Bitcoin’s failure to follow the broader risk-asset rally for two straight weeks is causing pain for short-term ETF holders, but a very different signal is emerging from the off-exchange market. Market maker and OTC desk Wintermute says long-term funds have started buying BTC in tranches through OTC trading, seeing current levels as attractive on an 18-month horizon. The observation comes as roughly $2 billion flowed out of spot Bitcoin and Ether ETFs while fresh capital chased AI stocks and small caps, according to the original market update.
Steady block-by-block demand on OTC desks often reflects deep-pocketed investors building positions without tipping on-exchange order books. This kind of accumulation can reduce liquid supply long before it shows up in spot price action. Wintermute frames the buying as conviction-driven rather than a short-term trade, the kind of positioning you see when funds allocate with a multi-year view.
Wintermute flagged a key downside support band between $60,000 and $65,000. That zone matters not just as a technical level but because it likely aligns with on-chain cost bases and historical accumulation ranges. If it holds, the OTC buying could act as a floor beneath the current grinding price. But the flip side is clear: a macro shock that slices through that band would test the conviction of even the most patient allocators.
For those tracking institutional infrastructure, the very fact that OTC desks are channeling steady demand is a mark of market maturation. The institutional toolkit has expanded rapidly, with weekly tokenization and on-chain settlement milestones making it easier for traditional funds to get crypto exposure without touching centralized exchanges. OTC trading is a natural next step for allocators who want to avoid slippage and front-running.
The $2 billion that bled from BTC and ETH ETFs over the same two-week stretch tells a parallel story. ETF flows are heavily influenced by sentiment and momentum; OTC buying by long-term funds is not. The divergence doesn’t guarantee an imminent price reversal, but it does raise questions about how much of the recent selling is transient.
When ETF outflows dominate headlines, it’s easy to miss the quiet positioning happening away from the tape. That doesn’t mean the caution is unwarranted. If the AI trade continues to suck liquidity out of risk assets, Bitcoin could struggle to hold $60,000 regardless of how much smart money is accumulating. The 18-month time horizon is a luxury most ETF traders don’t have.
The capital moving from crypto ETFs into AI stocks and small caps has parallels inside the digital asset space itself. Projects bridging AI and Web3 continue to attract developer and venture attention even as liquid crypto assets stall. One recent example is the partnership between UXLINK and Origins Network to deliver scalable AI infrastructure on decentralized rails, a sign that the AI theme is far from exhausted within crypto.
At the same time, regulatory friction in Washington adds another layer of uncertainty. With banks trying to derail the biggest crypto bill in US history just days before a Senate vote, the gap between near-term policy noise and multi-year accumulation widens. Long-term buyers may be pricing in a future where the regulatory picture is clearer, but the path there remains bumpy.
For now, the market is navigating two conflicting currents. Visible ETF flows scream risk-off, while OTC desks report patient buying at what Wintermute considers attractive levels on a long view. The next test is whether Bitcoin can hold the $60,000 to $65,000 support without a fresh catalyst. OTC volume data isn’t public in real time, so the only timely gauge will be how spot price reacts around that zone.
The divergence also highlights how crypto is maturing from a single-threaded narrative into a market where different cohorts of capital operate on entirely different timeframes. That makes the top-line ETF numbers less useful as a standalone signal. The smart money appears to be buying the dip quietly; whether that translates into a floor or a trap depends on macro forces that are currently tipping toward AI.


