On October 15 — the wire copy carries no year, which is its own small confession about how this market now metabolizes exchange product news — Coinbase added two perpetual contracts to its derivatives roster: PONS and CASHCAT.
That is the entire payload. Two tickers. One venue. No leverage caps. No funding interval. No named legal entity. No jurisdiction. No indication whether the book is run by Coinbase Financial Markets under CFTC oversight or by the offshore International Exchange — two very different machines wearing the same logo.
Everyone is watching the price. Nobody is watching the plumbing.
I have been reading plumbing for nineteen years, most of them from a desk in Istanbul where the Bosphorus fog rolls in at 6 a.m. and the FX forwards quote before the coffee does. In 2017 I was handed a junior quant seat and a miserable assignment: model the velocity of funds through the Ethereum ICO boom. Four months, 500 token sales, one on-chain dataset that smelled like a laundromat. What I found was that roughly 60% of "raised" liquidity was recycled back into the same handful of addresses within four hours. The demand was not organic. It was a loop. Tracing the liquidity ghosts through the ICO fog taught me the single most useful habit in this business: never ask what an asset is worth before you ask who is paying whom, how fast, and with what collateral.
So let's do that here.
Context: what a perpetual contract actually is
A perpetual futures contract is a three-legged machine. Leg one is the index — a composite price stitched from the spot markets of a set of constituent venues, usually a median or volume-weighted average with outliers trimmed. Leg two is the mark price, the index plus or minus a clamped basis, used for margining and liquidation. Leg three is the funding rate, a periodic payment between longs and shorts that tethers the perp to the index.
Only one of those three is an opinion. The funding rate is a payment. It is the only honest number on the screen.
For BTC, the index is deep. Dozens of venues, billions in daily volume, outlier trimming that actually works. Mark price and last print converge within a few basis points, and funding oscillates in a boring band around zero. That is what a mature market looks like from inside the machine.
For a newly listed small-cap, none of that holds. The index is thin. The constituent venues may number three or four, and their combined daily volume may be less than a single mid-tier NFT flip. The mark price is not a price; it is a rumor with a timestamp. And the gap between the rumor and the last print is where every liquidation cascade is born.
This is my standing objection to how the industry talks about oracle latency. Everyone treats it as a DeFi problem — a stale feed on a lending market, a mispriced collateral ratio. But the same latency lives inside every centralized venue's index engine, and there it is worse, because it is opaque. On-chain, at least, you can read the feed contract. In a CEX index methodology PDF, you get a paragraph and a prayer. The oracle problem did not get solved by decentralization. It got relocated.
Core: the arithmetic of leverage on a thin float
Here is the part the listing announcement does not tell you, and that the price chart will never tell you either.
A perpetual contract does not need the underlying asset to exist in size. It needs a reference price and two counterparties willing to post margin. That is the whole trick. Leverage does not multiply the float; it bypasses it entirely.
Run the arithmetic on a hypothetical. Suppose a token has $50 million of circulating float and trades $8 million a day across three venues. Coinbase opens a perp with 10x maximum leverage. It takes very little time for open interest to reach $200 million. At that point, notional exposure in the derivative market is four times the size of the asset the derivative is supposed to track. The tail is now bigger than the dog, and the dog is on a leash made of a three-venue median.
Now add the second-order effect. In a market where open interest exceeds float by a multiple, the short side cannot actually deliver the asset in a squeeze scenario. It doesn't have to — perps cash-settle — but it does have to post margin. So a squeeze stops being a scramble for inventory and becomes a scramble for collateral. That is a different animal. It resolves faster, it is more violent, and it produces funding rates that look like typos.
I watched this exact geometry in 2020, when I was running Uniswap V2's constant product formula against FX forward curves to hunt a temporal arbitrage in cross-border settlement. I calculated a 15% risk-adjusted edge and wrote three threads on how impermanent loss correlated with fiat volatility. Then I killed my own bot, because the operational complexity was distracting me from the actual insight: DeFi was building parallel central banks, and parallel central banks have parallel funding markets. The funding rate is the policy rate of that shadow system. When it spikes to 300% annualized, someone is being liquidated, and it is usually not the person who set the parameters.
Which brings us to the listing effect — the belief that Coinbase adding a ticker is a bullish signal.
It was, once. In the late 2010s, a Coinbase spot listing produced a median 24-hour pop in the tens of percent. By the early 2020s that had compressed hard, and by now the effect is a rounding error for any asset with real depth. What remains is a template: announce, pump, distribute. The template works best on assets with thin float and no institutional short base — precisely the profile that gets a perp listing rather than a spot listing.
