Greenpool Greenpool

Assessments

Skygard: the batch's clearest sovereignty case, and its least-documented one

PRE-REVIEW DRAFT — right of reply not yet sent

Bottom line: Skygard is the clearest Sovereign-axis case in the entire Norwegian batch — Norwegian state and municipal capital, a founding CEO drawn from the national security regulator, explicit “data on Norwegian soil” positioning — and, on paper, the largest heat-reuse claim we’ve found anywhere in this dataset. It is also the earliest-stage operator here: no capacity figure for any of its three sites is public, and its efficiency claims are unquantified. The strong sovereignty story should not be allowed to paper over either gap.

A note on which site this assessment centers

This assessment treats Hovinbyen as Skygard’s primary site, and its facility-page link points there. That’s a deliberate choice, not an arbitrary one: Hovinbyen is the one Skygard purpose-built, the one Telenor anchors as a tenant, and the one carrying the heat-reuse story below. Grorud and Lørenskog are acquired legacy facilities — bought from Orange Business Digital Norway (formerly Basefarm) in January 2026 — real parts of the Skygard portfolio, but not the site that defines what the company is building. Skygard’s own site names all three internally as OSL1 (Hovinbyen), OSL3 (Grorud), and OSL5 (Lørenskog) — codes it also inherited from Orange’s prior naming. Those codes are not unique across operators: Vaultica separately uses “OSL01” and “OSL02” for unrelated sites in Oslo, and STACK Infrastructure separately uses “OSL03” for an unrelated campus. Green Mountain shares the “OSL” prefix at two Oslo-area sites but never numbers them — its own names are “OSL-Enebakk” and “OSL-Hamar”. Read any bare “OSLx” code together with the company name.

What checks out

Testing the sovereignty claim, not repeating it

Skygard’s own about page states plainly that data is stored and processed in Norway, “utenfor rekkevidde av utenlandsk lovgivning” — outside the reach of foreign legislation. That’s a strong claim, and it’s worth testing rather than quoting approvingly. On the ownership side, it holds up better than most operators in this batch: Norwegian state and municipal capital, Norwegian-domiciled legal entities, a board dominated by representatives of Norwegian shareholders. On the operational side, we can’t yet verify the claim beyond ownership — we don’t know Skygard’s data-residency contract terms, its cross-border support arrangements, or whether any of its software vendors could themselves create foreign-legal exposure regardless of where the servers sit. “Outside the reach of foreign legislation” is a legal proposition, not just a geography one, and Skygard’s ownership structure is necessary but not sufficient evidence for it. We treat it as a strong starting hand, not a settled case.

What isn’t there yet

No capacity figure is public for any of the three sites. We looked and found none — not on skygard.no, not in the sourcing behind this assessment. The NOK 2.4 billion figure is the total joint-venture investment commitment across all three sites combined, not a per-site capacity or spend figure. This is marked UNVERIFIED and stays that way until Skygard publishes something more specific or answers directly.

The efficiency claim is a marketing number, not a measurement. Skygard states its facilities use 50–70% lower energy than older data centres. No baseline, no methodology, and no third-party figure accompanies this — it reads as a directional claim, not something we can grade as evidence.

The Analysys Mason Nordic ownership question is unresolved, not dismissed. See above — we’re not guessing at an explanation, we’re flagging it for Skygard to answer.

Why the sovereignty story doesn’t resolve the whole picture

It would be easy to let Skygard’s ownership structure carry the entire assessment — it is, after all, a genuinely rare thing in this batch: state and municipal Norwegian capital, not private foreign capital, and a founding CEO whose background was literally in national communications security. That’s real, and it matters for the Sovereign axis. But an operator with no public MW figure and no measured efficiency data is, on the Green axis specifically, one of the least documented in the dataset — the opposite of Green Mountain’s position, where the ownership is foreign but the operational evidence is unusually strong. The two axes are independent for a reason, and Skygard is the clearest illustration of why: strong on one, thin on the other.

Confidence

Ownership and its state/municipal character: A (primary, company-confirmed on multiple sides of the transaction). The Analysys Mason Nordic discrepancy: B — both the founding press release and Skygard’s current about page are primary sources that disagree with each other; unresolved, not a grading judgment on either. Current board composition: A (named, dated, company’s own current page). That Lindeberg led Skygard from founding to July 2026 and returned to Nkom: C (single named press source — Techwatch, 6 July 2026 — not yet independently corroborated or confirmed by Skygard). That Johannessen is the current CEO: A (primary, Skygard’s own site). The sovereignty claim (“outside the reach of foreign legislation”): tested above on the ownership evidence available, not graded as a single fact — it’s a legal proposition we can’t fully verify from ownership structure alone. The 12,000–15,000-homes heat-reuse figure: B (reported, not yet independently audited or cross-confirmed against a utility-side source). The Nkom launch-event datapoint: B (reported, single event, not independently re-confirmed). Capacity and efficiency claims: absent — no grade to assign where no figure exists. Skygard has not yet been contacted for right of reply; see the right-of-reply log for current status.