Sanders Capital 2025Q4: 44 Holdings, $87B AUM, AI Stack Bets

Lewis Sanders has spent the back half of his career proving that "concentrated value" is not a contradiction. The 2025Q4 13F from Sanders Capital, LLC — the firm he founded in August 2009 after his run as CEO of AllianceBernstein — sits 44 positions deep against $86.82 billion of disclosed equity value. The top five names carry 42.7% of the portfolio; the top ten carry 61.8%. There is no closet indexing here. Every position is doing real work.

The Q4 book is also unusually self-consistent. The top two holdings — Alphabet at 11.53% and Taiwan Semiconductor at 11.19% — are two ways of expressing the same view: that AI demand is sticky, but that the most defensible economic rent in the AI stack lives at the foundry and search/cloud level rather than at the model-training layer. The portfolio's third-through-fifth weights (Meta, Microsoft, HCA Healthcare) reinforce that pricing-power-with-pricing-discipline thesis, and the late-2025 addition of a 5.98% beneficial-ownership block in Seagate Technology — filed via Schedule 13G/A on April 16, 2026 — extends it into the AI-storage layer.

The portfolio at a glance

The full top-eight weight stack, sized at 2025Q4 reported values:

The single largest active bet is the GOOG + TSM pair, jointly running 22.7% of the book. Sanders's track record at AllianceBernstein was built on long-horizon contrarian-value calls — he resisted the late-90s tech bubble and reaped the 2000-2002 unwind — and the current pair-trade implementation is recognizably in that tradition. Search and foundry are both businesses where competitive moats are widening with AI workloads rather than narrowing.

HCA Healthcare at 5.63% is the cleaner contrarian read. The position size implies Sanders Capital is buying the demographic-tailwind / aging-population narrative against the policy-overhang concerns that have kept the hospital operator's multiple capped for much of 2024-2025. The Bank of America position at 2.87% — sized roughly the same as Berkshire's BAC weight in its public 13Fs — implies a parallel read on the post-rate-hike normalization in regional banking earnings power.

Concentration as a discipline, not an accident

A 44-holding portfolio with 42.7% concentrated in five names looks aggressive against the multi-cap-value peer group, which typically runs 70-90 holdings with top-five weight closer to 20%. But the comparison is misleading. The 13F Insight platform tags Sanders Capital with a Whale Score of 81.5 — top-decile across the active-manager universe — driven precisely by the concentration. The firm publishes a hit rate (per filings on AUM growth) that suggests the concentrated book is the source of alpha, not a risk-management afterthought.

The concentration profile also explains the 2025 AUM trajectory. From $66.91 billion at the end of Q1 2025 to $86.82 billion at the end of Q4 — a 30% lift in three quarters, almost entirely from mark-to-market on the existing book rather than client inflows. The fund's GOOG, TSM, and META positions all printed 30%+ returns across that window. When the concentrated bets work, they compound the firm's economic capital faster than dilutive inflows would.

The Seagate block is the latest signal

Sanders Capital filed a Schedule 13G/A on April 16, 2026 reporting beneficial ownership of 13,035,275 shares of Seagate Technology Holdings — a 5.98% stake with 7,510,117 shares carrying shared voting power. The 13G/A position implies Sanders Capital has now crossed the 5% threshold on STX, which triggers a structural change in how the firm has to disclose subsequent moves on the name.

The thesis is the AI-storage adjunct to the HBM supercycle: as hyperscaler AI training and inference workloads ramp, the volume of cold-tier and archival storage demand grows in lockstep — and Seagate's positioning in high-capacity HAMR drives gives it a structural edge in that market. Cross-referencing the position with the 11.19% TSM weight, Sanders Capital is now expressing the AI thesis across three layers: foundry (TSM), search/cloud (GOOG), and high-capacity storage (STX). The model-training layer (NVDA, which we don't see in the top 20) is conspicuously absent. The implicit Sanders Capital house view is that the model-training spend is over-capitalized and the durable economics live elsewhere in the stack.

The AUM history, in context

Sanders Capital's reported 13F AUM across 2.5 years of recent history:

The line tells the story: steady $50-60B base across 2023, a structural rise to $65-75B across the 2024 AI re-rating, and the $86B Q4 2025 print that puts the firm in the top-100 13F filers by AUM. With 44 positions deep, that translates to roughly $1.97 billion in average position size — a sizing that affords Sanders Capital meaningful influence on the names it carries, without requiring activist engagement.

Cross-checking the conviction read

Four facts anchor the Sanders Capital institutional read:

• $86.82B AUM: 30% growth across 2025 driven primarily by mark-to-market, not inflows.
• 44 holdings: Top-decile concentration discipline across the active-manager universe.
• 42.7% top-5 weight: No closet-indexing. Every top position is doing real work.
• WhaleScore 81.5: Top-decile institutional conviction rating per the 13F Insight active-manager score.

The portfolio also implicitly answers the question every multi-cap manager is asked: where in the AI stack is the durable economic rent? Sanders Capital's bet — foundry (TSM 11.2%), search/cloud (GOOG 11.5% + MSFT 6.8% + META 7.6%), storage (STX 5.5%), with NVDA and the AI-trainer layer absent — implies the answer is "everywhere except the trainer layer." That is a contrarian position against the consensus narrative of the trailing 18 months. Whether it converts to outperformance over the next 18 will be the most important test of Sanders's late-career strategy.

What to watch from here

• 2026Q1 13F filing (mid-May 2026): Whether the GOOG + TSM pair holds at 22.7% combined weight or trims. Sanders Capital's rebalancing cadence is typically quarterly.
• STX 13G/A amendments: Any subsequent amendment moving the Seagate stake above or below 5.98% would signal the firm's conviction trajectory on the AI-storage thesis.
• HCA conviction read: Whether the 5.63% Healthcare position survives the next policy-cycle headlines.

The full Sanders Capital filing history and signal flow tracks on the Sanders Capital filer page. The broader concentrated-value cohort — and the institutional read on companion AI-infrastructure names like Taiwan Semiconductor and Alphabet — is on the institutional signal feed. For a primer on how to evaluate concentration as a discipline versus a risk factor, see the explainer hub.