No Q4 Filing: Burry Ended Scion After Concentrating 96.7% Into PLTR+NVDA Puts

Michael Burry''s Scion Asset Management filed its last known 13F on November 3, 2025. Seven days later, Scion deregistered as an investment adviser. As of February 19, 2026, after the Q4 filing deadline window, there is still no Scion Q4 2025 13F in our production database or SEC-indexed records used for this research. The practical conclusion is clear: Q3 2025 was the public end of Scion''s filing era.

TL;DR Summary

• No Q4 filing: No Scion 13F for period ending 2025-12-31 is visible by 2026-02-19.
• Extreme concentration: Scion''s top two positions (PLTR puts + NVDA puts) totaled $1.0987B, or 96.72% of reported Q3 value.
• Portfolio shock: Reported portfolio value jumped from $578M (Q2) to $1.136B (Q3), a +96.41% QoQ increase while holdings count fell from 15 to 4 in our DB.
• Timing: Scion filed on 2025-11-03 near the local peak zone in AI momentum names, then deregistered on 2025-11-10.
• Interpretation risk: 13F option values are not premium paid; they reflect reportable market-value exposure of underlying shares. Media often overstates “bet size” if this is ignored.

If you followed our earlier institutional deep-dives on Berkshire, Bridgewater, Renaissance, Pershing, and Viking, Scion''s final print sits at the far end of the conviction spectrum: less diversification, more directional optionality, and then institutional silence.

Final Filing Anatomy: 96.7% in Two AI-Sensitive Put Buckets

Scion''s Q3 2025 filing (period end 2025-09-30) reports four rows in our holdings table, but two rows dominate almost all disclosed exposure:

• Palantir (PLTR) puts: $912.1M reported value (80.30% in-row share of portfolio)
• NVIDIA (NVDA) puts: $186.58M reported value (16.43%)
• Molina Healthcare (MOH): $23.92M (2.11%)
• SLM (SLM): $13.29M (1.17%)

Top-two concentration is not just high. It is effectively binary: $1.0987B out of $1.1359B. For readers tracking manager behavior, that usually means one thing: the manager no longer optimizes for smooth quarterly optics, only for thesis payoff.

Q2 to Q3: From Diversified Book to One Macro Expression

In Q2 2025, Scion reported 15 lines across healthcare, consumer, China ADRs, and mega-cap tech—including UNH, REGN, LULU, META, BABA, JD, and ASML. In Q3, that diversified surface disappears in our DB snapshot, replaced by concentrated downside convexity plus two small longs.

My read as a researcher: this is less a “rebalance” and more a structural statement that Burry saw regime risk as dominant over stock-picking dispersion. The portfolio construction itself becomes the argument.

Deadline Window Context: Peers Filed, Scion Stayed Dark

Around the 2026-02-14 deadline window, large and high-attention managers were active in the tape and in filings. In our database snapshot, Renaissance filed on 2026-02-12 and Soros Fund Management filed on 2026-02-13; Two Sigma, Viking, and Elliott appeared shortly after.

Scion did not. That contrast matters for narrative integrity: this is not a data-lag anecdote from a quiet corner of the manager universe. It happened during the most visible filing window of the quarter.

PLTR/NVDA Price Path and the Option-Reporting Trap

Late-2025 price action amplified attention. PLTR printed around the low-$200s in early November before a sharp post-peak drawdown. NVDA similarly rolled over from late-October highs into a deep November pullback before stabilizing. This sequence made Burry''s disclosed put exposure look prescient in headlines.

But this is where many public takes lose precision: 13F option lines describe reportable exposure mechanics, not premium paid. A $900M+ “put position” in 13F framing does not mean $900M in cash outlay. Analysts who ignore this overstate leverage and misread risk budget.

What I Think This Final Filing Actually Signals

Three conclusions stand out.

1. It was a closure trade, not a franchise trade. The concentration profile is inconsistent with preserving external-AUM relationship volatility. It is consistent with a manager positioning for a final thesis expression before strategic exit.
2. The informational edge was in scenario asymmetry, not precision timing. Even if entry/exit windows were imperfect, put convexity can still monetize volatility bursts.
3. Scion''s silence after Q3 is itself data. In a cycle where peers kept printing forms, no Q4 Scion filing by 2026-02-19 reinforces that deregistration was operationally real, not rhetorical.

For readers building cross-manager views, compare this endpoint against the broader Q4 cluster in our whale funds overview and our consensus report. The contrast is the point: most large managers optimized continuity; Burry optimized a final, concentrated message.

Whether that message proves ultimately right is still an open market question. But as portfolio communication, this was unmistakable: concentrate, file, deregister, disappear.