Bitcoin’s hard cap is easy to understand: there will only ever be 21 million coins.
What’s hard to understand is that the marginal market is allowed to trade far more than 21 million coins worth of exposure, because most of that exposure is synthetic and cash-settled, and it can be created or reduced in seconds.
That distinction has become Bitcoin’s core paradox in the past year or so.
Scarcity is a property of the asset, while price is a property of the market microstructure that dominates the next aggressive order. When derivatives volume and leveraged positioning become the dominant arena, Bitcoin can trade like an asset with a tight supply and, at the same time, like an asset with effectively elastic exposure.
21 million coins, but a much larger marginal market
Spot is the only venue where a trade necessarily moves actual BTC from one owner to another.
Perpetual and dated futures don’t mint coins, but they do create a second market that can become larger, faster, and more reflexive than spot. Perps are designed to track spot through a funding mechanism and can be traded with leverage, which means a relatively small amount of collateral can control a much larger notional position. That combination tends to pull activity into derivatives when traders want speed, leverage, shorting ability, and capital efficiency.
Price discovery is simply where the next meaningful market order lands. If most urgency lives in perps, then the path of least resistance is set there, even if long-term holders never touch leverage and even if the underlying supply is fixed. In that regime, moves are frequently driven by changes in positioning: liquidations, forced de-risking, hedging flows, and the rapid repricing of leverage. Those flows can overwhelm the much slower process of spot accumulation, because the marginal actor isn’t choosing whether to buy coins but whether to add or reduce exposure.
This is also why visible order book support is a weaker concept than it looks on a chart. Displayed bids can be real, but they’re conditional. They can be pulled, layered, refreshed, or simply outpaced by the volume coming from the larger derivatives complex. Order books are records of resting intent, not execution guarantees.
What the data shows
The Binance BTC/USDT perpetual futures versus spot volume ratio is the cleanest starting point because it quantifies where activity is concentrated.
On Feb. 3, the perpetual-to-spot volume ratio read 7.87, with $23.51 billion in perpetual volume against $2.99 billion in spot while BTC traded around $75,770. On Feb. 5, the ratio was still 6.12, with $15.97 billion in perps volume against $2.61 billion in spot, and the price near $69,700.
The ratios matter because they’re not a minor skew; they describe a market where the dominant source of turnover is a leveraged, shortable venue. In that setup, the next tick is more likely to be set by the repricing of exposure than by incremental spot buying.
The aggregated order book liquidity delta adds a second layer: not just where volume traded, but where liquidity accumulated near price. CoinGlass defines depth delta as the imbalance between bids and asks within a specified range, here ±1% around the current price, which is a way to summarize whether the visible book is bid-heavy or offer-heavy.
The biggest footprint appears on the derivatives side right as the market was entering the drawdown window. Futures liquidity delta printed +$297.75 million on Jan. 31 at 14:00 with BTC around $82,767. Spot later showed +$95.32 million at 18:00 around $78,893. Even by Feb. 5 at 14:00, spot delta still showed +$36.66 million with BTC near $69,486.
This data shows a market where spot bids existed and, in some moments, grew, but price still kept sliding. Once you accept the hierarchy where derivatives are the dominant class, this stops being a contradiction. Displayed liquidity near spot can improve while the larger derivatives venue continues to force repricing through leverage reduction, short pressure, or hedging. When perps dominate turnover, the marginal seller isn’t a real person that’s lost conviction, it’s just a manager managing positions.
Now add the third channel that investors tend to treat as the definitive spot proxy: US spot Bitcoin ETFs. The flow sequence we’ve seen in last week looks like a tug-of-war rather than a steering wheel aimed at the cliff.
Heavy outflows hit on Jan. 21 at about -$708.7 million, then Jan. 29 at about -$817.8 million, then Jan. 30 at about -$509.7 million. Feb. 2 flipped sharply positive at about +$561.8 million, then reverted to -$272.0 million on Feb. 3 and -$544.9 million on Feb. 4.
Public flow tallies like these are widely tracked through aggregators such as Farside and are frequently referenced in market coverage, but they fail map one-for-one to intraday price when the derivatives venue is setting the marginal trade.
It’s also worth being precise about what an ETF flow is and is not. Creations and redemptions are executed through authorized participants. Depending on the product and regulatory permissions, those processes can be cash-based or in-kind, which changes how directly ETF activity translates into spot market transactions in BTC.
In mid-2025, the SEC approved orders permitting in-kind creations and redemptions for crypto ETPs, which is specifically about allowing authorized participants to create or redeem shares using the underlying crypto rather than only cash, bringing the operational structure closer to other commodity ETPs. (SEC) Even with that structure, ETF flows still sit alongside derivatives positioning, dealer hedging, and exchange liquidity, which can dominate short-horizon price formation.
Finally, exchange reserve data anchors this abstract data into something more tangible: the amount of BTC sitting on exchanges, which is a proxy for immediately tradable inventory.
From Jan. 15 to Feb. 5, all-exchange BTC reserves rose by 29,048 BTC, a 1.067% increase, reaching just over 2.75 million BTC.
This matters because it separates two ideas that are often blended together.
Bitcoin can be scarce in total supply and still feel well supplied at the point of transaction if exchange inventory rises into a risk-off window. ETF inflows can be positive and yet the tradable float can expand via deposits, treasury moves, or repositioning by large holders. And even if the tradable float tightens, derivatives can still amplify volatility because exposure can be added or removed faster than coins can move.
A scarcity model that matches how Bitcoin trades
A useful way to reconcile all of this is to treat Bitcoin scarcity as a stack of time horizons rather than a single number.
At the slowest layer is protocol supply, which is fixed by design. That’s the layer the 21 million cap describes.
At the middle layer is the tradable float, which is what can realistically hit the market without friction. Exchange reserves aren’t the best proxy for this, but they’re directionally useful because they measure coins that are already sitting on a platform built for rapid transaction.
At the fast layer is the synthetic exposure: perps, dated futures, and options. This layer can expand or contract extremely quickly because it’s constrained by collateral and risk limits, not by coin movement. When activity concentrates here, a large share of the market is expressing views through leverage and hedges, not through coin acquisition.
At the final layer is the marginal trade itself: the next forced buy or sell that clears through the most active venue. The perpetual-to-spot volume ratios that have been sitting between roughly 6 and 8, combined with the larger liquidity delta prints on futures, show a market where that marginal trade was happening in derivatives, not in spot.
That framing tells us that scarcity is real, but it doesn’t guarantee day-to-day tightness. The market can trade scarce assets through abundant exposure, and the venue with the most urgent flow tends to set the next price.
That’s why we need to treat ETF flows, exchange reserves, and derivatives dominance as three separate lenses that can disagree in the short run. When they line up, moves tend to be cleaner. When they diverge, you can see exactly what the charts show: bids appear, narratives whip around, and price still bleeds because the marginal market is elsewhere.