SM-012 — Energy Allocation and Fiscal Sustainability

The allocation question

Finland has approximately 3 GW of zero-carbon electricity capacity that is, or will be, contested between several large-scale use cases. The publicly documented cases include: SGFA-type distributed green hydrogen and synthetic fuel production (electrolysis, Power-to-X, CCU chains), foreign-owned hyperscale data centres with 24/7 baseload demand, domestic industrial loads (Outokumpu, Neste, SSAB) seeking long-term price certainty, and offshore wind development seeking grid connection.

SM-011 analysed the allocation mechanism from a grid capacity and flexibility perspective. SM-012 analyses the same allocation decision from a fiscal sustainability perspective: what does each outcome imply for Finland's ability to fund public services, service sovereign debt, and maintain economic resilience over a ten-year horizon?

The question is not which option produces more global economic activity. It is which option produces more Finnish fiscal capacity — the precondition for the state to fulfil its functions.

The institutional layer determines value retention

CN-013 established that the same physical electricity produces radically different economic outcomes depending on the institutional layer through which it passes. The physical constraint (3 GW, 18 TWh/year) is identical across scenarios. What differs is who captures the value, where it is booked, and how much remains within the Finnish fiscal perimeter.

In the SGFA scenario, the entire value chain — electrolysis, compression, distribution, synthetic fuel synthesis, fertiliser production — involves Finnish entities subject to Finnish corporate taxation, Finnish employees subject to Finnish payroll taxation, and Finnish consumers paying Finnish VAT. The institutional layer retains value domestically.

In the data centre scenario, the Finnish subsidiary typically books only a thin operating margin (empirically 1–3% in comparable transfer pricing structures), while the service revenue — generated by international cloud computing, AI inference, and storage — is booked in low-tax jurisdictions (Ireland, Luxembourg). The institutional layer exports value. The Finnish fiscal perimeter captures the electricity cost, the employment cost, and a thin margin, while the bulk of value flows to the parent entity.

Annual fiscal comparison (3 GW / 18 TWh)

Scenario A — SGFA / Hydrogen + CCU/PtL

Based on hydrogen production at 4.2 €/kg (378,000 t H2/year at 70% electrolysis efficiency) and synthetic fuel production at 2.1 €/litre (CCU/PtL chain), combined with the full domestic value chain:

Scenario B — Foreign-owned data centres

Based on hyperscale colocation economics (60 €/kW/month service revenue to Finnish entity), with transfer pricing reducing the Finnish subsidiary margin to approximately 2%:

Ten-year cumulative fiscal position

The 44 billion euro difference over ten years from a single 3 GW allocation decision is not an argument that data centres produce no value. They produce significant global value. The question is where that value lands — and the answer depends entirely on the institutional layer, not on the technology.

Implication for state fiscal capacity

The SGFA scenario produces approximately 3.5 billion euros in tax revenue over ten years that the DC scenario does not. In the context of Finland's 2026 fiscal position — approximately 2.5 billion euro annual deficit, interest costs rising toward 6.3 billion by 2030 (SM-001 §04) — this difference is material. It represents one year of additional deficit service capacity, or approximately 15% of the annual healthcare budget.

More significantly, the 22.7 billion euro trade balance improvement in the SGFA scenario directly reduces Finland's external financing requirement. A country that produces its own transport fuel needs fewer euros to buy it from abroad. This reduces the current account deficit and the structural pressure on sovereign debt accumulation.

The regulatory arbitrage mechanism

The data centre fiscal outcome is not produced by any individual firm acting improperly. Transfer pricing between a Finnish subsidiary and an Irish or Luxembourg parent is legal, common, and entirely rational from the parent's perspective. The problem is structural: Finland's regulatory framework does not require that the fiscal value generated by Finnish electricity, Finnish grid infrastructure, and Finnish grid stability services be retained within the Finnish fiscal perimeter.

This is the same mechanism that SM-011 identifies in the grid connection context and CN-013 identifies in the spatial value capture context. The physical layer (electricity, grid capacity, hydrological endurance) is Finnish. The institutional layer (corporate structure, tax jurisdiction, transfer pricing) is designed elsewhere. Value flows through the institutional layer outward.

