Coinbase’s prediction markets have crossed $200 million in annualized revenue in the third quarter of 2026 — one of the few bright spots in an earnings report that showed the exchange’s traditional trading business still shrinking.
The figures, published October 29, 2026, come at an awkward moment for the prediction-market industry. Just weeks earlier, a $10 undercover bet triggered 13 felony charges against rival platform Kalshi in Utah — a reminder that the regulatory perimeter around event contracts is still being drawn. Coinbase’s numbers now land in the middle of that debate.
## The headline: $200 million and doubling
According to company data, prediction markets have grown at an 82 percent compounded quarterly rate since launch, making them among the fastest-growing products in Coinbase’s history. Annualized revenue doubled from roughly $100 million in the previous quarter to more than $200 million in Q3.
Bernstein analysts credited the expansion of prediction markets and derivatives with offsetting weakness in spot trading. For an exchange that spent a decade depending on volatile crypto trading fees, a product line less tied to crypto volatility is strategically significant — a point Coinbase Chief Financial Officer Alesia Haas underscored when she described prediction markets as an additional revenue source less correlated with crypto price swings.
## The rest of the ledger tells a harder story
The prediction-market surge sits inside a mixed quarter. Total revenue came in at $1.22 billion, essentially flat compared to the previous quarter, while overall revenue fell 35 percent year over year.
Net income told a different story: $719 million, up 66 percent from $433 million a year earlier, with earnings per share of $2.51. That ended a run of three consecutive quarterly losses — but the improvement rested heavily on an accounting item. Coinbase booked a $715 million gain on crypto held for investment purposes, up from $424 million in the same quarter of 2025. Adjusted net income was far more modest at $78 million.
Spot trading continued to lose ground. Spot volume declined 8 percent from the previous quarter to $135 billion, and the platform’s share of global crypto trading fell to 10.7 percent.
## Subscriptions quietly carry the model
Beneath the prediction-market noise, the quieter transformation is the steady rise of recurring revenue. Subscriptions and services revenue grew quarter over quarter and now accounts for nearly half of net revenue, according to Haas — a proportion that would have been unthinkable for Coinbase in its trading-fee heyday. Stablecoin revenue rose 3 percent from the prior quarter to $328 million, of which $21 million was attributable to USDC held in corporate balances. Average USDC holdings hit a record $22 billion.
The significance is resilience. Subscription income, stablecoin float revenue and prediction-market fees are all structurally less dependent on spot-trading volumes, which have historically collapsed whenever crypto prices go quiet. In a quarter where spot volume fell and the Fed held rates, the fact that Coinbase still printed flat revenue and record USDC balances suggests the diversification strategy is doing real work.
## The regulatory shadow
The commercial success arrives while the sector’s legal footing remains unsettled. The Utah criminal case against Kalshi — triggered by an undercover bet on state election markets — showed how quickly state authorities are willing to escalate from fines to felony charges when prediction platforms push into contested event categories.
For Coinbase, the calculus is different but the exposure is real. Its prediction markets sit inside a publicly listed company with a market cap measured in the tens of billions, a shareholders’ report to file every quarter, and far more to lose from a regulatory misstep than any startup rival. That scale cuts both ways: it makes Coinbase a tempting target for regulators wanting to set precedent, but also a more careful operator with stronger compliance resources than the platforms already facing charges.
What the Q3 numbers establish is that prediction markets are no longer an experiment. A $200 million annualized run-rate, doubling quarter over quarter, means event contracts are becoming a load-bearing pillar of the exchange-earnings model. The question for the next several quarters is whether regulators will let that pillar keep being built — and on whose terms.
For everyday traders, the practical takeaway is simpler: the products driving Coinbase’s growth are exactly the ones drawing the most legal scrutiny. If the sector’s expansion continues at anything near its current pace, the collision between prediction-market revenue and prediction-market regulation is only a matter of time.
kalshi eats 13 felony counts over a literal $10 undercover bet while coinbase doubles to $200M annualized in the same quarter. the regulatory asymmetry is the real story here
^ the utah case is basically the test balloon. if a $10 bet becomes 13 felonies, every compliance team in this sector is rewriting playbooks tonight
asymmetric is right. coinbase lobbied first, kalshi bet on the courts. one of those strategies just ate 13 felonies in utah
doubling to 200M annualized while your rival eats 13 felonies over a 10 dollar bet is the most coinbase sentence possible
Bernstein crediting derivatives and event contracts for offsetting weak spot volume tracks with what the earnings showed. Question is what happens to that $200M run rate the moment a few states decide event contracts are gambling.
the second a state copies the utah playbook that 200M run rate reprices. event contracts are one prosecutor away from a compliance wall
already circling tbh. utah went first because a 10 dollar sting was easy, but state lottery boards have been eyeing event contracts all year. one more copycat and that run rate gets repriced fast
lottery boards are the real threat, a few already ruled event contracts need gaming licenses. the 82 percent compounding assumes that moat holds
one more copycat and it reprices is right, but state regulators move in years. by the time lottery boards wake up coinbase will have lobbied half of them
lottery boards dont need years though. utah built 13 felonies off one 10 dollar undercover bet, any ambitious state AG can copy that playbook in a month once the political wind shifts
the run rate holds while the sports calendar stays loaded. the quiet risk is q4 into q1 when event volume thins and that 82 percent compounding hits a real denominator
^ exactly, the 82 percent compounds off a tiny launch base. nfl season is doing the heavy lifting, rerun that number in february
82 percent quarterly compounding since launch is wild. kalshi catches 13 felonies over a 10 dollar bet and coinbase just prints. regulatory arbitrage is the real product
worth noting the 715M was a paper gain on held crypto, not prediction revenue. strip it out and the quarter looks way less shiny
strip the paper gain and the quarter is derivatives and prediction markets carrying spot. that mix works until volatility dies
volatility dying is the exact scenario that kills it. spot fees need chop, prediction markets need events, a quiet quarter starves both at once
right, and bernstein still had to lean on derivatives plus event contracts to make the quarter read well. the 200M is real but it is carrying a shrinking spot book almost single handed
strip the paper gain and prediction markets are the only line that grew without volatility help. thats the part bernstein is actually underlining
a 10 dollar undercover bet triggering 13 felony counts. utah really out here running sting operations on prediction markets lmao
haas calling it uncorrelated revenue is doing marketing work. sports markets are 100 percent correlated with nfl seasonality, the q4 run rate flatters the whole thing
200M annualized on event contracts while kalshi fights felonies over a 10 dollar bet. same product, different lobbyists
82 percent quarterly compounding with a felony case as your competitive moat. coinbase basically let kalshi absorb the legal arrows and monetized the safe lane. cold but it worked
the felony moat is real though. kalshi spends q4 in a utah courtroom, coinbase spends it shipping more nfl markets. that gap compounds faster than 82 percent
82 percent quarterly compounding on event contracts while spot shrinks is coinbase becoming a sportsbook with a custody license. the 715M paper gain just hid how fast the mix flipped
the moat talk ignores that kalshi fighting 13 felonies builds case law for everyone. if they win in utah every exchange benefits, coinbase gets the legal work done for free