Bitcoin held on centralized exchanges has fallen to approximately 2.56 million BTC as of June 15, 2026, the lowest level recorded since early 2020. The drawdown represents a net withdrawal of roughly 440,000 BTC — approximately $28.2 billion at current prices — from exchange wallets over the past...
"The supply is just not available on exchanges the way it used to be. Between ETFs, corporate treasuries, and self-custody, you're seeing a structural shift in where Bitcoin sits." — Martin Hiesboeck, Head of Research, Uphold
Bitcoin held on centralized exchanges has fallen to approximately 2.56 million BTC as of June 15, 2026, the lowest level recorded since early 2020. The drawdown represents a net withdrawal of roughly 440,000 BTC — approximately $28.2 billion at current prices — from exchange wallets over the past 12 months, according to on-chain data aggregated by CryptoQuant and Glassnode.
The decline runs counter to typical late-cycle behavior, where exchange balances tend to rise as holders prepare to sell. Instead, Bitcoin is migrating into three distinct custody channels: spot ETF trusts holding 1.28 million BTC ($84.7 billion in AUM), corporate treasuries led by Strategy's 846,842 BTC position, and self-custody wallets bolstered by regulatory clarity from both the SEC and a January 2025 executive order affirming self-custody rights.
At $63,961 per BTC on June 18, 2026, the price sits 1.8% below the prior day's open, pressured by the Federal Reserve's hawkish stance under Chair Kevin Warsh. The paradox: exchange supply is thinning while price action remains flat. The data suggests a market where available liquidity is shrinking, but incoming demand has not yet materialized to absorb the reduction.
On-chain data from multiple analytics providers paints a consistent picture of sustained exchange outflows:
The trajectory reversed sharply starting mid-2025. Net outflows of approximately 45,277 BTC were recorded in a single 30-day period in April 2026 alone, worth roughly $3.4 billion. Since 2021, approximately 780,000 BTC have moved off exchanges in aggregate.
The scale of this drawdown is notable in context: with 20.04 million BTC in circulation, exchange-held supply now represents approximately 12.8% of total mined coins. At the 2020 peak, that figure exceeded 16%.
The 440,000 BTC that left exchanges over the past year did not vanish. It migrated into identifiable pools:
1. Spot Bitcoin ETFs U.S. spot Bitcoin ETFs collectively hold 1,283,847 BTC as of mid-June 2026, representing approximately 6.2% of total circulating supply. BlackRock's IBIT leads with roughly $67 billion in AUM, followed by Fidelity's FBTC at $17 billion. The combined AUM of all U.S. spot Bitcoin ETFs stands at $84.7 billion — down from a peak of $96.5 billion in April 2026 due to price depreciation and a record nine-day outflow streak in late May that saw $2.8 billion exit.
2. Corporate Treasuries Public companies holding Bitcoin as a treasury asset represent a growing share of supply absorption. Strategy (formerly MicroStrategy) holds 846,842 BTC at an average cost of $66,384.56 per coin, with a total cost basis of $33.1 billion. A June 2026 crypto rout erased $62 billion in combined market capitalization from public Bitcoin treasury companies, including Strategy, Tesla, and Marathon Digital. In a notable shift, Strategy disclosed a sale of 32 BTC during June — the first departure from its "never sell" policy.
3. Self-Custody The remaining outflows point to individual and institutional self-custody. The January 2025 U.S. executive order explicitly affirmed the right to self-custody digital assets and conduct peer-to-peer transactions, providing regulatory backing for a practice that had previously existed in a gray area.
The migration of Bitcoin into ETFs has created a custody concentration that merits attention. Over 80% of ETF-held Bitcoin assets flow through Coinbase Prime as custodian, creating a single point of dependency worth approximately $74 billion. Coinbase Prime secures roughly 12% of total crypto market capitalization and supports over 470 assets.
BlackRock's IBIT prospectus names Coinbase as its Bitcoin custodian, with Anchorage listed as an available alternative custodian — though BlackRock has disclosed no current plans to diversify. Fidelity, by contrast, self-custodies through its own digital asset subsidiary, Fidelity Digital Assets.
The custodial concentration raises operational risk questions. If exchange-held supply continues to decline while ETF custody remains concentrated, a single custodian failure — however unlikely — would represent a systemic risk to a material portion of Bitcoin's accessible supply.
OTC desk fee structures for institutional transactions generally converge around 0.1–0.5% of total cost, though actual rates vary based on relationship volume, asset liquidity, and negotiation. This cost structure incentivizes large holders to move off exchange and into direct custody arrangements.
The corporate treasury model, popularized by Strategy starting in 2020, has expanded to at least 154 public companies. The model's fragility was exposed during June 2026: a $62 billion market capitalization wipeout across public Bitcoin treasury holders underscored the leverage embedded in these positions.
