13F Season: What the Smart Money Owned (As of 12/31/25) — and What to Do With It

If 13F season feels like “following the smart money,” that’s because it is — with two important caveats: it’s delayed, and it’s incomplete. The filings can still be useful for spotting institutional conviction and broader positioning themes, but investors (and especially traders) need to treat this as a map of where managers were, not where they are now.

What a 13F Is (and Why the Fine Print Matters)

A Form 13F is a quarterly SEC filing required for institutional investment managers with at least $100 million in qualifying assets. It discloses long positions in U.S.-listed equities (and certain other reportable securities) as of the quarter-end date.

The limitations matter. A 13F does not show shorts, most derivatives exposure, or what managers did after quarter-end. That’s why the timestamp is the headline: this data is as of 12/31/25, meaning it can be stale the moment it prints — particularly in fast-moving sectors like AI, software, semis, and high-multiple consumer.

The Big Picture: Platform Winners, Select Cyclicals, and Some De-Risking

The overarching pattern is not a single sector stampede, but a preference for high-quality, scale-driven “platform” businesses alongside selective cyclicals. The counterbalance is that several managers appeared to trim crowded megacap/AI exposure into year-end or rotate within it rather than add broadly.

In plain English: some managers leaned into durability and “winner-takes-most” economics, while others reduced concentration risk, harvested gains, or repositioned for a different 2026 setup.

Communication Platforms and Mega-Cap Tech: META Was a Crowd Add… With Notable Exceptions

Meta Platforms was one of the clearest “multiple managers touched this” names in your notes.

Bill Ackman’s Pershing Square initiated a large new META position (2.67 million shares) while also adding to Amazon and dramatically cutting Alphabet exposure. ValueAct increased META (+36% to roughly 1.05 million shares). David Tepper’s Appaloosa also added META (+62% to about 600K shares).

But it wasn’t one-way traffic. Dan Loeb’s Third Point exited META, which stands out precisely because it runs counter to the broader sponsorship trend.

Takeaway: META looked like a popular year-end “platform + AI adjacency” expression. The presence of at least one high-profile exit suggests META also functioned as a funding source for some books or was considered fully valued versus alternative opportunities.

E-Commerce and Cloud: Amazon Was Both a Buy and a Funding Source

Amazon’s activity showed classic “everyone owns it, but not the same way” behavior.

Pershing Square added aggressively (+65%). Baupost initiated a new AMZN position (2.1 million shares). Viking initiated AMZN as well (~3.13 million shares). On the flip side, Berkshire Hathaway reduced AMZN sharply (-77%), while Third Point and Appaloosa trimmed to varying degrees.

Takeaway: AMZN remains a core long-duration compounder in many portfolios, but it also gets used as a rebalancing lever. When both new buying and heavy trimming appear in the same quarter, it often signals a mature “core position” name where managers are expressing different risk budgets rather than different views on the business.

Housing and Financial Services: Rocket Companies Became a Surprisingly Shared Bet

Rocket Companies (RKT) was one of the more interesting overlaps because it appeared as a meaningful add across very different styles.

ValueAct increased RKT (+55% to ~14 million shares). Third Point also added aggressively (+138% to ~5.5 million shares). Leon Cooperman initiated a very large new position (21.02 million shares).

Takeaway: When multiple managers size up the same mid-cap name into quarter-end, it often reflects a shared thesis — either on housing/mortgage cyclicality, rate dynamics, market-share positioning, or company-specific catalysts. Regardless of the exact rationale, the clustering is notable and worth tracking for follow-through in subsequent filings and price action.

Industrials and Infrastructure: UNP and DHR Drew Fresh Interest

Several managers leaned into industrial/steady cash-flow franchises.

Third Point increased Union Pacific (+107%) and Danaher (+1100% from a small base), alongside other adds that fit the “durable operators” bucket. Baupost also increased UNP. In general, the filings showed scattered signs of managers favoring businesses with resilience, pricing power, and long-duration compounding — but without the same multiple risk as frothier software.

Takeaway: This reads like a year-end tilt toward “real economy” quality — a hedge against software multiple compression and macro uncertainty, rather than a pure growth bet.

Consumer Discretionary: Chipotle Was a Battlefield Name

Chipotle (CMG) is a perfect example of why 13Fs are better at showing disagreement than consensus.

Third Point initiated a large CMG position (~4.73 million shares). Pershing Square, meanwhile, exited CMG in size. Viking also exited CMG.

Takeaway: When you see large buying and large selling in the same quarter, it’s often a rotation battleground rather than a clean fundamental consensus. Practically, these names can trade choppy because flows and positioning can dominate the narrative more than incremental fundamentals.

Semiconductors and AI Infrastructure: Selective Rotations and Risk Management

The AI/semi complex showed more nuance than a simple “risk-on” stampede. There were selective adds, but also notable trims and even exits.

Third Point made a small add to NVDA (+3%). Appaloosa trimmed NVDA (-10%) and cut AMD sharply (-65%) while adding Micron (+22%). Druckenmiller exited SMTC and trimmed TSM. Tiger Global trimmed NVDA, MSFT, AMZN, and TSM. SoftBank exited NVDA entirely — a headline-worthy move given the prior scale.

Takeaway: Into 12/31/25, a number of managers appeared to be managing concentration and crowding in AI bellwethers, rotating within the complex (e.g., MU vs AMD), or taking down exposure where valuations and positioning risk were perceived to be high.

Energy, Materials, and Metals: Divergent Views, With Some De-Emphasis on Gold

A few filings suggested reduced enthusiasm for parts of the metals complex. Greenlight reduced positions including GLD and GDX, and exited OIH, while initiating or adding in a mixed basket spanning energy-related equities and idiosyncratic value names.

Meanwhile, Berkshire increased CVX, which reads more like a targeted conviction position than a “sector call.”

Takeaway: This bucket didn’t show a single unified institutional “flow,” but there were signs of selective positioning rather than broad commodity beta.

Activists and Event-Driven: Focused Bets and Big Cuts

Some of the cleanest “conviction tells” often come from outsized reductions.

Trian cut Invesco (IVZ) by roughly 80%, which rarely looks like simple trimming. Starboard initiated CWAN and FLR while adding to RIOT — a mix of activist/event-driven posture with some higher beta spice.

Takeaway: Activist books can create idiosyncratic “flow stories” that don’t map neatly onto sector narratives, but the position sizing (especially major reductions) can be a useful flag.

How to Use This (Without Letting the Timestamp Trick You)

The right way to treat these filings is as a source of themes and watchlists, not as a trading system.

First, anchor on overlaps: repeated activity across managers tends to be more durable than a one-off move. In your set, META and RKT were among the clearest overlaps. Second, mark battlefield names where big adds and big exits coexist (CMG, AMZN, META); those often bring volatility as flows fight. Third, keep the timestamp front and center: these are holdings as of 12/31/25, and many of these portfolios likely changed meaningfully during January and February.

Bottom Line

These 13Fs suggest year-end positioning leaned toward platform winners and selective cyclicals (mortgage, industrial durability, and consumer experiences), while several managers trimmed or rotated within crowded mega-cap and AI exposures. The setup for readers is straightforward: use this as a roadmap for what institutions wanted to own heading into 2026 — but don’t confuse it for what they own today.