The deadline was May 15. The filings went public. And within 48 hours, it became clear that some of the most-watched portfolios on Wall Street had made significant, and in several cases, contradictory moves during the first quarter of 2026.

This is 13F season. Every institutional manager running more than $100 million in qualifying U.S. equity positions is required to file a Form 13F with the SEC within 45 days of quarter end — meaning what you’re reading about now reflects positions as of March 31. There’s a built-in lag. That’s the first thing to understand before drawing any conclusions. What happened between April 1 and today is not in these filings, and in some cases — Alphabet being the most notable example — the price of the stocks these funds sold has moved substantially since the exits occurred.

With that context in place, here’s what the filings actually showed.

Ackman Exits Alphabet, Enters Microsoft

The headline trade of the week belonged to Bill Ackman and Pershing Square Capital Management. On May 15, ahead of the firm’s quarterly filing, Ackman disclosed a new position in Microsoft ($MSFT) — and simultaneously revealed that Pershing had nearly exited its Alphabet stake.

The numbers: Pershing Square disclosed ownership of approximately 5.65 million Microsoft shares valued at roughly $2.09 billion at the end of Q1 2026. By market close on May 15, that position had climbed to approximately $2.3 billion. Ackman said the firm began accumulating shares in February after Microsoft’s stock declined following its fiscal second-quarter earnings, with the position established at around 21 times forward earnings — broadly in line with the broader market multiple and below Microsoft’s own historical trading average.

On the other side: Pershing reduced its Alphabet Class C shares from more than 6.1 million shares in Q4 2025 to roughly 312,000 shares in Q1 2026. Class A holdings fell from approximately 678,000 shares to about 32,000 shares over the same period — a 95% reduction across both share classes. In a follow-up post the next day, Ackman was careful to frame the Alphabet sale as a capital reallocation decision rather than a bearish view on the company: he described Alphabet as a finite-capital trade made to fund the Microsoft position, not a vote against Google’s long-term prospects.

The Microsoft thesis rests on three legs. Azure cloud revenue, which grew 40% year-over-year in Q3 FY2026 according to the company’s own 10-Q and 8-K filings — a sequential acceleration from 31% growth in the same period the prior year. Microsoft 365 Copilot monetization, which Ackman values at $30 per user per month in enterprise adoption. And the firm’s OpenAI stake, which Ackman suggested could be worth $200 billion on its own — an asset he described as undervalued by the market.

Microsoft’s most recent quarterly report confirmed Q3 FY2026 total revenue of $82.9 billion, up 18% year-over-year, with EPS of $4.27, up 21%. The Intelligent Cloud segment — which houses Azure — posted $34.7 billion in revenue, up 30%. Capital expenditure came in at $31.9 billion for the quarter, a figure that initially spooked the market when guidance was raised, contributing to the stock’s 15% year-to-date decline that created the entry point Ackman exploited.

The pattern here is worth noting. Ackman bought Alphabet in 2023 when the AI narrative around Google was at its weakest. He bought Amazon and Meta during similar periods of skepticism. Now he’s applying the same framework to Microsoft: find a dominant, durable franchise at a moment of temporary market doubt, and size in. Whether the thesis holds depends on how quickly Copilot revenue scales and whether Azure capacity constraints — which Microsoft management acknowledged will persist through 2026 — resolve without compressing margins further.

Druckenmiller Moves Fast — Again

If Ackman’s move was methodical, Druckenmiller’s was characteristically aggressive.

Duquesne Family Office — Druckenmiller’s vehicle, which has generated three-year annualized returns of approximately 39% on an equal-weighted basis — sold its entire 385,000-share Alphabet Class A position during Q1 2026. The striking part: that stake had been built up in Q4 2025, when Druckenmiller increased the position by 277% from 102,000 shares to 385,000. He held it for one quarter, the position appreciated more than 50%, and he exited. Between January 1 and March 31, Alphabet shares had fallen 8%, meaning the exit came during a down period — and since then, Alphabet closed at $396.78 as of May 18, up 27% for the year. The missed gain has already drawn attention.

Duquesne also cut its Amazon position from 737,940 shares down to just 9,539 shares — a 99% reduction. That move is consistent with a broader rotation out of large-cap consumer and internet platforms and into more targeted AI infrastructure plays.

Where the capital went: Druckenmiller opened a fresh 195,955-share Broadcom ($AVGO) stake, alongside a new position in cancer-diagnostics company Caris Life Sciences (1.89 million shares), clinical-stage drug developer Revolution Medicines (315,860 shares), and smaller new stakes in Seagate Technology (50,700 shares), Twilio (182,000 shares), and Humana (138,000 shares). The Broadcom bet is the most structurally legible — Broadcom has become one of the most important custom silicon suppliers in the AI build-out, behind Nvidia, and the stock currently trades at roughly 31 times forward earnings with a June 3 earnings report as the next major catalyst. Watch the $10.7 billion Q2 AI revenue guidance line; a miss against that number compresses the multiple quickly.

