Three of the best-known addresses in American asset management reported the same stock on 14 August. All three held it on the 30 June 2026 reporting date, all three had changed the position substantially during the quarter, and all three filed within hours of each other.
The stock is Micron Technology. At Coatue Management it accounted for 7.46 percent of the reported book. At Appaloosa, 14.57 percent. At Situational Awareness, 27.54 percent. Coatue had increased the position by 1,793.7 percent, Situational Awareness by 27,712 percent, and Appaloosa had cut it by 41.4 percent.
Anyone trying to extract a signal from filings like these runs into a problem that better data cannot solve. Three experienced firms, one stock, one quarter, three directions. The usual question asked of a 13F is who is buying what. This quarter shows that is the wrong question. The right one is: at what dose.
What Coatue did in the second quarter
Philippe Laffont’s Coatue Management reports an equity book of $48.63 billion across 66 positions. That is more than twice what Situational Awareness disclosed, spread across almost three times as many names.
At the top sit Taiwan Semiconductor at $4.26 billion and Lam Research at $4.09 billion, both trimmed slightly. Then comes Micron at $3.63 billion. The position grew from 165,931 to 3,142,269 shares. It is the same build-out other firms undertook in the second quarter — it simply ends at a 7.46 percent weight rather than more than a quarter of the book.
Applied Materials follows at $3.05 billion, reduced by 19.7 percent; GE Vernova at $3.01 billion, essentially unchanged; and Amazon at $2.82 billion, increased by 49.2 percent. New to the portfolio are Intel at $1.69 billion and 12,084,027 shares, and AI chip designer Cerebras Systems at $1.55 billion. On the other side, ASML was cut by 40.5 percent and Netflix by 31.7 percent. Coatue exited only five positions entirely, among them Visa and the iShares bitcoin fund.
The resulting picture is not a firm that changed its mind. It is a rotation within the same thesis: away from the equipment makers of chip fabrication, toward memory, compute capacity and power. GE Vernova, Eaton, Equinix and Constellation Energy sit unchanged or slightly larger in the book. Those are turbines, switchgear, data centres and power stations.
The position that could not have existed three months earlier
Fourth in the portfolio, at $3.17 billion and 18,561,780 shares, is a name that could not have appeared in any earlier 13F: Space Exploration Technologies. SpaceX listed on Nasdaq on 12 June 2026 at an offer price of $135. Eighteen days later the quarter closed.
The valuation price can be reconstructed from Coatue’s filing: $3.1715 billion across 18,561,780 shares implies roughly $170.86 per share. Cathie Wood’s ARK Investment Management reports the same CUSIP with 4,478,013 shares and $767 million, implying about $171.3. Two independently submitted forms, effectively the same price, roughly 27 percent above the offer price. It is the first quarterly snapshot of the shareholder base after the listing.
Here the form shows its useful side. For a newly listed company with no history, the list of institutional holders at the first reporting date is genuine information. It says who entered at or near the offer price, and at what size. That Alphabet holds 7.2 percent of the Class A shares according to a separate late-July filing rounds out the picture from the capital side.
Why the same line means three different things
Back to Micron. All three firms read the second quarter correctly: memory was tight, prices rose, the stocks ran. Micron had gained roughly 304 percent in the first half, SanDisk 858 percent. Buying in April looked right by June.
July broke it. Korean competitor SK Hynix reported a record quarter and simultaneously guided 2026 capital expenditure fifty percent higher, to at least $31 billion. The market read that as the end of scarcity. Micron fell about twenty percent in July, SanDisk more than thirty-five, and Korea’s benchmark index dropped 29 percent in a month.
For the three firms, that identical price move meant something entirely different.
For Coatue, at a 7.46 percent weight in a book of 66 positions, it was a bad month in one holding. A twenty percent decline costs roughly 1.5 percent of the reported portfolio value. Unpleasant, nothing more.
For Appaloosa, which had already cut the position by 41.4 percent to 975,000 shares, it was a partial retreat that paid off — but not completely: Micron remained the second-largest holding at 14.57 percent. David Tepper took profits in the second quarter and stayed invested anyway. That is the middle ground between the two extremes.
