Q1 2026 13F Roundup: What 91 Superinvestors Bought and Sold

Ninety-one of the ninety-four managers tracked on this site filed a 13F for both the March 2026 quarter and the December 2025 quarter, which makes Q1 2026 the most recent quarter where a genuine cross-manager comparison is possible. This is what those 182 filings say when you line them up.

Two things stand out before any individual name does. The first is that the single most widely "bought" stock of the quarter was not bought by anybody. The second is that the biggest dollar move in the whole dataset came from the manager least associated with buying technology.

Coverage, and why this is a Q1 roundup

13F filings are due 45 days after quarter end. The Q2 2026 deadline is August 14, 2026, and as of this article's publication date only three of the 94 managers we track had filed for the June quarter. A cross-manager roundup built on three filings would be noise, so this piece covers Q1 2026 instead. The Q2 version follows once the August deadline passes.

Within Q1, coverage is close to complete: 91 of the 94 managers filed both quarters and are included below. The three left out are all stale filers whose most recent 13F predates the March quarter entirely: Michael Burry and Jeff Ubben last filed for September 2025, and Whitney Tilson has not filed since 2015. Everything in this article is calculated only on the 91 with a clean quarter-over-quarter pair.

The most-bought list is a trap

Rank every stock by how many managers opened a brand-new position in it during Q1, and two names sit at the top that have no business being called consensus buys.

Sunbelt Rentals: ten new holders, zero buyers

Sunbelt Rentals appears in ten managers' Q1 filings. It appears in nobody's Q4 filing, or any filing before that, in the entire history we hold. Not one manager sold a share. Ten separate value investors independently discovering the same equipment rental company in the same 90 days, all at once, with no seller on the other side, is not how markets work.

What actually happened is a listing change. Ashtead Group, the UK-listed parent of the Sunbelt Rentals business, moved its primary listing to the United States and now reports under the Sunbelt Rentals name with a US CUSIP. That made it a 13F-reportable security for the first time. Managers who had owned it for years through its London listing had nothing to report before Q1. Now they do.

Our own data shows the seam. Tom Russo held 6,111,758 shares of Ashtead's UK-listed line at the end of December and 6,049,133 shares of Sunbelt Rentals at the end of March, with the Ashtead line gone. That is a swap, not a purchase. The other nine holders never showed an Ashtead line at all, because a plain London listing is outside the 13F perimeter entirely.

The position sizes make the point about scale. Dodge & Cox reported 53.1 million shares worth $3.46 billion, which was the largest single "new position" opened by any manager in the quarter. It was not a decision made in Q1.

Waters: six of seven new holders got the shares in the mail

Waters Corp went from three holders to ten. That looks like a stampede into lab instruments until you check who the seven new holders are. Six of them, including First Eagle, Bill Nygren, Steven Romick, Robert Olstein, FPA Queens Road and Mairs & Power, also hold Becton Dickinson. Waters combined with BD's biosciences and diagnostics business in a Reverse Morris Trust, so BD shareholders received Waters stock while keeping their BD shares.

The tell is in the BD numbers: the count of managers holding BD did not fall, and their aggregate share count barely moved, going from 16.53 million to 16.22 million. Nobody swapped out of one into the other. They were handed the second one.

Exactly one of the seven, Andreas Halvorsen at Viking Global, opened Waters without holding BD. That is the only new Waters position in the quarter that reflects a decision.

Both cases are worth remembering, because the pattern repeats. Any screen that counts "number of managers who added a position" will put corporate actions at the top of its list, every quarter, forever. Spinoffs, redomiciliations, share-class conversions and merger consideration all look identical to buying if all you compare is two lists of holdings.

What managers actually bought

Strip out the corporate actions and the picture is less dramatic and more useful. Here are the stocks with the most managers on the buy side, along with how many were selling the same name in the same quarter.

Read the third column. Microsoft was the most-bought stock of the quarter and also among the most-sold, with 21 managers on one side and 24 on the other. Only Alphabet's two share classes had more sellers. Amazon was 19 against 18. Visa was 16 against 16. These are not consensus trades. They are crowded names where the crowd disagreed, which is a different and much weaker signal than it looks like in a headline.