Slight tangent, but it’s worth sitting with: Druckenmiller built and exited the Alphabet position in a single quarter, reportedly locking in meaningful gains, then rotated into Broadcom — which had already run — at a much higher multiple. Whether that’s macro-driven timing or something more granular in Broadcom’s AI backlog is impossible to know from the 13F alone. Duquesne did not comment publicly on the reasons for the Alphabet exit.

Tepper Bets on Amazon. Berkshire Walks Away.

This one is worth reading carefully, because two of the most respected capital allocators in the world took exactly opposite positions in the same stock during the same quarter.

David Tepper’s Appaloosa Management nearly doubled its Amazon stake in Q1 2026 — a 98% increase that added 2.1 million shares, bringing the total to 4.3 million shares valued at roughly $900 million. Amazon now represents approximately 15.2% of Appaloosa’s disclosed portfolio, up from 7.3% at year-end 2025, making it the fund’s largest disclosed holding by a substantial margin. Tepper also boosted his Uber stake by 242%, raised Vistra Energy by 114%, added 18% to Taiwan Semiconductor, initiated a sizable new position in Sandisk valued at approximately $179 million, and increased Micron by 11%. The direction is clear: AI infrastructure, data storage, and power generation.

Meanwhile, Greg Abel — Warren Buffett’s hand-picked successor at Berkshire Hathaway — fully exited the conglomerate’s Amazon position during Q1. The likely explanation is administrative rather than analytical: Todd Combs, who had managed the Amazon position, left Berkshire in January 2026 to lead JPMorgan Chase’s $10 billion Strategic Investment Group. Berkshire appears to have unwound the positions Combs oversaw following his departure. That context matters when interpreting the exit — this may not reflect a view on Amazon’s fundamentals so much as a restructuring of the book under new management.

Amazon stock has risen more than 20% since the end of Q1, suggesting Tepper’s timing was sharper. The February pullback — which pushed shares below $200 — appears to have been the entry window. Amazon delivered its highest operating margin ever in Q1 2026, with AWS continuing to hold the leading share of the cloud services market and the company’s AI infrastructure investment beginning to produce measurable returns.

Other Moves Worth Tracking

TCI Fund Management (Chris Hohn) — The London-based fund, which manages a $45 billion disclosed U.S. equity portfolio, made a notable addition in Q1: a new Class A Alphabet position, while also adding to its Class C shares by 16.55%. TCI also expanded S&P Global by 19%, added 9.9% to Visa, and increased Moody’s by 7.71%. At the same time, TCI slashed its Microsoft stake from 16.78 million shares down to 2.73 million. Hohn’s top five holdings remain General Electric (29.85% of portfolio), Visa (20.39%), Moody’s (13.84%), S&P Global (13.22%), and Canadian Pacific Kansas City (8.1%). This is a concentrated, quality-compounder book — and the Microsoft trim alongside a new Alphabet stake is a direct mirror image of what Ackman did, but in reverse.

Third Point (Daniel Loeb) — Loeb’s $2.1 billion disclosed portfolio made several new entries in Q1: Meta Platforms, Alphabet, aerospace components maker TransDigm, crypto miner Hut 8, and the SPDR Gold ETF. Third Point simultaneously exited PG&E, Microsoft, and Brookfield. The combination of AI platform exposure plus a gold ETF position alongside a crypto miner reflects something close to a barbell: high-growth AI names on one end, hard-asset inflation hedges on the other.

Searchlight Capital and INDUS Capital — Among the smaller but structurally meaningful moves in the Q1 filings: Searchlight Capital fully exited its 2.27 million share stake in Uniti Group ($UNIT), worth approximately $17.82 million, a shift representing 16.29% of Searchlight’s reportable 13F assets. On the other side of the ledger, INDUS Capital added 126,500 H World Group ($HTHT) shares worth roughly $6.45 million, bringing its total HTHT position to 682,224 shares valued at $34.31 million — an 86.37% concentration in a single name. H World shares traded at $45.51 as of May 18, up 21.1% over the prior year. That level of single-name concentration is unusual for a fund required to file a 13F.

Options Market Context

The 13F filings only show long equity positions. What they don’t show — and this is the part most investors skip — are the shorts, the options hedges, and the cash positions. A long disclosed at 5% of a portfolio might be fully offset by a put position or a short in the same sector that simply doesn’t appear on the form. Druckenmiller’s Broadcom buy, for instance, may be paired with protection against the multiple compressing around the June 3 earnings date.

