The quarter's filings point in different directions
Jump Financial reported 63,210 NVIDIA shares for 2026-Q2, compared with 4,047 in the preceding reporting period. SecBot's institutional change record classifies that as a doubled position and reports a 1,461.9% increase in shares. In the same quarter's data, Wolverine Asset Management reported just 5,374 shares, down from 113,866, a 95.3% reduction. Those two filings put the variety of reported portfolio changes into sharp relief: an increase at one manager sits alongside a near-exit at another.
This is a story about reported holdings, not a live account of trading. The records are quarter-end 13F disclosures, and the quarter label is 2026-Q2. The changes compare shares reported for the current quarter against shares reported for the prior period. They do not identify the individual transactions or their dates. The public records therefore establish reported position differences, not a precise account of when or why a manager changed exposure.
The SecBot data shows other substantial additions and reductions in NVIDIA positions, too. Van Eck Associates Corp reported 75,849,721 shares, compared with 56,327,417, and its reported share change was 34.7%. Allianz Asset Management GmbH reported 30,069,187 shares against 23,939,334, with a 25.6% increase. On the reduction side, Whale Rock Capital Management reported 377,204 shares, down from 1,041,128, and a 63.8% decrease. This mixture is more informative than treating the filings as a single institutional vote on the stock.
Selected position changes
| Manager | Previous shares | 2026-Q2 shares | Reported share change |
|---|---|---|---|
| Jump Financial | 4,047 | 63,210 | +1,461.9% |
| Van Eck Associates | 56,327,417 | 75,849,721 | +34.7% |
| Allianz Asset Management | 23,939,334 | 30,069,187 | +25.6% |
| Wolverine Asset Management | 113,866 | 5,374 | -95.3% |
| Whale Rock Capital Management | 1,041,128 | 377,204 | -63.8% |
| Brevan Howard Capital Management | 2,960,428 | 657,548 | -77.8% |
The standout percentage increase needs scale context. Jump Financial's change is a very large percentage because the preceding reported position was 4,047 shares. Its current reported position, 63,210 shares, is far smaller in absolute share count than Van Eck's 75,849,721 or Allianz's 30,069,187. Percentage change captures the movement relative to the prior holding; it does not rank managers by total shares held.
Van Eck's entry in the local record was dated 2026-08-22. Allianz's was dated 2026-09-02. Jump Financial and Wolverine Asset Management entries were dated 2026-08-22. These are record creation timestamps, not transaction dates. HRT Financial LP appears separately as a new position with 115,607 shares, dated 2026-10-24. The filing record gives its reported value as $23,131,000. It does not provide a prior share count for that row, consistent with the new-position classification.
Additions were not uniform
Other additions show that the percentage changes describe different starting points and position sizes. ABN AMRO Investment Solutions increased its reported count from 2,784,088 to 4,486,854 shares, with the row reporting a 61.2% increase. Centiva Capital reported 1,958,867 shares, up from 1,273,002, and a 53.9% increase. Vinva Investment Management reported 5,236,319 shares, compared with 3,450,141, for a reported 51.8% increase.
A separate record for ADAPT Investment Managers SA shows 295,573 shares, up from 68,865, and a 329.2% change. The same data labels this doubled. Those categories are SecBot's derived descriptions of the quarter-to-quarter reported share counts. They do not mean every share was purchased in the market during the quarter. Reporting can reflect other position changes and does not explain the manager's rationale.
Reductions also vary in magnitude. Brevan Howard Capital Management LP reported 657,548 shares against 2,960,428, with a 77.8% decline. Vident Advisory reported 2,167,442 shares, down from 3,349,355, and a 35.3% reduction. These moves differ from Wolverine's reported 95.3% decrease, which left 5,374 shares in the record. The data calls Wolverine's change a near-exit, not an exit, so the distinction matters: the reported position was reduced substantially but was not zero.
What the records can and cannot establish
A Form 13F is a delayed disclosure of qualifying securities holdings. In this dataset, the positions are reported by tracked managers for a reporting quarter. A quarter-end share count is a snapshot, not an intraday quote or a full portfolio history. It does not reveal the sequence of transactions within the period. Nor does the count alone establish whether a manager viewed a position as a hedge, a trading position or a long-term allocation.
The listed reported values are nominal dollar amounts in SecBot's stored records. They are not the manager's cost basis, profit or assets under management. Price changes can alter reported dollar values even when share counts move differently, so share-count comparisons are the clearer measure for this story. The filing data also does not say whether a manager's entire exposure is captured by the long-share rows. The query used here returns mapped non-option long holdings, and should not be read as a complete ledger of every financial exposure connected with NVIDIA.
The snapshot should not be generalized to all institutions. These rows cover locally tracked managers and the quarter specified. Several updates have different record dates, including entries created on 2026-08-22 and 2026-10-24. The reports are not all equally recent in the local system, and their creation dates do not transform quarter-end disclosure into a real-time position feed.
Where to inspect the data
The company page for NVIDIA provides the stock-level view, while the institutional holdings page covers manager disclosures. These records are useful for checking both the share count and the reported change classification. The most striking percentage move in this set is Jump Financial's 1,461.9% increase, but its 63,210-share position also shows why a percentage alone is not a measure of absolute exposure.
The next relevant point is a later reporting-quarter filing, when another set of quarter-end share counts can be compared with these 2026-Q2 records. SecBot's institutional data remains a delayed disclosure snapshot rather than a live holdings feed. For more SecBot research and data views, register here.