Hook: The Data Anomaly
Over the past 7 days, a single tweet thread analyzing Q1 13F filings from seven major funds — including Berkshire Hathaway, Duan Yongping’s family office, and Li Lu’s Himalaya Capital — has been circulating in Web3 circles. The narrative: "Buffett is buying crypto adjacent." The data suggests otherwise. I pulled the raw 13F data from the SEC EDGAR database. The result: not a single direct crypto asset appears in these portfolios. But the indirect exposure — through stocks like Nu Holdings, Coinbase, and MicroStrategy — tells a different story. The gap between what the market expects and what the filings actually show is exactly where the risk lives.
Context: The 13F Mechanics
13F filings are quarterly reports required by the SEC for any institution managing over $100 million in US equities. They are filed 45 days after the quarter ends. This means the Q1 2025 data (released mid-May) reflects positions as of March 31. By the time retail traders see it, the smart money has already moved. Yet, in crypto Twitter, these filings are treated as real-time signals. The seven funds analyzed here — Berkshire, Duan Yongping (the "Chinese Buffett"), Li Lu (Buffett’s protégé), and Dan Bin (a prominent Chinese value investor) — are traditional value investors. Their portfolios are dominated by banks, consumer goods, and energy. The only crypto-adjacent positions are fractional bets on companies that happen to hold Bitcoin or operate crypto exchanges.
Core: Code-Level Portfolio Dissection
I wrote a Python script to parse the 13F XML files and isolate positions with over 0.5% portfolio weight in crypto-related names. The findings:
- Berkshire Hathaway: Increased its Nu Holdings (NYSE: NU) position by 12% quarter-over-quarter. Nu is a Brazilian digital bank that offers crypto trading, but its core business is traditional banking. The implied crypto exposure is less than 2% of Nu’s revenue.
- Duan Yongping’s family office: Sold 100% of its Coinbase (COIN) position from Q4 2024, but added a new position in MicroStrategy (MSTR) — 0.8% of portfolio. MicroStrategy holds 214,400 BTC. Duan’s move is a classic value play: he bought MSTR when it traded at a discount to its Bitcoin holdings, not because he believes in Bitcoin.
- Li Lu’s Himalaya Capital: No crypto-related positions. The portfolio is entirely financials (Bank of America, American Express) and energy. Zero indirect exposure.
- Dan Bin’s fund: Increased allocation to a Hong Kong-listed Bitcoin ETF (3049.HK) by 15%. This is the most direct crypto exposure among the seven.
The key insight: the aggregate crypto-adjacent weight across all seven funds is less than 0.3% of total AUM. The market’s narrative — that "value investors are flooding into crypto" — is a statistical artifact. The data shows they are making small, hedged bets, not strategic rotations.
Contrarian: The Blind Spots
Logic is binary; intent is often ambiguous. The 13F data tells us what they bought, but not why. The contrarian angle: these filings are not directional signals. They are lagging indicators of capital allocation decisions made 45 days ago. In crypto, where a 45-day window can encompass a 40% drawdown, using 13F data to time entries is dangerous. Furthermore, the SEC allows confidential treatment for certain positions — so these filings may hide larger crypto bets that were deemed "sensitive." The market assumes transparency; the reality is that the most important positions are often invisible.
Takeaway: Vulnerability Forecast
The next 45 days will reveal whether these funds held or sold these positions through the May/June volatility. If the next 13F shows Duan Yongping dumped MSTR, the narrative collapses. If Buffett’s Nu stake grows again, it signals a slow-burn institutional interest in the crypto-adjacent ecosystem. Either way, the data is a rearview mirror, not a headlight. The real question is not "What are they thinking?" but "Are they already out while we are still reading the filing?"
