Crypto winter has a branding problem.
The phrase makes it sound like the chain goes quiet, wallets stop moving, and the whole machine turns cold. However, the cleanest proof of retail pulling back rarely lives on-chain.
The people who vanish first aren’t the power users bridging stables into DeFi or the long-term holders shuffling coins between cold storage addresses. They’re the casual participants who show up when risk feels fun, open a broker app, tap market buy, and then disappear without leaving a neat on-chain footprint.
That’s why the most usable retail barometer sits in an often overlooked place: the earnings lines of Robinhood and Coinbase.
When retail activity thins out, brokers feel it as fewer trades, lower notional, and less transaction revenue. When retail warms up, it shows up as higher engagement and higher take.
You can have a Bitcoin chart that looks alive while participation is shrinking, because price is now carried by a narrower set of buyers using ETFs, futures, and other structured products.
A participation recession can coexist with a price rebound. You only need to look at what these two companies just reported to see how that split looks in practice.
Robinhood’s fourth quarter made the point in numbers that are hard to argue with. Total net revenues rose 27% year over year to $1.28 billion, with transaction-based revenues up 15% to $776 million.
But the composition of that revenue is important.
Options revenue came in at $314 million, up 41%, and equities revenue hit $94 million, up 54%. Crypto revenue, on the other hand, fell to $221 million, down 38% YoY.
That’s what a retail rotation looks like.
Coinbase, which many still treat as a proxy for retail crypto demand, reported the same chill from a different angle.
In its Q4’25 shareholder letter, total revenue was $1.781 billion, with transaction revenue at $982.7 million and subscription and services revenue at $727.4 million. Consumer transaction revenue was $733.9 million for the quarter, down from $843.5 million in Q3. Institutional transaction revenue rose to $185.0 million from $135.0 million. The company also reported a $667 million net loss for the quarter.
Put those together, and you get the same problem as Robinhood: retail activity cooled, the business leaned harder on non-transaction lines, and the quarter made more from its services stack than trading.
The retail barometer lives in broker P&L
On-chain metrics can tell you whether whales are distributing, whether long-term holders are spending, whether stablecoin supply is expanding, and whether the base layer is busy.
But they can also mislead you about retail participation because the retail cycle is about people actively trading, not just coins moving.
A lot of today’s flow sits inside wrappers where the chain never sees it. If someone buys exposure through a broker, hedges it with listed options, or trades within an internal venue, the user experience is busy, but the chain can look calm.
Robinhood is built around that user experience, so we can look at its quarterly report like a behavioral survey with a P&L attached. The company ended Q4 with 27 million funded customers and an ARPU of $191.
Those might not be crypto-native metrics, but they’re exactly what you want when you’re trying to answer one plain question: are people still participating?
The participation answer in Robinhood’s case is yes.
But the risk answer is more specific: retail has leaned into instruments that offer defined outcomes and fast feedback, with options and event contracts being the most popular.
Operating data makes that clearer.
Options contracts traded hit 659 million in Q4, up 38% year over year. Crypto notional trading volumes were $82 billion, with $48 billion tied to Bitstamp and $34 billion on the Robinhood app, where notional fell 52% year over year. Event contracts traded reached 8.5 billion in Q4.
Robinhood can call 2025 a record year and still show you a crypto winter in the exact place it actually hurts a retail-facing broker: the crypto revenue line and the app’s crypto notional.
Transaction-based revenue got a lift from equities and options, while crypto lagged at $221 million and missed expectations that clustered higher. That helped explain why the quarter disappointed, even with record net revenue.
That matters because it frames crypto winter weakness as a participation issue, not a product failure. The platform kept its audience, but the audience just did less crypto trading.
Coinbase is different because it sits closer to the core venue economy. Retail and institutional flow share the same brand even when they behave differently.
The shareholder letter spells out the mix shift without needing any extra interpretation: transaction revenue for Q4 was $983 million, down 6% quarter over quarter.
Coinbase attributes the consumer decline to weaker consumer spot volume and mix shifts. Institutional transaction revenue rose quarter over quarter, even as institutional spot volume fell.
When a quarter looks like that, it means retail is stepping back while institutional flow becomes relatively more important.
It also means the business model is moving toward recurring revenue, so it doesn’t live and die on the next trading frenzy. That kind of winter-proofing is easiest to see in the subscription and services section.
Coinbase reported $727.4 million in subscription and services revenue in Q4 and $364.1 million in stablecoin revenue alone. Stablecoin revenue helped cushion the hit from weaker trading volumes.
That is, without a doubt, the most misunderstood part of the cycle, because the market assumes that crypto winter equals inactivity.
However, in practice, crypto winter often means that the business of crypto moves toward rails, custody, and yield-like revenue streams that keep working even when retail goes home.
Price can recover while participation stays thin
A crypto winter becomes easier to understand once you separate the price of Bitcoin from the breadth of participation around it. Price can be supported by a smaller set of buyers using regulated wrappers, hedging instruments, and institutional balance sheets.
That can keep the chart alive while the culture of participation feels muted. You see it when the big numbers concentrate in fewer pipes and the spillover into everything else fades.
Coinbase’s own operating notes hint at that concentration. Consumer spot trading volume was $56 billion in Q4, while institutional spot trading volume was $215 billion.
You don’t have to romanticize institutional adoption to see what that implies. In quarters like this, the market can function with fewer participants, but it behaves differently. It can rally on reallocations, hedge flows, and macro positioning, without lighting up the broader set of behaviors that people associate with a full mania.
Robinhood’s quarter gives you the retail version of that.
People are still trading, but crypto is no longer the default outlet for that energy. Options revenue was up 41% year over year, and event contracts became a central product line that the company chose to spotlight.
The appetite for action got redirected into instruments that feel more controllable, more game-like, or more legible in a market where sentiment turned sour.
That redirection also explains why staring at on-chain activity can be confusing.
On-chain can look stable because the users who remain are the ones who actually use the rails.
Meanwhile, the marginal participant who drives the emotional volume of a cycle can disappear without leaving a neat signature, because that participant’s entire relationship with crypto was mediated through apps, wrappers, and broker interfaces.
Coinbase tied its weak quarter to a broader crypto selloff and pointed to the way trading volumes can collapse quickly when risk sentiment breaks.
Robinhood made a similar point from the other side, showing that equities and options can keep the retail engine running even when crypto cools.
So where did retail risk go?
Robinhood’s numbers hand you three answers.
First, it went into listed options, with 659 million contracts traded in Q4. Second, it went into event contracts, with 8.5 billion traded in the quarter. Third, some of it just stopped expressing itself through crypto notional on the Robinhood app, which the company said fell 52% year over year.
Coinbase’s answer is that retail cooled, institutional flow held up better, and the company leaned harder on stablecoin-driven revenue and other subscription and services lines to keep the business less dependent on retail churn.
All of this tells us that when retail steps back, the industry rebalances around the parts that can keep earning.
However, markets can recover before people do, and price can stabilize while participation stays selective.
The first place you’ll see the crypto winter ending and the crowd coming back will be the earnings line that records whether people are clicking, trading, and paying spreads again.