Reviewed: November 22, 2026 · Data window: March 31 to September 30, 2026 · Sample: AHNIM STUDIO’s fixed 17 institutional 13F reporting groups, 1,179 clean manager–security observations present in all three quarters
Did widely shared securities stay split between manager share-count increases and decreases in Q3 2026? Yes. Of 15 securities continuously reported by at least four managers in our fixed panel, 12 were split in both Q1→Q2 and Q2→Q3. The camps were not stable: across 84 manager–security observations in those 12 repeated-split securities, 40 observations flipped direction, 38 continued the same non-zero direction and six involved an unchanged quarter.
TSMC was the clearest example. Its manager count moved from three increases and seven decreases in Q1→Q2 to seven increases and three decreases in Q2→Q3, with eight of 10 continuous managers reversing direction. Yet their combined disclosed TSMC shares still fell 6.2% in Q3 because the reductions were larger. A manager majority therefore answers a breadth question; it is not a capital-weighted consensus signal.
Three findings
- Direction splits persisted across quarters. Twelve of 15 securities with at least four continuous managers had both an increase and a decrease in each interval.
- The individual camps changed almost as often as they persisted. Among 84 observations in the repeated-split set, 47.6% flipped between increase and decrease, 45.2% kept the same non-zero direction and 7.1% involved an unchanged quarter.
- Manager count and aggregate magnitude disagreed three times. TSMC and Spotify had more increasing than decreasing managers but fewer combined shares; Uber had more decreasing managers but slightly more combined shares.
Twelve widely shared securities stayed split
| Security | Continuous managers | Q1→Q2 inc / dec | Q2→Q3 inc / dec | Managers that flipped | Q2→Q3 aggregate shares |
|---|---|---|---|---|---|
| TSMC ADR (TSM) | 10 | 3 / 7 | 7 / 3 | 8 | −6.2% |
| Meta Class A (META) | 10 | 3 / 7 | 6 / 4 | 7 | +16.3% |
| Amazon (AMZN) | 10 | 5 / 5 | 3 / 7 | 4 | −5.7% |
| NVIDIA (NVDA) | 9 | 3 / 6 | 7 / 2 | 4 | +8.8% |
| Alphabet Class A (GOOGL) | 8 | 4 / 4 | 1 / 6 | 2 | −7.0% |
| Microsoft (MSFT) | 8 | 3 / 5 | 1 / 6 | 2 | −0.5% |
| Broadcom (AVGO) | 7 | 3 / 4 | 3 / 3 | 2 | +12.5% |
| AMD (AMD) | 5 | 2 / 3 | 1 / 4 | 3 | −26.7% |
| DoorDash Class A (DASH) | 5 | 2 / 2 | 2 / 1 | 2 | +2.9% |
| Uber (UBER) | 4 | 2 / 1 | 1 / 2 | 3 | +2.0% |
| Spotify (SPOT) | 4 | 1 / 3 | 3 / 1 | 2 | −44.2% |
| ASML ADR (ASML) | 4 | 2 / 2 | 1 / 3 | 1 | −39.3% |
The table is deliberately security-specific. Alphabet Class A and Class C remain separate because they have different CUSIPs, even though they share an issuer. That avoids treating unlike filing lines as one share-count unit. It also means the result is narrower than an issuer-level portfolio view.
TSMC and Meta changed camps most often
Eight of the 10 continuous TSMC managers reversed direction. ARK, Bridgewater, Coatue, Light Street, Soros and Whale Rock moved from decreases to increases; Appaloosa and Third Point moved from increases to decreases. Altimeter increased in both intervals, while Tiger Global decreased in both.
The count flipped from a 3–7 decrease majority to a 7–3 increase majority, but combined disclosed shares moved the other way. The seven increases added about 694,000 shares in aggregate. The three decreases removed about 1.88 million, including roughly 1.75 million from Tiger Global. The net result was a 1.19 million-share decline, or 6.2%, among the continuous TSMC group.
Meta also showed unstable camps: seven of 10 managers reversed direction. Unlike TSMC, the direction count and magnitude agreed in Q3. Six managers increased, four decreased, and combined shares rose 16.3%. Coatue, Whale Rock, Bridgewater and Altimeter supplied much of the positive share-count change, while Fundsmith, Tiger Global, Appaloosa and Pershing Square reported fewer shares.
