Over the past twelve months, the number of Bitcoin addresses with non-zero balances that have never received funds from a centralized exchange has increased by 28%, according to data from River's latest self-custody report. That's a jump from approximately 12 million to over 15 million addresses. The figures are stark: while total Bitcoin holders have grown modestly, the proportion of self-custodied coins has risen to an estimated 42% of the circulating supply, up from 34% in 2023. This is not just a statistic—it is a signal about how the market's trust architecture is being renegotiated.
Why now? The context is impossible to ignore. The collapse of FTX in November 2022 was a watershed moment for crypto's relationship with custodians. Since then, every major exchange breach, regulatory clampdown, or audit failure has reinforced a simple lesson: not your keys, not your coins. River's report, released this week, captures the culmination of that sentiment. It aggregates on-chain data from over 60,000 nodes and cross-references it with exchange withdrawal patterns to estimate the share of Bitcoin held in self-custody versus third-party custody. The report's methodology is rigorous—it filters out dust addresses, pooled mining payouts, and known service wallets to isolate genuine individual holdings.
The core insight is that the shift is accelerating. In the first quarter of 2025 alone, over 200,000 BTC moved from exchange wallets to addresses that have never interacted with a known custodial service. That's roughly 1% of the total supply in just three months. The immediate impact is twofold: it reduces the available liquidity on exchanges, which can exacerbate price volatility during sell-offs, and it decentralizes the security of the network, distributing key ownership across a broader base of individuals. As an exchange market lead, I've watched this trend from the front line. During the 2022 bear market, I ran 'Transparency Tuesdays' to show our cold wallet reserves. I saw firsthand how fear of insolvency drove users to withdraw. The River data confirms that this behavior is not temporary—it is a structural shift.
But the numbers also reveal a more nuanced story. The growth in self-custody is concentrated among holders with balances between 0.1 and 1 BTC. This 'middle class' of Bitcoiners—neither whales nor dust holders—is driving the movement. They are likely retail investors who have been burned by exchange failures and are now taking responsibility for their own keys. The report shows that addresses in this range have increased by 22% year-over-year, while those with over 100 BTC have remained flat. This suggests that the self-custody trend is not about the ultra-wealthy moving coins off exchanges; it is about the democratization of security.
The ethical pulse of the decentralized economy beats strongest here. When ordinary people choose to hold their own keys, they are not just protecting their wealth—they are asserting a principle: that trust should be minimized, not centralized. This aligns with the core ethos of Bitcoin, but it also introduces new risks. The same report notes that the number of addresses with balance errors due to lost keys has increased by 15% over the same period. The human cost of self-custody is real. In my early days as a community liaison for Icon Foundation, I fielded hundreds of calls from users who had lost access to their wallets. The pain of a lost key is not just financial; it is emotional. As a compassionate narrative guide, I believe we must balance the celebration of self-custody with a sober acknowledgment of its challenges.
Now, let's dig into the data. River's report provides a breakdown by geography. The highest growth in self-custody is in Asia-Pacific, particularly in South Korea and Japan, where regulatory uncertainty around exchange licenses has prompted users to move coins offline. In contrast, North America has seen a slower but steady increase, likely because institutional products like spot ETFs offer a regulated alternative that feels safer. The report also highlights a correlation with Bitcoin price movements: self-custody addresses tend to accumulate during bear markets and distribute during bull runs. This is intuitive—people buy the dip and hold, then sell when prices rise. But the current cycle is different. Even as Bitcoin has rallied 60% from its 2024 lows, the flow of coins into self-custody has not reversed. That is a divergence from historical patterns.
Building bridges in a fragmented digital frontier requires us to ask why. One explanation is that the ETF narrative has changed the calculus. Institutional investors can now gain exposure to Bitcoin without taking custody, so retail investors who want to own the asset directly are doubling down on self-custody. Another is that the rise of Layer 2 solutions like Lightning Network has made self-custody more practical. With Lightning, you can hold your own keys while still participating in payments and DeFi-like services. The River report does not explore this directly, but based on my experience auditing custodial setups, I've seen a clear uptick in Lightning-enabled wallets that support non-custodial channels. The technology is catching up to the ideology.
