Viska Sjóðir · Macro book · Strategy memorandum

The Four Themes of the Macro Book

The fund presents itself as a vehicle for access to four core thesis themes — the fixed majors. Every holding and every sub-theme belongs to exactly one of them, and daily intelligence is produced per theme. This memorandum states each thesis, how the four differ, how they interlock, and where every dollar of the invested book sits today.

book as-of 2026-07-24 · broker-verified classifications as-of 2026-07-27 weights: % of invested capital
Part I

The four theses


The boxed statements are the approved client-facing thesis texts (operator, 2026-07-15), shown verbatim. The commentary beneath each is the strategy desk's reading of what makes the thesis distinct.

Hrávörur & málmar

Minerals

71.1% of invested book
Approved thesis text Hard-asset scarcity: monetary metals (gold, silver, PGM) hedge fiat and geopolitical risk while industrial metals (copper, aluminum, uranium) face structural supply deficits against electrification and nuclear-renaissance demand.

This is one theme with two engines. The monetary sleeve (gold, silver, PGM, and the royalty houses) is the fund's most direct expression of the fiscal-dominance view: central banks accumulate gold, real assets reprice under financial repression, and miners lever that repricing. The industrial sleeve (copper, aluminum, uranium) is a supply-deficit position: electrification and the AI buildout demand metal that a decade of underinvestment cannot deliver on time. Both engines argue for owning things that come out of the ground — which is why this is the heavyweight sleeve of the book.

Orkuinnviðir

Energy Infrastructure

19.6% of invested book
Approved thesis text The grid is the bottleneck of electrification: everything outside the datacenter fence, transmission, distribution, and energy transport and storage, must be rebuilt for AI-era load growth. The fund owns the buildout.

Distinct because it owns the delivery problem, not the energy itself and not the computer. Turbines, switchgear, transformers, transmission construction — the equipment and labor between generation and load. Its cycle also runs latest: order books (GE Vernova's ~100 GW backlog class) convert to revenue over years, so this sleeve peaks after the first AI capex wave. Its named risk is the same event that flatters it — a pause in AI capex deflates the crowded end of the trade.

Orkuframleiðsla

Power Generation

6.1% of invested book
Approved thesis text Years of underinvestment in energy production meet structurally rising demand. The fund owns generation capacity and its suppliers across gas/LNG, oil majors, and short- and long-term energy production dynamics.

Distinct because its scarcity is molecules and megawatts, not equipment: upstream capacity that was starved of capital for a decade while demand — now amplified by datacenter load — kept compounding. It holds both the producers and their direct service suppliers (drilling, energy services), on the view that underinvestment pays the service layer first when activity returns.

AI & gagnaver

AI

3.1% of invested book
Approved thesis text The AI buildout is a multi-year capital-expenditure supercycle. The fund owns the physical value chain inside the datacenter that the buildout cannot proceed without: compute and memory silicon, in-rack power equipment and cooling, and the robotization wave that AI enables. Grid and generation outside the fence belong to the energy worlds.

The doctrine expresses AI through picks and shovels, not hyperscalers: the book deliberately owns none of the Mag 7 except Alphabet — the single sanctioned direct large-cap AI exposure. The small headline weight understates the fund's real AI exposure, which is spread through copper, uranium, and the power chain by design; this theme holds only what lives inside the fence.

Part II

One machine, four rooms


The four themes are not four unrelated bets — they are stations along a single causal chain, plus a monetary hedge on the regime that finances it. The AI capex supercycle creates power demand; power demand exposes a generation deficit and a grid bottleneck; and every stage of that buildout is metal-intensive. The doctrine's boundary rule keeps the rooms clean: AI is inside the datacenter fence, Energy Infrastructure is everything outside it; uranium belongs to Minerals even where the argument is nuclear power.

Minerals copper · aluminum · uranium supply the entire buildout
supplies
Energy Infrastructure grid, turbines, T&D — outside the fence
delivers power
the datacenter fence
AI compute, in-rack power & cooling
Power Generation gas/LNG, upstream capacity & services
fuels
generation feeds the grid; the grid feeds the fence — demand flows back the other way
The monetary-metals sleeve of Minerals runs on the second engine: it hedges the fiscal regime that funds this entire buildout (central-bank gold accumulation, real assets over financial). One sleeve, both engines — which is why Minerals is the anchor weight rather than a satellite.

Read as a matrix: what is scarce differs by room — metal in the ground (Minerals), megawatts (Power Generation), grid capacity (Energy Infrastructure), and datacenter capability (AI). When each pays differs too: production and metals reprice earliest on scarcity; infrastructure converts backlog latest; AI capex is the demand signal that drives all three. The book is therefore sequenced, not just diversified — the same story at four different clock speeds.

Part III

The invested book by theme


27 positions, IBKR statement of 2026-07-24, weights on invested capital (cash excluded; cash is separately 30.7% of total equity). Sub-themes are the strategy desk's classification, recorded with confidence and sources in the fund's classification store.

Minerals · Hrávörur & málmar

71.13%
CENXCentury AluminumAluminum7.22%
BBarrick MiningGold — major7.07%
SCCOSouthern CopperCopper7.00%
NEMNewmontGold — major6.83%
FCXFreeport-McMoRanCopper6.11%
AAAlcoaAluminum6.05%
RIORio TintoDiversified mining4.47%
SBSWSibanye-StillwaterPGM & gold4.45%
CCJCamecoUranium4.29%
FNVFranco-NevadaRoyalties & streaming4.19%
LUNLundin MiningCopper & base metals3.81%
HLHecla MiningSilver3.70%
WPMWheaton Precious MetalsRoyalties & streaming3.27%
AUGOAura MineralsGold — junior2.68%

Energy Infrastructure · Orkuinnviðir

19.65%
POWLPowell IndustriesSwitchgear & distribution6.80%
GEVGE VernovaTurbines & grid equipment2.97%
MYRGMYR GroupT&D construction2.78%
ETNEatonElectrical equipment2.45%
NVTnVent ElectricConnection & protection2.00%
AZZAZZGalvanizing & coatings1.46%
VRTVertivDatacenter power & cooling0.85%
VMIValmont IndustriesGrid structures0.24%
PRIMPrimoris ServicesEnergy construction0.09%

Power Generation · Orkuframleiðsla

6.11%
EQTEQTNatural gas4.40%
RIGTransoceanOffshore drilling services1.45%
AMRCAmerescoEnergy services & efficiency0.25%

AI · AI & gagnaver

3.12%
GOOGLAlphabetHyperscaler — sole Mag-7 exception3.12%