Swiss National Bank Doubles Down on AI Platforms, Adds Defensive Pharma

The biggest dollar moves are a clean read on conviction: Swiss National Bank is not rotating away from winners, it’s pyramiding into them. Nvidia, Apple, and Microsoft all saw share counts lifted around the mid‑single digits, adding roughly $824.2M, $626.3M, and $507.5M of exposure respectively, despite huge embedded gains versus average cost.

This is classic “own the rails” behavior in AI. Nvidia, Broadcom, and Micron sit at the semiconductor heart of the build‑out; Microsoft, Alphabet, and Meta monetize AI via cloud, productivity, and ad platforms; Amazon rides both cloud and commerce. Every one of those franchises had its share count bumped 6–8%, signaling that the bank views recent price action as noise inside a durable adoption curve.

The outlier, and the clearest new thesis, is AstraZeneca. A fresh $1.20B position drops straight into the top‑20, alongside existing scale in Eli Lilly, AbbVie, Merck, and UnitedHealth. Together, that cluster says the fund wants exposure to long‑duration drug and therapy pipelines — obesity, oncology, immunology — and is willing to pay today for cash flows that will matter long after the AI hardware cycle normalizes.

The trims widget is empty, but that absence is itself a signal: there were no visible top‑50 reductions this quarter. Funding for the new AstraZeneca stake and the broad set of adds clearly came from outside the disclosed list — either smaller positions sold down, or cash flows into the equity book.

Within the top‑50, the pattern is uniform: every listed line shows an increased share count, often around 6–7%. That homogeneity is not passive drift; it’s a deliberate top‑up of the existing playbook rather than a re‑write. The manager did not rotate from AI into health care; it layered health care on top of AI.

If conviction were cooling in the growth engines, you would expect profit‑taking in high‑gain names like Nvidia, Broadcom, or Lam Research. Instead, each saw more capital, even where gains versus average buy run into triple‑digit territory. The sell discipline, at least this quarter, is happening off‑stage in smaller or fully exited names we can’t see in the top‑50 snapshot.

Headline sector weights barely budged, but the internal tilt matters. Technology slipped from 59.33% to 58.58% — effectively unchanged — even as AI‑linked names soaked up most of the new dollars. That tells you the fund is concentrating within tech: more into semis and cloud platforms, less into non‑core tech further down the book.

The real rotation is into health care. Sector weight jumped from 8.88% to 10.02%, driven by the new AstraZeneca position and across‑the‑board adds to Eli Lilly, AbbVie, Merck, Johnson & Johnson, UnitedHealth, Amgen, and Abbott. Health care is now the clear second pillar behind tech, not an afterthought.

Consumer exposure stayed remarkably stable. Discretionary inched down to 13.59% while staples barely moved at 1.59%, but within those sleeves the manager reinforced dominant franchises: Walmart, Costco, Home Depot, McDonald’s, TJX, Coca‑Cola, and Pepsi all saw incremental buying. Energy, telecom, and basic materials weights each ticked down a hair, underscoring their role as ballast — Exxon Mobil, Verizon, AT&T, and Linde are there to stabilize, not drive, the P&L.

Taken together, the moves say Swiss National Bank is willing to wear near‑term volatility to stay overexposed to the biggest general‑purpose technologies of this cycle. The combination of heavy Nvidia, Microsoft, Alphabet, Broadcom, Amazon, Meta, and a deep bench of semicap (Applied Materials, Lam Research, KLA) is a bet that AI compute demand outlasts the inevitable correction scares.

At the same time, bulking up health care suggests an awareness that the AI trade is crowded. By scaling Lilly, AstraZeneca, AbbVie, Merck, and UnitedHealth, the fund is building a second growth engine whose fundamentals are tied to demographics and innovation, not GPU shipment cycles. That barbell should make drawdowns like this quarter’s -8.7% easier to digest.

Looking ahead, don’t expect a wholesale rotation out of tech unless the thesis on AI infrastructure truly breaks. More likely, future quarters will show fine‑tuning at the edges — trimming lower‑conviction cyclicals and telecoms — to keep feeding capital into whatever sits at the intersection of global scale, pricing power, and secular growth, whether that’s cloud GPUs or next‑generation therapeutics.