Tracing the liquidity ghosts through the perpetual fog, what you actually see is a venue adding marginal revenue at near-zero marginal cost. Coinbase already runs the matching engine, the risk engine, the clearing stack, the surveillance layer. Adding a ticker costs a config file and a compliance memo. The revenue, in the form of taker fees on a high-turnover instrument, is real and immediate. This is not a strategic bet on PONS or CASHCAT. It is a bet on the churn rate of the retail trader — the most reliable variable in the entire industry.
And understand the competitive frame. Binance holds somewhere in the range of 40-50% of global crypto derivatives volume. Bybit and OKX sit in the low-to-mid teens each. Coinbase, for all its spot-market compliance premium and its Nasdaq listing, has historically been a single-digit player in derivatives. Perpetual listings are how you close that gap without building anything new. The spot venue is the brand. The derivatives venue is the business.
The rented book
One more piece of plumbing before the contrarian turn.
Someone has to quote the market. On day one, the depth you see on a new perp is not owned by the market; it is rented from designated market makers under a rebate agreement. Negative maker fees, volume tiers, inventory caps, sometimes a floor on uptime. These contracts expire or get renegotiated, and when they do, the book thins.
This is why day-one depth is a liar. It measures the size of the subsidy, not the size of the community. If you want a leading indicator for whether a perp will still have a functioning order book in ninety days, ignore the volume print and watch the spread. A spread that stays tight after the rebate period is real liquidity. A spread that widens from one basis point to eight is a market maker walking out the door.
Contrarian: the listing is bearish for the token and bullish for the exchange
The consensus read is simple: Coinbase lists it, therefore it goes up. I want to argue the opposite, at least about the structure of price discovery.
First, a spot-only asset has a one-sided exposure profile. You can buy it, or you can not buy it. That is a crude but honest market. A perp listing imports a leveraged short side onto an asset that previously had none. Every basis trade, every delta-neutral desk, every funding-arbitrage fund now has a reason to hold the token short against a long elsewhere. The marginal price-setter stops being the token's holder and becomes a trader who does not care about the token at all.
Second, perps create a documented exit channel. Early holders of a thin-float asset face a brutal problem: selling into the spot book moves the price against them. A perp market lets them express the exit in the derivative while the spot price stays cosmetically intact. The chart looks fine. The float is leaving.
Third, and this is where the industry's favorite narrative gets it backwards: nobody who matters cares how many chains the contracts are deployed on. If PONS or CASHCAT is being pitched as an omnichain asset with a presence on eight networks, that is a fundraising slide, not a market structure. The perp prices on one index, sourced from a handful of venues, and the settlement layer underneath the collateral is an implementation detail until it isn't. Users do not pay for interoperability. They pay for depth and a funding rate that doesn't eat them alive.
Bear case: five ways this breaks
One. Index fragility. If the constituent venues are thin, a single wash trade can move the composite and trigger liquidations on a price that no one actually traded at.
Two. Funding whipsaw. New small-cap perps routinely print funding extremes in both directions within the same week, squeezing longs, then shorts, then longs again. The venue collects fees on all of it.
Three. Subsidy withdrawal. When the market maker rebate expires and the book thins, spreads widen by an order of magnitude and the effective cost of any position doubles.
Four. Reclassification. If either asset draws a securities designation, the perp is delisted and open interest is force-settled, usually into the worst possible liquidity.
Five. The unlock cliff. If a large vesting tranche hits a market already carrying open interest several times the float, the reflexive unwind is not a correction. It is a demolition.
None of these require PONS or CASHCAT to be fraudulent. They only require them to be small. Smallness is the risk factor, and it is the one thing the listing announcement never mentions.
Takeaway
There is a version of this story where Coinbase is quietly building the derivatives leg its shareholders have demanded for years, and PONS and CASHCAT are just two more config lines in a machine that already works. That version is probably true, and it is boring, and nobody will write about it.
The interesting version is the one where the venue gets paid regardless of direction, the market maker gets paid for showing up, the arbitrage desk gets paid for the basis, and the only participant holding a non-zero directional opinion is the retail trader who read a headline. Tracing the liquidity ghosts through the blob fog of the next fee-market cycle — because the collateral behind these positions increasingly settles on rollups whose blob space will not stay cheap forever — the question worth asking is not whether PONS and CASHCAT pump.
It is whether, ninety days from now, you can name the entity quoting the book, and whether the index that decides your liquidation is a price or a rumor.