Outokumpu CEO ter Horst stated in May 2026: "The question is what is the price of industrial electricity in 20 years — we need to see that investment environment." This is the long-horizon version of the same question: if the institutional layer that surrounds Finnish electricity is systematically designed to export fiscal value, the price of Finnish electricity in 20 years will reflect a tax base that has been progressively hollowed out to fund the public services that make Finland an attractive location for industry.

Sensitivity and limitations

The estimates in §03–04 are based on publicly available price data (hydrogen 4.2 €/kg, SAF 2.1 €/litre, colocation 60 €/kW/month) and standard Finnish tax rates. Key sensitivities:

• Transfer pricing margin: If the Finnish DC subsidiary books 5% margin rather than 2%, corporate tax rises by approximately 13 M€/year. The structural direction does not change.
• Hydrogen price trajectory: At 3 €/kg (pessimistic 2030 estimate), SGFA tax revenue falls by approximately 60 M€/year. The gap remains approximately 220 M€/year.
• Capacity factor: 6,000 hours/year (50% CF) is used throughout. At 4,000 hours (lower wind availability), all figures scale proportionally.
• CCU/PtL yield: Fischer-Tropsch yield at 0.45 kg fuel/kg H2 is conservative. At 0.55 kg/kg, fuel production volume rises 22%.

The 44 billion euro ten-year figure is robust to moderate variation in any single parameter. It is sensitive to the fundamental assumption that DC transfer pricing maintains the thin-margin subsidiary structure. If regulatory change forced full profit booking in Finland, the DC fiscal outcome would improve substantially — but this would require the same institutional design change that this memo argues is currently absent.

§ 06 — Portfolio-allokaatio: veropohjan diversifikaatio per TWh

Scenarios A and B in §03–04 compare two endpoints. The more useful question is how a 3 GW / 26.3 TWh electricity budget should be allocated across products to maximise fiscal stability — not peak return in a single scenario.

The core finding from portfolio analysis: SGFA does not maximise energy value. It maximises tax base diversification per TWh. Different products carry different market cycles — food (stable), fuels (cyclical), compute (globally growing), grid services (weather-dependent). Correlation between revenue streams is low. This is the structural advantage, not efficiency per MWh.

Optimaalinen allokaatio (3 GW SGFA, Suomi)

After subsidy costs (PtX support −150–250 M€, DC incentives −50–100 M€), net state cash flow is approximately 0.9–1.2 B€/year — compared to the 0–200 M€ range from single-product allocation.

The difference does not come from efficiency. It comes from breaking correlation between revenue streams.

Biogeeninen CO₂ Suomen rakenteellisena kilpailuetuna

The portfolio model assumes CO-route thermochemical conversion (TN-013) as part of the PtX block. A critical input parameter: Finland does not face a biogenic CO₂ supply constraint. Pulp mills, sawmills and CHP plants produce biogenic CO₂ as a continuous byproduct. This material has historically had no commercial value and has been vented or used in minimal internal applications.

This is a structural advantage that does not appear in standard CCU cost models, which typically assume either direct air capture (expensive) or industrial point sources with competing claims. In Finland's case:

• Biogenic CO₂ supply is geographically distributed across the same regions as SGFA node candidates (Pohjois-Savo, Kainuu, Pohjois-Pohjanmaa)
• The feedstock cost is effectively zero — the constraint is collection logistics, not availability
• CO-route economics improve further because biogenic carbon credit value does not need to be netted against capture cost

A second supply layer reinforces this advantage: Finnish municipal energy companies (kunnalliset energialaitokset) produce electricity and district heating predominantly from wood-based fuels. These plants are geographically distributed across exactly the same regions as viable SGFA node locations — Kuopio, Joensuu, Kajaani, Kokkola, Rovaniemi and others. Their biogenic CO₂ output is continuous, co-located with existing heat and power infrastructure, and currently has no utilisation pathway.

This means the Finnish biogenic CO₂ supply base is not concentrated in a handful of large industrial point sources but distributed across dozens of municipal CHP plants — a logistics advantage that further reduces SGFA node feedstock cost and removes single-source dependency risk.

This makes the Finnish PtX block in the portfolio structurally less subsidy-dependent than equivalent calculations for Central European or UK nodes, where biogenic CO₂ must be specifically sourced or replaced with DAC.

Public discourse: three positions, one missing variable

In May 2026, three senior Finnish voices addressed the data centre question in public. Their statements are reproduced here not as criticism but as empirical documentation of the analytical gap this memo addresses.