The genuinely lopsided names are further down. Micron had six managers buying and none selling, the only name in the top twenty with a clean sweep. David Tepper and Ray Dalio both added to existing stakes, and Philippe Laffont and Stanley Druckenmiller opened new ones. Netflix ran nine buyers to three sellers, MercadoLibre nine to three, and Spotify eight to three.

Among brand-new positions, the widest agreement outside the corporate actions was ASML, opened by five managers including Laffont, Daniel Loeb, Dan Sundheim and Wallace Weitz. Four managers each opened AstraZeneca, TSMC, LyondellBasell, Coupang, Spotify, SanDisk, Micron, Versant Media and Alphabet's A shares.

Alphabet: 28 sellers and one very large buyer

Alphabet's A shares were the single most-sold name of the quarter by manager count, with 27 trims and one full exit against nine buyers. The C shares tell the same story: 24 trims, one exit, eight buyers. On a naive reading, superinvestors spent the quarter getting out of Alphabet.

They did not. Warren Buffett took Berkshire's A-share position from 17,846,142 shares to 54,249,798, a tripling that added just over $10 billion of reported value and made Alphabet Berkshire's seventh-largest US equity holding at roughly 5.9 percent of the portfolio. He also opened a separate 3,585,215-share position in the C shares. That single manager's buying outweighed the combined trimming of everyone else.

Going the other way, Bill Ackman cut Pershing Square's C-share stake from 6,163,871 shares to 311,726, a 95 percent reduction, and opened a $2.09 billion position in Microsoft. Two very concentrated managers took opposite sides of the same trade in the same quarter, which is a better reminder of what 13F data is worth than any consensus table.

The clearest selling: consumer credit and big banks

Capital One is the closest thing to a unanimous verdict in the data. Eighteen managers reduced or eliminated it and two added, a ratio no other large-cap name comes close to. John Armitage exited 2,069,130 shares outright. Daniel Loeb cut from 1,100,000 shares to 140,000. Halvorsen roughly halved a 2.95 million share position. The two buyers were Dan Sundheim and Leon Cooperman.

Wells Fargo is nearly as one-sided in a quieter way. Seventeen managers trimmed it and not one added a share. There were no full exits, so this reads as broad de-risking rather than a thesis change: Dodge & Cox went from 32.31 million shares to 31.17 million, Nygren from 12.90 million to 10.98 million, Richard Pzena from 9.98 million to 8.77 million. Everyone shaved. Nobody left.

UnitedHealth: the most genuine disagreement of the quarter

UnitedHealth shows seven full exits, which sounds decisive until you look at the sizes. Only three were real positions: Buffett's 5,039,564 shares worth $1.66 billion, Halvorsen's 1,197,273 shares, and Torray's 59,153. The other four were rump holdings of a few thousand shares or fewer, the largest of them 6,170 shares, the kind of residue that gets swept out in routine housekeeping. Counting them equally with Berkshire's exit inflates the signal.

Meanwhile eleven managers bought. Dodge & Cox added to a position that ended the quarter at 9,466,756 shares and $2.56 billion, and Boykin Curry at Eagle Capital added to a $1.59 billion stake. Two of the largest and most patient shops on the list were buying what Berkshire was selling. That is a real disagreement between serious investors, and it is more interesting than any of the crowded tech names.

Berkshire's housekeeping quarter

Buffett's Q1 was unusually busy on the sell side. Berkshire fully exited sixteen positions, including Visa ($2.91 billion), Mastercard ($2.28 billion), UnitedHealth ($1.66 billion), Domino's Pizza ($1.40 billion), Aon ($1.27 billion) and Pool Corp ($702 million). The remaining ten include a full exit from Amazon ($525 million) and from HEICO ($327 million), alongside Allegion, Charter, Lamar Advertising, Diageo and a cluster of small Liberty Media stubs. Constellation Brands went from 13,000,000 shares to 632,890, effectively an exit executed over two lines. Chevron came down 35 percent, from 130.2 million shares to 84.4 million.

The buying was concentrated in three places: the Alphabet tripling described above, a new 39,809,456-share position in Delta Air Lines worth $2.65 billion, and a tripling of the New York Times stake from 5,065,744 shares to 15,146,535. Apple was untouched at 227,917,808 shares, still 22.0 percent of the reported portfolio. The full breakdown is on the Berkshire Q1 2026 recap page.