For traders looking at the options market around the names that dominated the filings this week, a few structural observations worth considering:

• Microsoft ($MSFT): The stock is up from Ackman’s estimated entry range of $381–$430 (February purchase window), trading at approximately $418 as of mid-May. If you believe the Azure acceleration is durable and Copilot monetization inflects in the next two quarters, a defined-risk call spread targeting the prior highs captures the upside without full equity exposure. For traders expecting continued capex-driven multiple compression, a long put or put spread with a June or September expiry gives defined downside exposure against the thesis.
• Amazon ($AMZN): Up more than 20% since end of Q1. IV rank and put/call behavior have been shifting toward calls as the stock has recovered from the February lows. For those who believe the AWS margin expansion cycle continues, long calls or a call spread structure above current levels captures additional upside. For those who think the 20%+ recovery has already priced in the AI capex thesis, a bear call spread or collar on an existing long position is a defined-risk hedge.
• Broadcom ($AVGO): Trading at approximately 31 times forward earnings with a June 3 earnings catalyst. Elevated IV around the earnings date typically makes long straddles or strangles relatively expensive. A defined-risk structure — such as a bull call spread for those expecting the AI silicon story to accelerate, or a bear put spread for those who think the multiple is stretched — limits both cost and exposure around the catalyst. Watch the Q2 AI revenue guidance against the $10.7 billion marker.

These are analytical frameworks, not instructions. If you believe X, a defined-risk structure would be structured as Y. The position sizing, strike selection, and expiry choice depend entirely on your own capital base, thesis duration, and risk tolerance.

What the Rotation Says

Pull back from the individual trades and look at the aggregate picture across the Q1 filings. Several things are happening simultaneously.

First, AI infrastructure is the dominant theme — but funds are increasingly distinguishing between layers of it. Ackman is buying the platform (Microsoft/Azure). Druckenmiller is buying the custom silicon (Broadcom). Tepper is buying the cloud infrastructure and data storage (Amazon, Sandisk, Taiwan Semiconductor). These are not the same bet dressed in different clothing. They reflect different views on where margin and growth will actually accrue inside the AI capital expenditure cycle.

Second, Alphabet is being actively debated. Ackman and Druckenmiller both reduced or exited. TCI and Third Point both initiated or added. The stock is up 27% year-to-date. Two highly experienced managers sold it, two others bought it, in the same quarter. That kind of divergence is worth tracking more closely than any individual trade.

Third, airlines are out. Tepper’s Appaloosa fully exited American, Delta, and United Airlines in Q1, with $217 million in American alone at year-end 2025 going to zero by March 31. The explanation given: elevated fuel costs tied to the ongoing Iran conflict and margin pressure across the sector. That’s a clean macro call, and it’s unambiguous.

The hard part is that by the time these filings are public, six weeks have passed. The stocks have moved. The hedge funds may have already changed their positions. The 13F is a rearview mirror — useful for pattern recognition, less useful as a real-time signal. What it does tell you is how some of the most analytically rigorous capital allocators on the planet were thinking about the market at a specific point in time. That’s not nothing. But it’s also not a roadmap.

Forward Watch List

• Broadcom (AVGO) earnings – June 3: The most immediate catalyst tied to this week’s filings. Q2 AI revenue guidance against the $10.7B target will either validate or stress-test Druckenmiller’s thesis. Multiple expansion or compression hangs on this number.
• Microsoft (MSFT) Copilot monetization: The next quarterly disclosure will be the first real read on whether enterprise M365 Copilot adoption at $30/user/month is scaling at a rate that justifies Ackman’s capex-adjusted valuation thesis.
• Alphabet (GOOGL): With TCI buying, Third Point buying, Druckenmiller selling, and Ackman selling — all in the same quarter — the next earnings cycle will be a stress test for which camp reads the AI competitive dynamic correctly.
• Amazon (AMZN) AWS margins: Tepper’s 98% position increase is a concentrated bet on Amazon’s AI infrastructure investments converting into margin expansion. AWS margin trajectory is the core variable to track.
• Airline sector: Tepper’s full exit signals a macro read on fuel costs and travel demand that the market hasn’t fully processed. Oil above $100 and margin compression across the carrier landscape may have more room to run.

The Q1 2026 13F cycle revealed something that doesn’t show up clearly in any single trade: the smart money is not unified. The most-watched funds in the world looked at the same macro environment — AI capex acceleration, elevated rates, oil above $100, a volatile tech sector — and reached different conclusions about which assets benefit most. That disagreement is interesting. It’s more interesting than any individual position.

The real question isn’t what Ackman bought or what Druckenmiller sold. It’s who gets the second-order effect right. AI infrastructure spending has to monetize somewhere. The funds filing this week each have a different answer for where that somewhere is.

– The Editorial Desk