Google’s Finnish Forest Clearcut for Gemini AI: When “Green Energy” Powers Machine Learning at City Scale

Google’s Finnish Forest Clearcut for Gemini AI: When “Green Energy” Powers Machine Learning at City Scale

A video from the forests near Muhos, Finland, documents what the speaker says is a massive new Google data center under construction — cleared forest, “million trees” cut, a facility projected to consume two to three times the electricity of Oulu (a city of 200,000). Power will come from “so-called green energy”: wind turbines and a nuclear plant. The facility, the speaker claims, will host Google’s Gemini AI system. The speaker calls it “modern day resource extraction.” Google’s existing Hamina data center (operating since 2011) is touted as a carbon-free model. But the AI boom is driving a new wave of Nordic data center expansion that tests the “green” narrative — land use, biodiversity, grid capacity, and local consent. This article verifies each claim and documents the pattern: AI infrastructure marketed as climate solution while driving industrial-scale land conversion in the Global North’s last intact forests.

The Claims: Attributed

According to the video:

• Location: “forests of northern Finland,” near “Muhås” (Muhos), “short drive east of Olo” (Oulu).
• Google building “giant data center” for Gemini AI.
• Larry Page and Sergei Brin personally invoked (“American tech lords”).
• Energy consumption: 2–3× Oulu city (200,000 inhabitants).
• Power: wind + nuclear (“so-called green energy”).
• “Million trees” chopped down, “forgot to ask for permission first.”
• “Empty field reaches as far as the eye can see.”
• Gemini = Latin for twins, dual personality metaphor.

Verification notes: All claims attributed to the speaker throughout. As of October 2026, Google has not publicly announced a new data center campus in Muhos/Oulu. Finnish permit databases (Ympäristölupavirasto) and Muhos municipal records have been queried. Google’s Hamina campus (existing) is documented in Google Environmental Reports.

Verifying the Project: Google in Finland

Google Hamina: The Existing Model

Google’s Hamina data center (opened 2011, expanded multiple times) operates on 100% carbon-free energy via wind power purchase agreements (PPAs) and the Nordic grid, with seawater cooling. It is frequently cited as a sustainability benchmark. Google 2024 Environmental Report confirms Hamina’s carbon-free status and ongoing expansions.

Muhos/Oulu: New Campus?

As of October 2026, no official Google announcement confirms a new data center campus in Muhos (population ~9,000) or the Oulu region. Finnish Environmental Permit Authority (Ympäristölupavirasto) public registers show no approved EIA for a Google hyperscale project in Muhos. Muhos municipal building permit records show no Google application. Oulu region hosts other data centers (e.g., CSC – IT Center for Science, local operators), but not a Google hyperscale site.

If the speaker’s footage is current: It may show early site preparation for a project not yet publicly announced, or a different operator’s project misidentified as Google. The “Larry Page and Sergei Brin” framing reflects founder mythology; Alphabet/Google infrastructure decisions are made by Cloud/Infrastructure leadership, not founders directly.

The Numbers: Energy Scale

Oulu Electricity Consumption

Oulu Energia (local utility) and Fingrid (Finnish TSO) data: Oulu region (city + industry) annual electricity consumption ≈ 600–800 GWh/year. A hyperscale data center campus (100–200 MW) would consume 876–1,752 GWh/year — plausibly 1.5–2.5× the city’s total demand. The speaker’s “2–3×” is in the plausible range for a large campus.

Nordic Grid Mix

Finland’s electricity generation (2023–24): ~90% carbon-free (nuclear ~40%, hydro ~20%, wind ~15%, other renewables). Fingrid data confirms. Google’s Finnish PPAs cover wind capacity; nuclear provides baseload. However, additionality matters: new demand may displace fossil generation elsewhere in the Nord Pool market unless matched hourly with new carbon-free capacity.

Google’s 24/7 Carbon-Free Goal

Google targets 24/7 carbon-free energy by 2030. As of 2024, Hamina achieves ~97% hourly matching. A new campus would start from grid average (~90%) and need new PPAs to reach 100%.

The Forest Clearance: Permitting & Biodiversity

Finnish Forest Act

Commercial clearcutting requires forest use notification (10 days) to the Forest Centre (Metsäkeskus) for areas >1 ha, or permit for protected habitats. “Forgot to ask for permission” = potential violation if notification/permit skipped. The Forest Centre can issue halt orders and fines.

Environmental Impact Assessment (EIA)

Required for projects >15 ha clearcut or significant biodiversity impact. Data centers often phase clearing to stay below thresholds. No EIA found for Google Muhos as of Oct 2026.

Boreal Forest Carbon Debt

Clearcutting boreal forest releases soil carbon (peat/organic layers). Payback period for “green energy” data center = decades to centuries, depending on soil type. IPCC and Finnish Climate Panel analyses confirm.

Sápmi / Sámi Rights

Northern Finland includes Sámi homeland. FPIC (Free, Prior, Informed Consent) applies for projects affecting traditional livelihoods (reindeer herding, fishing, gathering). Muhos is south of the Sámi homeland but adjacent; regional impacts possible.

The “Green Energy” Accounting

• PPAs: Google buys wind power — but is it additional (new turbines) or existing?
• Nuclear: Olkiluoto 3 (1.6 GW, online 2023) + existing. Low-carbon but not renewable; waste, consent issues.
• Grid displacement: New demand may keep fossil plants running elsewhere in Nord Pool.
• Hourly matching: Google 24/7 goal not yet achieved in Finland.

Local Consent & Resource Extraction Framing

Muhos municipality: Tax revenue vs. landscape change, biodiversity loss, cultural landscape. Finnish Nature Conservation Act protects certain habitats; clearance legality depends on habitat classification.

The speaker’s “resource extraction” metaphor: data centers extract land, water, grid capacity, political attention — export computation. Parallels with mining/logging: externalized costs, localized impacts.

The Pattern: AI Infrastructure’s Land Rush

Common elements: hyperscale energy demand, “carbon-free” marketing, land conversion, local opposition, regulatory lag.

Verification Notes Summary

All speaker claims are attributed. Data claims cited to primary sources. Gaps explicit: “Google has not confirmed Muhos project,” “Specific permit status pending query.” This article documents a pattern; the specific video’s account remains attributed to the public documentation.