Greenpool Greenpool

Assessments

Blix Solutions: the cleanest ownership in the dataset, and the weakest efficiency figure

PRE-REVIEW DRAFT — right of reply sent 27 August 2026, pending until 10 September 2026

Bottom line: Blix is, on the evidence we have, the strongest Sovereign-axis case in the entire dataset and the weakest Green-axis case among the Norwegian operators we have covered so far. Both halves of that sentence are meant literally, and neither is softened to make the other more comfortable. The ownership chain ends at one named person in Oslo — nothing else we have assessed comes close on clarity. The self-published PUE of 1.3 sits well behind the 1.08–1.20 range the MW-scale operators publish, and total capacity of roughly 4 MW is a different order of magnitude from Green Mountain’s 93 MW at Enebakk alone.

What checks out

Blix operates three sites and rents space in many more

This distinction matters and is easy to misread. Blix operates BDC, NR5 and CJH. It also sells co-location in Stockholm, Copenhagen, Helsinki, Amsterdam, London and New York using space inside other operators’ facilities — that international list is tenancy, not Blix-operated capacity, and should not be read as a Blix footprint.

One of those rented locations is in our own dataset: Blix’s co-location table lists Vaultica OSL01 among its available sites (ISO 14001, ISO 27001, 6 kW per rack). So a Norwegian-owned operator with the cleanest ownership chain we have found runs customer equipment inside a facility owned by Apollo Global Management, a US asset manager — both companies assessed separately by us, at different points on the Sovereign axis. This is the clearest illustration in our dataset of a limitation the Index openly carries: it scores the building and its owner, not the machines inside and who controls those. See What the Index does not measure.

The efficiency figure, stated without cushioning

Blix publishes PUE 1.3 at both facilities. For comparison, the Norwegian operators assessed in batch one publish 1.08–1.20 — Green Mountain across four sites, Lefdal at roughly 1.08–1.1, Nscale at approximately 1.1. Blix’s figure is materially worse, and it is worth being direct about that rather than burying it in a sentence about how PUE isn’t everything.

One qualification belongs alongside it, and it does not rescue the number: 1.3 is self-published and not third-party audited — but so are most of the figures it is being compared against, so the comparison is at least like-for-like. What we cannot tell is whether 1.3 is a measured annual average or a design figure, which is a materially different claim. That is on our question list.

A third operator whose own capacity figures disagree with themselves

Blix publishes conflicting capacity figures for its own sites, and we show all variants rather than choose:

After Vaultica’s OSL01 and COP01, this is the third operator in our dataset whose own published capacity figures contradict each other. We state that as a fact about the state of published data in this sector and leave it there; all four figures above are on our right-of-reply list.

Why the trade-off is the actual finding

It would be easy to write Blix up as a sovereignty success story and leave the efficiency number as a footnote, or to write it down as a small operator that can’t compete and leave the ownership as a curiosity. Both would miss what the pairing shows.

Blix runs converted urban buildings in Oslo. Green Mountain runs purpose-built halls in a mountain and a converted ammunition store with fjord cooling. The first structure gives you a short, legible ownership chain and a business that answers to one identifiable person; it does not give you the thermal envelope or the siting freedom that produce a 1.08 PUE. The second gives you the reverse: world-class efficiency figures under an ownership chain that runs to Tel Aviv.

That is not a moral about which operator is better — it depends entirely on what a buyer is optimising for. It is a structural observation about how the two axes of this Index tend to trade against each other in practice, and Blix is the clearest single illustration of it we have found. It is also why batch one’s MW-scale scope excluded operators like this, and why the inclusion threshold is now published rather than left implicit.

What we don’t know

Confidence

Ownership chain and its terminating individual: A (primary registry — Brønnøysundregistrene via proff.no, checked 25 August 2026, corroborated by the operator’s Nkom registration). Facility capacities: C at both main sites — not for lack of publication but because Blix publishes figures that disagree with each other, as set out above. PUE of 1.3 at both sites: B (operator-published, self-reported, not third-party audited, and not stated to be measured rather than designed). BDC heat reuse: B (operator-published on its own facility page; local building reuse confirmed, the district-heating supply described as future and planned rather than operating). NR5 heat reuse: absent — none listed. WUE: absent. AI/HPC relevance: absent, and flagged above as a genuine open question rather than a negative finding. CJH: essentially ungraded — we have a name, a certification and a rack density, and not much else. Right of reply was sent to Blix on 27 August 2026; the response window is open until 10 September 2026. That deadline is later than the 4 September cohort simply because the email went later. See the right-of-reply log for status by operator.

Corrections

27 August 2026 — we reported no heat reuse where the operator publishes some. This assessment previously stated, under what we don’t know: “Heat reuse: none found for either site. Absent, and on our question list.” That was wrong. Blix’s own BDC facility page states waste heat is reused locally to heat the building, with future planned remote supply to district heating. The source was the operator’s own published page — one we should have found before publishing, and not one that required any privileged access. The contrast that survives the correction, and is worth keeping: BDC re-uses waste heat and NR5 does not, so it is the newer and larger facility that lists none. Corrected across this assessment and both facility pages.