Elsewhere among the largest single moves, Chris Hohn cut TCI's Microsoft position from 16,781,953 shares to 2,728,412, and the Gates Foundation Trust exited Microsoft entirely, selling all 7,691,207 shares. The Trust also trimmed Berkshire B shares from 19,406,764 to 17,048,304. John Armitage exited his 1,748,656-share Microsoft position outright. Between Hohn, the Gates Trust, Chase Coleman, Laffont and Armitage, five managers each took more than half of a nine or ten-figure Microsoft stake off the table, in the same quarter Ackman opened one.

Prices did most of the work

Aggregate reported value across the 91 matched managers fell from $1.296 trillion to $1.198 trillion, down 7.6 percent. It would be easy to read that as heavy net selling. It mostly is not.

Because filers report both share counts and market values, you can isolate price. Across the 931 positions where a manager's share count was identical in both quarters, aggregate value fell only 2.9 percent, from $294.7 billion to $286.3 billion. The market as a whole was down modestly. The damage was concentrated in exactly the names these managers hold most.

Implied quarter-end prices, derived from filers' own reported values, show where:

• Microsoft: $483.62 to $370.17, down 23 percent
• Capital One: $242.36 to $182.43, down 25 percent
• American Express: $369.95 to $302.48, down 18 percent
• UnitedHealth: $330.11 to $270.59, down 18 percent
• S&P Global: $522.59 to $425.34, down 19 percent
• Visa: $350.71 to $302.24, down 14 percent
• Meta: $660.09 to $572.13, down 13 percent
• Alphabet A: $313.00 to $287.56, down 8 percent
• Occidental Petroleum: $41.12 to $65.00, up 58 percent

Occidental is the outlier in both directions. It was the largest positive contributor to aggregate value in the entire dataset, roughly $7.8 billion, and essentially none of that was buying. Berkshire's 264,941,431 shares did not change at all. This is the main reason we do not publish a "biggest dollar inflows" table: in a quarter with double-digit price moves, dollar change measures the market, not the manager.

Breadth: slightly more building than pruning

Across the 91 managers there were 982 brand-new positions and 923 complete exits, against 2,038 adds and 2,527 trims to existing holdings. Forty-four managers ended the quarter holding more distinct positions than they started with, 34 held fewer, and 13 were unchanged. Portfolios got marginally wider while prices fell, which is roughly what you would expect from a group that mostly buys weakness.

You can follow the same underlying moves as they land on the activity feed, or see the current standings on most bought and most sold.

How this was calculated

Every figure above comes from SEC Form 13F filings for period 2026-03-31 compared against the same manager's 2025-12-31 filing. Only managers with a filing in both quarters are included, which is 91 of 94.

Positions are matched on CUSIP and security type rather than ticker symbol, because ticker symbols change, share classes share names, and symbol mapping can collapse two unrelated issuers into one row. Only common stock is counted. Puts, calls and debt lines are excluded, since option rows report the value of the underlying shares rather than capital at risk and would badly distort any "biggest position" ranking. A small number of symbols were left out of the leaderboards where our own symbol mapping is ambiguous enough that a count could not be trusted.

Implied prices are a filer's reported market value divided by reported share count, not a market data feed. They are consistent across dozens of independent filers for each name, which is why we are comfortable citing them, but they are quarter-end snapshots and nothing more.

We do not publish a sector rotation table for this quarter. Our sector mapping leaves roughly 40 percent of reported value unclassified, and a rotation number computed on the classified 60 percent would be more misleading than useful.

What this quarter is worth

Three honest conclusions. Managers disagreed sharply on the largest technology names, so anything framed as a "consensus" position in Microsoft, Amazon or Visa is an artifact of counting. There was real one-sided selling in consumer credit and banks, with Capital One at eighteen sellers to two buyers the clearest example. And the loudest apparent buys of the quarter, Sunbelt Rentals and Waters, were paperwork.

None of that is a reason to buy or sell anything. These filings describe positions as of March 31, disclosed up to 45 days later, and read here four months on. Their value is as a source of questions. Why were Dodge & Cox and Eagle Capital buying the healthcare insurer Berkshire was leaving? That is a better starting point than any table on this page.