 2030. Voluntary, market-based. Not a binding obligation.	TEM 2025
2025	Finland raises data centre electricity tax to standard rate (2.24 snt/kWh). Government freezes increase after Google signals investment uncertainty.	Finnish Government 2025
Dec 2025	DIESL publishes empirical study: data centres (except crypto) have near-zero short-run price elasticity. Confirms inelastic load assumption.	DIESL 2025
Apr 2026	Siilasmaa: green electricity shortage risk if data centre growth continues unchecked. Orpo: no investment has been blocked yet.	HS 16–17.4.2026
17.4.2026	Trast (CPC Finland): system will hit a wall within two years. Requires additionality PPA as condition for grid connection. First explicit public additionality demand from industry actor.	HS 17.4.2026
17.4.2026	10 GW wind+solar milestone passed. Nominally sufficient for typical demand — not for cold, calm winter peaks.	Suomen Uusiutuvat 2026

Load type	Flexibility	PPA coverage	Stress event behaviour
Data centre	Near-zero (DIESL 2025)	Typically high — pre-committed	Continues at full load
Industrial consumer	Moderate — some interruptibility	Partial	Partial reduction possible
Household / small business	High in theory — low in practice	None — spot-exposed	Faces price spikes, rationing
Critical infrastructure	None — must operate	None — state obligation	Protected by statute

Requirement	Status	Gap
Additionality PPA — connection requires contract financing new generation	Not required	Non-additionality PPAs accelerate market residualisation
Flexibility obligation — binding demand response or irruptibility	Voluntary (800 MW proposal)	Voluntary mechanism does not guarantee response during stress
Energy reserve contribution — MWh-dimensioned buffer, not just MW	Not in scope	72h+ adequacy gap unaddressed
Priority inversion framework — explicit load shedding order	Not defined	Stress event allocation left to market price signals — which fail during Shift III
Grid upgrade cost allocation — data centres fund their transmission impacts	Partial — connection fees apply, but system-wide reinforcement socialised	Cost transfer to all users continues

EPP (W168): 0.253 · Tight · FS(p): 42.6% (hydro_RF 0.568) · SP (W168): 27.4% · Tight
DC scenario +3000 MW: EPP → 0.509 · Elevated · SP → 100%

§10 scenario analysis: adding 3,000 MW data centre load (2030 scenario) to current system parameters pushes EPP from 0.253 to 0.509 — crossing the Elevated threshold. Stress persistence reaches 100%. This is the quantified endurance cost of Shift I without compensating capacity or flexibility obligations.

7 · Status

Active. This log will be updated upon: binding additionality requirements introduced, flexibility obligation legislation, grid connection policy changes, significant data centre connection approvals, or Fingrid/TEM statements on demand-side obligations.

Key trigger events to monitor: TEM's 800 MW flexibility mechanism legislation, Fingrid's updated grid connection criteria, EU Cloud and AI Development Act requirements, any Maine-style moratorium discussion in Finnish political debate.