For Situational Awareness, at 27.54 percent in Micron, a further 28.03 percent in SanDisk and reported gross leverage of around 400 percent, the same July was existential. The fund lost roughly 67 percent, its prime brokers called for margin, and in late July Citadel absorbed positions worth about sixteen billion dollars at a discount of around ten percent.
The same stock three times, the same market move three times: once an irritation, once a missed opportunity, once the end of the fund. The difference was not in the analysis. It was in position size, and in how much borrowed money sat on top of it.
The rest of the filing season
Micron was not the only case where prominent firms took opposite sides in the same quarter. The most-bought name of this round was Alphabet, and there too the line runs straight through the industry.
On the buying side: Berkshire Hathaway, which increased its Class A holding by 45.2 percent and its Class C by 658 percent; Dan Loeb’s Third Point, up 485.7 percent; Seth Klarman’s Baupost, up 16.1 percent; Chris Hohn’s TCI, up 12.2 percent; Coatue, up 12.6 percent; and Appaloosa, up 6.8 percent. Li Lu’s Himalaya Capital holds Alphabet across both share classes at almost 48 percent of its entire reported book and did not touch it.
On the selling side: Chase Coleman’s Tiger Global, down 45.4 percent; Ray Dalio’s Bridgewater, down 33.8 percent; and Bill Ackman’s Pershing Square, which cleared the position entirely. Ackman has said publicly that he still considers Alphabet attractive but that his capital is limited and better deployed elsewhere. That too is a reminder that a sale looks like a change of mind in the form when it is often only a capital decision.
It is equally instructive how differently the three best-known students of Julian Robertson behaved. Tiger Global cut across its top holdings — all ten largest names are down. Coatue added and built three sizeable new positions. Andreas Halvorsen’s Viking Global rotated: Meta up 75.8 percent, Sherwin-Williams up 24.5 percent, Air Products up 15.5 percent, while Taiwan Semiconductor was cut by 29.1 percent and Charles Schwab by 27.4 percent. Same school, same quarter, three different conclusions.
At Pershing Square, the striking detail is that three new positions of the same character appear at once: Visa at $1.12 billion, Mastercard at $1.09 billion and S&P Global at $1.06 billion — together nearly seventeen percent of a fourteen-position book. Payment networks and a ratings agency, all three high-margin businesses with defensible positions. When a concentrated fund adds three names from the same category simultaneously, it is one of the few configurations in which a 13F genuinely shows a recognisable intent.
The metric that is in the filing and usually overlooked
That is the practical yield of this filing season. A 13F answers “what does someone own” well, “what does someone think” poorly, and “what does someone risk” not at all. The one quantity that can be reliably derived from the document and that genuinely says something about conviction is a position’s weight in the overall book.
That number appears in no column. It has to be computed from position value and total value, which is exactly why it is missing from many automated screens. Those report that three prominent firms hold Micron. That it was every fourth dollar at one of them and every thirteenth at another disappears in the same list.
A second figure of the same kind is overall concentration. Coatue spreads $48.63 billion across 66 positions, an average of $737 million per name. Chris Hohn’s TCI holds $52.8 billion in eleven positions, 33.59 percent of it in GE Aerospace alone. Joel Greenblatt’s Gotham spreads $43 billion across 1,791 positions. All three are successful managers with comparable book sizes. Their filings are completely different documents, and a single line from each means something different.
What this means for reading such filings
Three rules follow from this quarter, and none of them requires specialist software.
First: always read positions as weights, never as absolute amounts. A billion dollars is two percent of the book at Coatue and half the business at a smaller firm. The absolute figure measures the size of the manager, not the strength of the conviction.
Second: read the position count alongside it. A portfolio of eleven names is a set of decisions each thought through individually. A portfolio of 1,791 names is a systematic model in which no single line expresses an opinion. Both can work, but only the first yields traceable individual ideas.
Third: percentage changes without a base are worthless. A gain of 27,712 percent sounds like maximum conviction but merely describes building up from a residual 17,362 shares. Conversely, a ten percent cut in a very large position can move more capital than a tripling of a small one. Only the change in weight carries meaning.
Read that way, a 13F delivers exactly what it can: a reliable snapshot of how strongly someone was committed to something on one specific date. No more, but no less either. This quarter that single metric was enough to separate three firms holding the same Micron position — and it explains why the same six weeks of July were an irritation for one of them and the end for another.
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.
