Norges Bank 13F Portfolio

Because all top positions are new in this 13F sample, the interesting question is where within tech Norges concentrated. The answer: a barbell of AI compute, hyperscale platforms, and the tools that make the chips.

On the compute side, Nvidia at 6.50% and Broadcom at 2.48% anchor a clear bet that data‑center silicon remains the choke point in the AI boom. Micron at 1.75%, AMD at 1.35%, Intel at 0.86%, Marvell at 0.42%, Texas Instruments at 0.39%, and a legacy SanDisk line at 0.47% round out a memory-and-connectivity thesis: AI workloads need not just GPUs, but high‑bandwidth DRAM, NAND, and custom accelerators.

The other half of the stack is software and hyperscale distribution. Microsoft at 3.82%, Alphabet’s two share classes at 2.88% and 2.63%, Meta at 1.84%, Amazon at 3.34%, and Oracle at 0.37% form a dense cluster of cloud, search, and enterprise workloads. Palantir at 0.36% is a more idiosyncratic nod to applied data analytics — small in size, but symbolically consistent.

Even within tech, Norway is not just buying the headline tickers; it is paying for the plumbing. Applied Materials at 0.81%, Lam Research at 0.78%, KLA at 0.59%, Cisco at 0.64%, and Digital Realty at 0.35% show a studied preference for the picks-and-shovels of fabs and data centers, not just the layer‑one brands.

The fact pattern is stark: in this top‑50 slice there are no reported trims at all, only new lines. That doesn’t mean Norges Bank didn’t sell anything elsewhere, but within its largest US holdings it chose not to crystallize gains.

Many of these positions show triple‑digit gains versus reported average cost — Nvidia with +656.3%, Micron with +659.1%, Broadcom with +296.1%, and Lam Research with +373.4%. Yet there is no sign of scaling them back at quarter‑end, which reads as an explicit rejection of the “take profits into strength” playbook.

The same applies outside semis. Apple’s stake sits roughly +194.7% above cost, Alphabet’s main line at +227.8%, and Eli Lilly at +249.7%, but all appear as intact, untrimmed exposures in this filing. For a conservative institution, that’s a notable signal: the gains are being treated as proof of thesis, not as a reason to cut risk.

With tech at 61.41% of the book, sector allocation is no longer about fine‑tuning; it’s about supporting a dominant AI thesis without letting macro risk run wild. Everything non‑tech is effectively portfolio ballast.

Consumer names get 10.63%, but skew defensive and scale‑driven: Amazon, Walmart, Costco, Home Depot, Procter & Gamble, Netflix. Healthcare at 7.96% centers on large‑cap pharma and managed care — Eli Lilly, Johnson & Johnson, AbbVie, AstraZeneca, UnitedHealth, Merck — giving exposure to another secular growth engine with very different cycle drivers than semis.

Financials at 6.03% (JPMorgan, Bank of America, UBS, Wells Fargo, Royal Bank of Canada, Morgan Stanley) and energy at 2.17% (Exxon, Chevron) read like macro hedges on higher nominal growth and rate regimes. Berkshire Hathaway’s two share classes at a combined ~1.21%, plus GE Aerospace, RTX, and GE Vernova, add an industrial and conglomerate overlay: if the AI cycle spills into capex and real‑asset demand, these names should already be in position.

Real estate at 3.42% is tellingly narrow: it’s mostly about payments quasi‑toll‑roads (Visa, Mastercard) and data‑center/healthcare REITs (Digital Realty, Welltower) — again, things that scale with transaction volume, aging demographics, and data intensity, not just GDP.

Taken together, the book says Norges Bank is not trying to out‑smart the AI cycle; it is trying to own as much of its infrastructure and cash flow as a sober, rules‑bound allocator can justify. The scale in semis, cloud platforms, and enabling equipment leaves little doubt about where they think equity value will accrue.

The deliberate barbell to banks, big pharma, energy, staples, and Berkshire‑style conglomerates hints at how they want to survive being wrong. If AI multiples compress or capex normalizes, the non‑tech sleeve should keep the fund tethered to broad nominal growth, healthcare demand, and commodity pricing.

For outside observers, the key takeaway is that this portfolio treats mega‑cap tech dominance as structural, not a late‑cycle anomaly. Unless regulation, geopolitics, or a genuine technological miss undermines the AI investment case, the most likely adjustments from here are within the tech stack — more emphasis on under‑owned enablers, less on crowded winners — rather than a wholesale retreat from the theme.