Legal & General Group 13F Portfolio

The biggest dollar add is Western Digital, up 17.9% in shares and now a $1.60B, 0.33% holding. That is a bold call that AI’s next bottleneck is storage: hyperscale build‑outs will need far more capacity and bandwidth than the market is currently willing to pay for in equity valuations.

The same logic shows up in Marvell, where the fund lifted exposure by 4.0%. Marvell is a clean way to play data‑center networking and custom silicon — the connective tissue between GPUs rather than the GPUs themselves — signalling a belief that interconnect and bandwidth are underpriced legs of the AI stack.

Texas Instruments is another quiet but telling add, with shares up 1.5%. TI’s analog and embedded footprint ties into power management, sensing, and industrial automation — a “real economy” complement to cloud AI — suggesting the manager wants durable cash flows that still monetize semiconductor intensity.

Outside core tech, small increases in Tesla, Costco, and Bank of America are notable. Tesla (up 0.5% in shares) extends the bet that AI bleeds into autos and energy systems, Costco adds to best‑in‑class defensiveness with real pricing power, and Bank of America reflects selective confidence in scale US banks as beneficiaries of a still‑resilient macro backdrop.

The funding sources are as instructive as the buys. Nvidia, still the single largest line at 6.63% and $31.8B, was trimmed by 6.1% — a textbook partial harvest after a 539.4% gain versus the fund’s average cost. Apple and Microsoft saw similar, if milder, treatment: Apple shares were cut 2.7%, Microsoft 2.4%, both off enormous embedded gains.

Within semis, the fund is clearly upgrading where it wants cyclicality. Intel was hit with an 11.5% reduction, and Lam Research and Applied Materials were each trimmed around 5–6%. Taken together, this is a move away from more mature or capacity‑heavy exposures toward names like Marvell, Western Digital, and Texas Instruments, where incremental AI capex can drive more operating leverage per dollar.

Outside technology, the sharpest signal is in energy and old‑line defensives. Exxon Mobil was slashed 16.3%, and Coke, Procter & Gamble, and Home Depot all absorbed mid‑ to high‑single‑digit cuts in shares. In financials, JPMorgan and Morgan Stanley were pared back meaningfully, even as Bank of America was added to, underscoring a preference for specific balance sheet and funding profiles rather than blanket sector exposure.

Technology is an overwhelming 63.56% of the disclosed book, up slightly from an estimated 63.31%. That stability at a very high level masks a clear internal re‑tilt: away from the most crowded, fully rerated mega‑caps and toward under‑owned enablers of AI infrastructure — semis, storage, and security.

Health care has quietly edged up to 8.86% from an estimated 8.78%, via a diversified basket rather than a single hero name: Eli Lilly, Johnson & Johnson, AstraZeneca, AbbVie, UnitedHealth, and Merck. This looks like a volatility dampener and inflation hedge dropped into an otherwise growth‑heavy book, especially given Lilly’s very large unrealized gain.

Finance has drifted down to 6.37% from 6.48%, but the nuance is intra‑sector. The fund is trimming JPMorgan, Morgan Stanley, Citi, and Royal Bank of Canada, while nudging up Bank of America, effectively consolidating big‑bank risk rather than exiting it. Energy (Exxon) and consumer staples (Coke) are being moderated, while industrials tick up to 4.24% on the back of Tesla and a still‑sizable Caterpillar stake, aligning cyclical exposure with AI, infrastructure, and reshoring themes.

The result is a book that’s not just “overweight tech,” but specifically wired to semis and infrastructure that monetize AI compute over many cycles, with healthcare and select financials as ballast.

Taken together, the quarter telegraphs a manager that believes the AI cycle is real, long, and only partially priced — but that the most obvious winners have already done a lot of the easy work. Hence trimming Nvidia, Apple, Microsoft, Alphabet, and Intel, while using that cash to deepen exposure to memory, storage, networking, and analog names that will feel the demand curve later and harder.

There is no evidence of a macro panic here. The slight reductions in energy, consumer staples, and some banks aren’t a retreat from risk; they’re a rotation from relatively static cash‑flow stories into operating leverage where incremental data‑center dollars translate more cleanly into earnings growth.

The incremental adds to Tesla and Costco, plus the durable allocation to healthcare majors, suggest the fund still wants diversification that can defend in a downturn without giving up structural growth. Expect future quarters to keep this pattern: modest trimming of mega‑cap AI and broad beta, with proceeds re‑cycled into the underbuilt parts of the AI supply chain and high‑quality cyclicals that benefit from digitization.

For observers, the key tell to watch will be whether Western Digital, Marvell, and Texas Instruments continue to climb the rankings from their current sub‑0.5% slots. If they do, this book will be less about owning “AI brands” and more about owning the infrastructure bottlenecks that AI spending can’t avoid.