Almost everything written about perpetual futures treats them as a trading venue: where you go for leverage, where volume concentrates, where the next hot asset gets listed. That's true, and it's the least interesting thing about them. The more important story is that perps built the most underrated interest rate in crypto, and the capital priced off that rate has grown far faster than the attention paid to it.
Start with the funding rate. A perpetual future has no expiry to pull it back to spot, so the two sides exchange a periodic payment, funding, to keep the perp anchored to the underlying. When longs are crowded, longs pay shorts; when shorts are crowded, shorts pay longs. It looks like plumbing, but it is the closest thing crypto has to a live, market-clearing price for leverage: what it costs, right now, to hold a levered position. It updates continuously, and it is set by actual positioning.
Funding is not crypto's only interest rate. The market has quietly assembled a young, fragmented rates complex: floating rates on lending protocols, funding on perps, term rates implied by dated-futures basis. Each prices something different. Lending rates set borrowing demand against supply, but they respond less directly to speculative positioning than funding does, and displayed yields can be distorted by temporary token incentives that have nothing to do with real borrowing demand. Funding, by contrast, is speculation's own price, marked to market every interval. The practitioner's edge isn't declaring one rate real and the others fake, it's reading the divergences. When funding rips away from lending rates, that gap tells you exactly where leverage demand lives, the way the TED spread once flagged stress in dollar funding.
A serious amount of capital is already priced off funding, and most people don't see the wire. The delta-neutral, or "basis," trade, long spot, short the perp, collect funding, is the engine behind billions of dollars in onchain yield strategies whose return stream is, mechanically, the funding rate.
The distinctions matter here, because the market has gotten sloppy about them. Most major stablecoins are backed by Treasury bills; their yield is a T-bill yield. Tokenized Treasury, credit and gold products derive their returns from the assets they hold. None of that is funding-rate exposure. But a distinct class of products, basis vaults, delta-neutral yield strategies, synthetic dollars that hedge crypto collateral with short perps, earns funding as a core revenue stream. A T-bill yield and a funding-rate yield behave nothing alike when markets turn.
Funding is positive most of the time, because most of the time crypto is net-long and hungry for leverage. So a basis strategy prints a steady, boring-looking return, and it's tempting to call that return "safe." I'd put it differently: it's a rate, and rates are conditional. When positioning flips, a sharp deleveraging, a risk-off week, a crowded long unwinding, funding can go negative, and the position that was quietly earning is quietly paying. The trade is also only as neutral as its execution: the short leg has to be margined and rebalanced across venues, and in a real dislocation, mark-price divergence, collateral mismatch, or an inability to move margin can break the hedge operationally even when it remains economically sound.
None of this is a criticism of basis trades. They are among the oldest and most legitimate strategies in finance, and I say that as someone who runs them. The problem is narrower: a conditional return being sold as an unconditional one.
The thing to watch is crowding. When too much capital chases the same basis, it competes away the funding it came to collect, right up to the point where the return no longer compensates for the tail. Compressed funding against steadily climbing strategy assets isn't a systemic alarm; it's a positioning signal that expected returns are falling. The real danger begins if managers respond by adding leverage, thinning their liquidity buffers, or moving to weaker venues to defend the same headline yield.
And funding is about to matter well beyond crypto. Perps are onshoring: the CFTC has approved the first US-regulated bitcoin perpetual and is now consulting on how far these contracts should extend into physically delivered commodities like crude oil. At the same time, perps are eroding the closing bell: when traditional markets are shut, 24/7 perps keep absorbing information and setting the price before the reopen confirms it. The open question is no longer whether markets should trade continuously, but how to keep thin overnight liquidity from producing manipulated prints and forced liquidations.
Which points to the most underrated implication of all: the first trillion dollars of real-world assets to arrive onchain may not be tokenized, they may be perpified. A perp needs no custody, no minting, no corporate actions, not even the issuer's cooperation. Hyperliquid's HIP-3 and the new wave of stock and pre-IPO perps are rapidly expanding the number of permissionless listing venues, and the market often comes onchain before the asset does. The sequencing serves both sides: perps bootstrap liquidity and price discovery, while fully backed tokenized products, which take longer to launch properly, deliver ownership. Tokenization creates true ownership; perps create a capital market. Once both exist, arbitrage between tokenized spot and perps connects their liquidity, and funding keeps the derivative anchored to the underlying.
As larger, more traditional balance sheets arb the same basis, funding gets more efficient and more compressed, good for market structure, uncomfortable for any product that quietly assumed fat, permanent funding. The strategies that survive will be the ones honest that their yield is a rate, that the rate is conditional, and that conditional rates go down as well as up.
Perps built crypto's rates layer. Funding rates are its most important real-time benchmarks: fast-moving, market-clearing, and highly revealing about leverage demand and positioning. It's time we started reading them that way.