U.S. spot Bitcoin ETFs logged back-to-back record redemption streaks in May and June, flipping year-to-date flows negative for the first time since launch.
Key takeaways
U.S. spot Bitcoin ETFs suffered their two longest redemption streaks on record in rapid succession during May and June 2026. Combined, the two windows drained an estimated $7.2B from the products and pushed 2026 year-to-date cumulative flows into negative territory for the first time, a milestone confirmed by Bloomberg ETF analyst Eric Balchunas. The selldown came against a total cumulative inflow base of ~$58.72B since the products launched in January 2024, per SoSoValue.
The first streak ran May 15 through approximately May 28-30, producing 10 consecutive outflow sessions. Bloomberg's compiled data put net withdrawals for that window at ~$2.8B; Farside Investors and SoSoValue figures cited across multiple outlets ranged from $2.8B to $3.5B for the same period.
The second streak followed in June, running 13 consecutive trading days for ~$4.37B in net outflows per Cobo citing SoSoValue. That streak ended June 5 with a $3.05M net inflow. Ethereum ETFs ended a separate 17-day outflow streak the same day.
IBIT absorbed the heaviest blow. Farside Investors data puts BlackRock's net outflows at approximately $3.3B during the extended streak. Fidelity's FBTC shed ~$456.6M and Grayscale's GBTC ~$303.6M across the same period. Bitcoin's price dropped from roughly $80,000 to $73,000 during the May window, then fell an additional ~17% during the June streak before stabilizing near $61,000, $63,000.
The catalyst stack is worth laying out explicitly because it clarifies what actually moved the needle.
The May streak was triggered by the May 12 CPI print, which showed reaccelerating U.S. inflation. 10x Research's Markus Thielen attributed the concurrent Bitcoin selloff below $60,000 to institutional selling through spot ETFs following that data, per CoinDesk. The Federal Reserve held rates at 3.5%, 3.75% through the period, keeping the opportunity cost of holding a non-yielding asset elevated.
The equity backdrop made the rotation cleaner. The S&P 500 pushed above 7,568 on AI and semiconductor momentum, offering a high-momentum alternative for multi-asset allocators. Escalating U.S., Iran tensions layered on a broad risk-off impulse. The result: marginal ETF buyers who entered during the 2024-2025 inflow surge rotated into that trade instead. None of those triggers represent a change in Bitcoin's supply schedule, monetary properties, or long-term debasement case. They represent rate-sensitive capital doing exactly what rate-sensitive capital does.
This is the tradeoff that came with ETF approval. More institutional access means more correlated drawdowns. ETF mechanics create structural buyers on the way in and structural sellers on the way out, and those sellers move in herds with macro sentiment. The global bond rout and Iran-driven Treasury yield pressure that have been building throughout 2026 are the same macro environment pressuring these marginal holders. Self-custody holders face no forced redemption pressure from that dynamic.
The $58.72B in cumulative inflows since January 2024 looks durable on the surface. But that aggregate does not distinguish between pension funds and endowments with multi-year lockups and fast-money hedge funds that will rotate again next quarter. The data split is not public. That opacity is itself worth tracking.
The June 5 inflow did not resolve the structural question; it paused the bleeding. The thesis that these outflows represent macro-tourist rotation rather than structural breakdown rests on one condition: the coins leaving ETFs need to be absorbed by self-custody accumulators, not dumped to exchanges.
Glassnode's long-term holder (LTH) accumulation and distribution data is the metric to watch. If LTH wallets show net accumulation concurrent with the ETF outflow windows, on-chain structure is intact and this is a dip absorbed by conviction holders. If on-chain data shows coins flowing toward exchanges and net LTH distribution, treat that as a top signal, not rotation noise. The sentiment data around peak-fear periods has historically coincided with local bottoms. Whether that pattern holds here depends entirely on what the on-chain absorption picture shows.
The 14-day moving average of ETF flows is the secondary signal. A sustained return to positive weekly flows after June 5 would confirm the streak was a macro-driven purge. A rollover back into outflows without a macro catalyst would raise the distributional top question again.
The primary drivers were macro: a reaccelerating U.S. CPI print on May 12, the Federal Reserve holding rates at 3.5%, 3.75%, S&P 500 all-time highs above 7,568 drawing capital into AI and semiconductor equities, and U.S., Iran tensions triggering a broad risk-off move. Bitcoin-specific fundamentals were not the catalyst.
Not necessarily. Total cumulative net inflows since the January 2024 ETF launch stood at approximately $58.72B through early May 2026 per SoSoValue. The YTD-negative milestone reflects the 2026 outflow streaks drawing against that base, but the figure conflates long-duration structural holders with short-duration macro allocators. The current data does not cleanly separate the two.
The on-chain signal matters more than the ETF flow data here. If Glassnode's LTH accumulation data shows self-custody holders absorbing the float that ETF sellers shed, the on-chain structure is intact. If coins exiting ETFs are flowing to exchanges and LTH wallets are net distributing, the distributional top case strengthens. ETF flow reversals alone are insufficient confirmation in either direction.