Strategy's 846,842 BTC position at an average cost of $66,384.56 leaves the company above water at current prices of $63,961 — but only barely. At current spot, Strategy's unrealized loss per coin is approximately $2,424, representing a total unrealized loss of roughly $2.05 billion. The sale of 32 BTC in June, while immaterial in size, was symbolically significant: it marked the first time the company acted against its "buy and hold" thesis.
Corporate treasuries represent a category of exchange withdrawal that is unlikely to reverse in the short term. These positions are held in institutional-grade custody solutions and governed by board-level decisions, making them less responsive to short-term price signals than retail exchange balances.
Self-custody growth introduces a secondary supply dynamic: permanent loss. Analysts at Chainalysis estimate that between 2.3 and 3.7 million BTC are permanently inaccessible, representing 11–18% of Bitcoin's 21 million fixed maximum supply. Causes include forgotten passwords, destroyed hardware wallets, deceased owners without succession plans, and early-era coins that have never moved — including an estimated 1 million BTC attributed to Satoshi Nakamoto's early mining activity.
The inheritance problem is becoming more acute. A 2026 report from the Gannett Trust framed this year as a turning point for early Bitcoin adopters beginning to address succession planning. The challenge is structural: unlike a bank account recoverable with a death certificate, self-custodied Bitcoin is permanently inaccessible once private key access is lost. Legal authority alone cannot move coins.
The first notable decline in self-custodied Bitcoin in 15 years has been documented, according to Uphold's Martin Hiesboeck. Some long-term holders are migrating assets into ETFs, motivated by tax advantages and simplified administration — effectively reversing the self-custody trend for a subset of holders.
This creates an unusual dynamic: exchange supply is declining, but so is a portion of self-custodied supply. The net effect is a shrinking effective float.
Several regulatory developments in 2025–2026 have reshaped the custody landscape:
SEC Staking Clarity: The SEC declared that proof-of-stake staking activities and liquid staking do not constitute securities transactions. This interpretive rule covers protocol staking, protocol mining, airdrops, and token wrapping of non-security crypto assets, removing a source of regulatory uncertainty that had previously deterred institutional custody.
FDIC Custody Standards (April 2026): The FDIC approved a proposed rule establishing custody and reserve standards for any FDIC-supervised institution providing crypto safekeeping services. Requirements include treating crypto as customer property, segregating it from institutional assets, protecting it from creditors in bankruptcy, and publishing monthly audit-verified reports. Banks can begin submitting applications as early as July 2026.
Federal Agency Joint Guidance: The Federal Reserve, FDIC, and OCC issued a joint statement clarifying that banks must follow existing risk management rules when providing Bitcoin custody services. This cleared a path for traditional financial institutions to enter the custody market.
The regulatory framework is channeling Bitcoin away from unregulated exchange wallets and into supervised custody structures. The trend reduces counterparty risk for individual holders but concentrates systemic risk among a smaller number of institutional custodians.
The supply picture sharpens when losses and locked positions are factored in:
| Category | BTC | % of Max Supply | |---|---|---| | Total mined (circulating) | 20,043,290 | 95.4% | | Remaining to mine | ~956,710 | 4.6% | | Estimated permanently lost | 2.3–3.7M | 11–18% | | Held in ETFs | 1,283,847 | 6.1% | | Strategy (corporate) | 846,842 | 4.0% | | Exchange-held | 2,560,000 | 12.2% | | Daily new supply (mining) | ~450 BTC/day | — |
The effective float — coins readily available for trading — is substantially smaller than the headline circulating supply suggests. If the midpoint estimate of 3 million BTC is permanently lost, and ETF/corporate positions remain largely static, the tradeable supply may be closer to 14–15 million BTC, of which only 2.56 million sits on exchanges.
At current mining rates of 3.125 BTC per block (post-April 2024 halving), approximately 450 new BTC enter circulation daily. Against a backdrop of 440,000 BTC in annual exchange withdrawals, new supply replaces less than half of what is being withdrawn.
The Bitcoin exchange supply drawdown is not a single-cause phenomenon. It reflects converging forces: ETF-driven institutional accumulation, corporate treasury lockup, regulatory frameworks that incentivize supervised custody over exchange storage, and a self-custody culture that simultaneously grows the long-term holder base while permanently removing coins from circulation through inheritance failures.
The market implications are mechanical rather than predictive. A thinner exchange float means that when demand does arrive — whether from macro shifts, ETF inflow reversals, or institutional mandates — it will encounter less available supply at prevailing prices. Conversely, the concentration of custody among a small number of entities introduces fragility that did not exist when Bitcoin was more broadly distributed across exchanges.
The data is clear on what is happening. What it does not resolve is whether the shrinking exchange float represents conviction, complacency, or simply the migration of risk from one set of intermediaries to another.