NVIDIA had a similar majority reversal but less manager-level turnover. The count shifted from three increases and six decreases to seven increases and two decreases. Four of nine managers flipped direction, while five kept their prior direction. Combined shares rose 8.8%, led by increases at Whale Rock, ARK and Tiger Global. The difference between TSMC’s eight flips and NVIDIA’s four shows why an aggregate majority change does not reveal whether the same managers changed course.
Three count majorities disagreed with aggregate shares
| Security | Q2→Q3 inc / dec / unchanged | Count direction | Aggregate share change | What the mismatch means |
|---|---|---|---|---|
| TSMC ADR | 7 / 3 / 0 | Increase majority | −6.2% | Three reductions outweighed seven increases. |
| Uber | 1 / 2 / 1 | Decrease majority | +2.0% | One increase outweighed two reductions. |
| Spotify | 3 / 1 / 0 | Increase majority | −44.2% | Tiger Global’s reduction dominated three increases. |
Spotify is the sharpest magnitude warning. Three of four continuous managers reported more shares, yet their combined shares fell 44.2%. The single decrease was Tiger Global, whose disclosed Spotify shares fell from roughly 1.17 million to 2,561. The other three increases did not offset that change. This is exactly why “three of four managers increased” should not be rewritten as “institutional exposure rose.”
What repeated disagreement adds to the Q3 picture
Our Q3 holder-breadth analysis counted additions and departures, including SK Hynix and Bending Spoons moving from zero to four managers. This article asks a different question: among securities that stayed in several managers’ disclosures for all three quarters, did managers move reported share counts in the same direction?
The answer is mostly no. Repeated splits appeared in 80% of the widely shared continuous-security set. Moreover, the 40 direction flips nearly matched the 38 same-direction observations. The division was therefore not just a stable set of managers accumulating while another stable set reduced. In TSMC and Meta, especially, the membership of the increase and decrease camps changed substantially from one interval to the next.
That complements the three-quarter expansion-persistence test, which measured whether prior increases continued, and the Q3 quantity-versus-valuation bridge, which separated share-count and unit-value effects. The current statistic is cross-manager: it tests whether several reporting groups agreed on direction for the same security and whether that agreement persisted.
Method, scope and limitations
- Periods and panel: March 31, June 30 and September 30, 2026 quarter-end holdings for AHNIM STUDIO’s fixed 17 institutional reporting groups. The politicians in our separate PTR panel are not included.
- Starting universe: 1,179 clean manager–security observations with the same CUSIP present in all three quarters after the standing instrument, identifier, reporting-group and corporate-action controls.
- Wide-sharing threshold: at least four continuous managers. Fifteen securities qualified; 12 had at least one increase and one decrease in both intervals. The threshold is a transparent sample definition, not a claim that four managers constitute market consensus.
- Direction transition: a flip means increase→decrease or decrease→increase for the same manager and CUSIP. An unchanged quarter is reported separately and is not forced into either direction.
- Share-class treatment: different CUSIPs remain different securities. Verified split adjustments in the inherited clean comparison are preserved; no unresolved corporate-action candidate enters this set.
- Interpretation: Form 13F reports quarter-end long positions and fair value. It does not provide transaction dates, prices, motives, cost basis, short exposure, total assets or complete portfolios. A missing line is not converted into a known sale date. Direction counts are not performance, conviction or future-return rankings.
The Guru research hub keeps the fixed-panel definitions, individual manager reports and related statistical articles together.
Sources and original documents
SEC, Frequently Asked Questions About Form 13F (updated March 6, 2026); SEC Form 13F and instructions; SEC Form 13F data sets.
Direct Q3 2026 filing indexes:
Altimeter; Appaloosa; ARK; Berkshire Hathaway; Bridgewater; Coatue; Dalal Street; Greenlight / DME; Fundsmith; Icahn reporting group; Light Street; Pershing Square 13F-NT notice and combination 13F-HR; Situational Awareness; Soros Fund Management; Third Point; Tiger Global; Whale Rock.
The normalized comparison preserves every Q1, Q2 and Q3 direct information-table URL at manager–security level. The public Q1/Q2 filing index list is also preserved in the sources section of our Q2 manager quantity-effect analysis.
AHNIM STUDIO calculation from the cited filings. This article is for information and education, not investment advice. Data and assumptions can be revised if a filing amendment or verified identifier or corporate-action change alters the normalized comparison.

Leave a Reply