Risk analysis on the largest individual ETH book

Disclosure: I'm part of the team behind Sentralis (portfolio risk and scenario analysis), which produced this. Not advice; all figures are model estimates under stated assumptions.

We ran a full scenario suite on Vitalik Buterin's public three-address book — the Safe holding 224k + 7k staked ETH plus the ENS hot wallet, not the $0 poisoned lookalike that circulates, and not the Ethereum Foundation (separate entity). Book: $443.9M at July-15 closes, 99.98% ETH.

A one-asset book defeats most portfolio analytics by construction (crisis-correlation stress moved 1-day VaR by exactly $450, reported as the non-result it is). What survives:

• The model-risk gap: same book, same seed — zero-drift GBM median year ≈ 0%, trailing-365d bootstrap median −46% (1-in-20: −65% vs −81%). A single VaR number for a book like this quietly picks a side.
• The scenario ordering: the FTX replay is the mildest thing we ran (−30%, vs −54% for the Oct-25→Feb-26 bear leg replayed from today). Self-custody has no counterparty leg; macro is what bites.
• Exit arithmetic: the position is ~0.9 days of the entire measured ETH tape; disciplined full exit ≈ 21 days normal / 70 crisis. No obligation forces any exit.
• The January 16,384-ETH open-source funding commitment buys $31.5M today, $5.9–16.9M across the pessimist model's year.

Full analysis (address verification, charts, method, limitations): https://sentralis.io/teardowns/vitalik/

First entry in a regular series of risk teardowns. Public figures' and organizations' portfolios, free, publicly disclosed positions only. World Liberty Financial is next; suggestions for after that welcome in the comments. Corrections to the data basis too.