Bitcoin Custody Data Cited by CZ Shows 1.57M BTC Lost

Bitcoin News

The custody debate around Bitcoin (BTC) sharpened on Aug. 4 after Binance founder Changpeng Zhao argued that centralized exchanges may be statistically safer than holding coins directly, using a River 2025 industry data set circulated by analyst Willy Woo. The figures estimate 1.57 million BTC lost through self-custody against 1.51 million BTC lost on exchanges, a narrow gap of less than 60,000 BTC. Zhao’s core point was not that exchange failures are minor, but that the public record captures them more completely. A large platform breach usually becomes a visible, documented event, while misplaced seeds, forgotten passwords, broken hardware devices and inheritance failures often disappear without a public trace. That reporting asymmetry can make the self-custody loss tally look smaller than the real operational burden borne by individual holders. The discussion matters because Bitcoin’s security model assumes users can protect keys, yet the practical skill required remains high. For newer participants, the choice between running a personal node, using a hardware device or relying on a regulated venue is less ideological than operational. Zhao also noted that some recorded exchange losses are tied to platforms that have since exited the market, which complicates direct comparisons between active custodians and defunct businesses. His framing turned the custody conversation away from a simple moral ranking and toward probability, process and recovery. In a market where a single lost key can permanently remove coins from circulation, the difference between a documented hack and an undisclosed personal error is important for any realistic risk model. The data also underscores that Bitcoin’s supply is not merely capped by protocol rules; it is reduced in practice by irreversible loss events, whether they occur in private wallets or on centralized infrastructure. That distinction is especially relevant during periods of heightened volatility, when users may move assets quickly and operational mistakes become more likely.

The second layer of the discussion centers on how exchange risk is managed and disclosed. Zhao pointed to Binance’s compensation record and its Secure Asset Fund, which the exchange recently expanded into a $1 billion Bitcoin reserve intended as a buffer against user losses tied to platform-side breaches. His argument is that a centralized venue can internalize certain failures, while a private wallet error usually has no recourse mechanism. The broader incident backdrop is mixed. Industry hack tracking showed incident counts increased about 50% during the first six months of 2026, even though total stolen value declined, suggesting more frequent but smaller attacks. At the same time, some historical exchange losses belong to businesses no longer operating; BitMEX, for example, announced in July that it was closing after an 11-year run. That means aggregate custody statistics can blend active risk with legacy failures. Self-custody risk also moved from theory to practice in early August, when a Coldcard hardware-wallet vulnerability was linked to BTC being drained from users. In one reported case, a victim lost $1.6 million within minutes despite following standard security procedures. Zhao used that episode to caution that no wallet configuration can promise absolute protection. He stopped short of telling users to abandon private wallets, instead describing the decision as a question of risk tolerance and product suitability. For Bitcoin holders, the practical takeaway is that custody is not a binary choice between trust and trustlessness. It is a spectrum of key management, backup discipline, operational security and, where relevant, the legal and financial protections offered by a service provider. This matters for institutional and retail allocators because custody assumptions shape portfolio construction, insurance decisions and the operational threshold at which an investor prefers convenience over direct control. Those choices become more visible when Bitcoin experiences sharp price swings and users reassess where assets should sit during stress.

COINOTAG’s proprietary 42-indicator composite S/R scoring engine rates Bitcoin’s nearest resistance at $63,953 with a 75/100 score, driven by Flip S→R and SMA 20 confluence, while stronger supply sits at $66,879 with 77/100 from EMA 100 and Donchian Upper. Spot traded near $63,598, up 1.83%, with funding at 0.0049%, open interest at $12.73 billion and a 1.52 long/short ratio, indicating modest long bias. The 25/100 Extreme Fear print, far from the optimism of an all-time high phase, and BTC’s 69.6% share of COINOTAG’s tracked market echo bear market caution, leaving limited room for altcoin leadership. A break above $63,953 could open $65,611, while losing $62,948, a 66/100 support backed by POC and Donchian Lower, would invalidate the bullish case.