CZ's viewpoint is not an isolated statement, but rather a consensus on "security layering" after the maturity of crypto infrastructure in 2026: Events from 2022-2025, such as FTX, Celsius, and 3AC, exposed the risks of exchange custody, but by 2026, leading platforms like Binance and Coinbase will rebuild trust through "proof of reserves + insurance funds + compliance audits", forming the statistical fact that "leading CEXs are safer than individual self-custody".

Similar cases can be compared to 2023 Ledger CTO Charles Guillemet's comments on Coldcard firmware vulnerabilities—hardware wallets are open-source but not verified by anyone, reflecting another side of "self-custody ≠ security": technical transparency does not equal operational safety, and user errors (loss of mnemonic phrases, phishing attacks) are more likely sources of loss.

In terms of capital pathways, CZ's "balanced approach" points to three types of asset allocation strategies: 1) Large long-term holdings → self-custody hardware wallets; 2) Medium trading funds → leading CEXs (Binance, Coinbase); 3) Small high-frequency → software wallets/on-chain DeFi, forming a "layered risk + purpose matching" hybrid architecture.

Essentially, this is a "trust reconstruction" before regulatory changes: when leading exchanges establish "bank-like" trust through compliance audits, insurance funds, and proof of reserves, retail investors shift from "distrusting any intermediaries" to "choosing regulated intermediaries", similar to the paradox after the 2008 financial crisis where "big banks became safer".

ABAB News · Law of Cognition

Statistical safety does not equal psychological safety
Exchange hacks make headlines, while self-custody losses go unnoticed
A sign of maturity: no longer asking which is safer, but which is more suitable