National Pension Service 13F Portfolio

The biggest dollar adds are almost embarrassingly straightforward: keep pressing the winners that own the rails of digital demand and U.S. consumption rather than searching for the next hero stock.

• Apple, Microsoft, and Alphabet (GOOGL line) all saw meaningful adds, with Apple’s stake up 1.6%, Microsoft up 2.2%, and GOOGL up 2.4%. These aren’t speculative reversals; they’re continued endorsement of compounders already showing triple-digit gains versus NPS’s cost basis.
• Amazon was increased by 2.6%, and Meta by 3.3%, extending the bet that cloud, advertising, and e-commerce volumes will keep compounding off a structurally higher digital baseline.
• Broadcom, Micron, and AMD were all nudged higher, signaling that NPS still wants exposure to AI hardware, but prefers balance-sheet-resilient suppliers and memory leverage over blindly adding to the most crowded GPU trade.
• On the consumer side, Walmart and Costco saw some of the fastest percentage adds in the top-50, with PG, Home Depot, and Netflix also increased. That combination is a clear call that real-world spend — from staples and warehouse clubs to housing-related DIY and streaming — remains underpriced versus the earnings power NPS sees.

This "platform-plus-consumer" bias says they’re betting the AI boom will translate into broad demand and advertising spend, not just semiconductor multiple expansion.

The sells are measured, not panicked. NPS is shaving the most extended AI-adjacent winners at the margin, using strength as a liquidity event.

• Nvidia, still the largest single position at 6.56%, was trimmed by 0.9%. With the position up more than 400% versus cost, this is a risk-budget decision, not a thesis break.
• Palo Alto Networks took a more notable -9.1% cut, and Western Digital, Marvell, and Intel were all reduced by low-single-digit percentages. Together, this looks like a move away from higher-beta, more cyclical or security-specific AI plays toward more diversified earnings streams.
• The small reductions in GOOG (class C), Lam Research, and Goldman Sachs are micro-adjustments: NPS kept overall exposure but is clearly comfortable banking some gains and redeploying into names with cleaner demand visibility.

The key point: they are not de-risking AI at the sector level. They are simply trimming the sharpest edges of the trade and recycling into platform behemoths and real-economy beneficiaries they regard as less fragile if the AI narrative cools.

At the sector level, the book barely budged — technology ticked from 58.63% to 58.45%, health care from 6.86% to 6.89%, and consumer discretionary from 9.63% to 9.81%. The message is that allocation is stable, but security selection is doing the work.

Inside tech, NPS is subtly rebalancing from more idiosyncratic, higher-volatility names toward the mega-cap platforms and diversified chip suppliers. The adds to Apple, Microsoft, Alphabet (GOOGL), Meta, Broadcom, Micron, AMD, and Texas Instruments, alongside trims in Nvidia, Palo Alto, Marvell, Intel, Western Digital, and Lam Research, collectively flatten single-name risk without giving up the AI and cloud cycle.

Outside tech, consumer discretionary is the quiet winner. Increases in Amazon, Walmart, Costco, Procter & Gamble, Home Depot, and Netflix inch the sector higher, effectively expressing confidence that real disposable income and pricing power in staples and big-box retail will hold up.

Energy (via Exxon and Chevron), industrials (Tesla and Caterpillar), and health care (Eli Lilly, Johnson & Johnson, AbbVie, UnitedHealth, Merck, Philip Morris) all saw modest builds, nudging their combined weights higher as a hedge against inflation and policy uncertainty.

Unclassified sleeves — notably broad ETFs like PBUS, IVV, TGRW, INFO, and LSVD, plus Berkshire — remained unchanged, anchoring the portfolio in low-churn beta and external active managers.

The pattern over 2026-Q2 is that NPS is not chasing the latest AI headline; it is methodically re-underwriting the beneficiaries of a durable U.S. growth-and-consumer cycle funded by AI-driven market gains.

Expect them to keep three pillars intact. First, a dominant core in U.S. platform tech and AI infrastructure, with ongoing preference for diversified cash machines over narrow concept plays. Second, a steadily rising allocation to consumer, health care, and real assets that monetize both nominal growth and demographic tailwinds. Third, a meaningful ballast in ETFs and quasi-index proxies that dampen idiosyncratic risk.

If volatility returns to AI hardware or cybersecurity, this quarter’s trims suggest they will harvest gains rather than defend every basis point of exposure. Conversely, continued strength in U.S. consumption and services would likely see further incremental adds to Amazon, Walmart/Costco-style retailers, and health-care franchises rather than a wholesale sector shift.

For observers, the signal is clear: NPS is running a patient, barbelled book. AI hype is welcome as a source of mark-to-market gains, but the real bet is that those gains can be converted into long-duration exposure to the global platforms and consumer franchises that will still matter a decade from now.