But here is the contrarian angle that most coverage will miss. The River report's data might be overstating the self-custody trend. Its methodology relies on heuristic clustering—identifying addresses that have never received funds from a known exchange. But what about users who use decentralized exchanges (DEXs) or peer-to-peer platforms? Those transactions do not appear as exchange inflows, so the report might classify coins held by a DEX user as self-custodied when they are actually sitting on a trading platform's smart contract. Moreover, the report does not distinguish between individual self-custody and custodial services like BitGo or Coinbase Custody that are not classified as 'exchanges' in the data set. These services hold billions in Bitcoin for institutions, and they are not captured in the 'exchange' bucket. So the 42% figure might include a significant portion of institutional custody, muddying the narrative of retail empowerment.
Another blind spot: the report's definition of 'self-custody' excludes multi-signature setups where keys are shared among multiple parties. Many advanced users employ multisig for security, but those addresses may be flagged as 'shared ownership' and excluded. The report acknowledges this limitation in a footnote, but it is easy to miss. As someone who has spent years in the trenches of DeFi governance, I've learned that the devil is in the footnotes. The true percentage of Bitcoin held by individuals who control their own private keys without any third-party dependency is likely lower than 42%—perhaps closer to 30%. That is still a significant shift, but it is not the revolution it appears to be.
Furthermore, the report's data on 'lost coins' is alarming but incomplete. It estimates that 2.5% of all Bitcoin is in addresses that have been inactive for over five years. Some of these are lost, but others are held by long-term hodlers who do not move their coins. The report cannot distinguish between a lost key and a patient investor. This is a methodological challenge that all on-chain analysts face. In my own work covering the NFT ethics investigation, I learned that on-chain data is never neutral—it is a map, not the territory. We must always ask who is not represented in the data.
The ethical pulse of the decentralized economy demands that we acknowledge these gaps. The River report is a valuable contribution, but it is not the final word. It tells us that self-custody is growing, but it does not tell us whether that growth is sustainable. What happens when the next bull market peaks and people rush to sell? Will they move their coins back to exchanges, or will they hold and use DEXs? The infrastructure for selling from self-custody is still clunky. Most users still need to deposit to a centralized exchange to cash out. That creates a bottleneck and a potential point of failure.
So what is the takeaway? The River report is a snapshot of a trend in motion, but the real story is the tension between ideology and pragmatism. Self-custody is a powerful expression of the Bitcoin ethos, but it is not a silver bullet. The market is fragmented: some users are moving to full self-sovereignty, others are opting for regulated custodians, and many are doing a hybrid. As a market lead, I see this every day. The key is to provide infrastructure that respects both choices. We need better key recovery mechanisms, more user-friendly multisig, and clearer education about the risks of self-custody.
Building bridges in a fragmented digital frontier means we don't have to choose between centralized trust and decentralized chaos. The River report shows that the market is already building those bridges. The question is whether we are building them fast enough. The next time a major exchange fails—and there will be a next time—the self-custody trend will accelerate again. But those who lost their keys will not be there to benefit. As I tell my team, 'Trust is the only currency that matters, but it must be earned with transparency.'
Looking ahead, I predict that the next phase of this trend will be driven by regulatory clarity. The European Union's Markets in Crypto-Assets (MiCA) regulation, which takes full effect in 2026, will require exchanges to offer segregated accounts and proof of reserves. This could reduce the incentive for self-custody among risk-averse holders. Conversely, the U.S. SEC's ongoing enforcement actions could push more users toward self-custody as a form of protest. The River report does not address these regulatory dynamics, but they will shape the future of the data.
In the end, the River report is a powerful reminder that Bitcoin's most important innovation is not the technology itself, but the trust architecture it enables. Self-custody is not just a technical choice; it is a statement about who you trust. The numbers are moving, but the direction is clear: the decentralized economy is being built by the people who hold their own keys. The question is whether we can build the bridges to make that path safe for everyone.