TL;DR:

• Coinbase reported $719 million in net income for Q3 2026, marking a 66% increase year over year. However, total revenue declined 35% to $1.22 billion, while prediction markets generated more than $200 million in annualized revenue as the exchange expanded beyond traditional crypto trading.
• Coinbase reported a significant increase in third-quarter net income, returning to profitability after a challenging period for its core trading business.
• According to its October 29 earnings report, Coinbase generated $719 million in net income, compared with $433 million during the same quarter last year.

Total revenue reached approximately $1.22 billion, representing a 35% decline year over year. The results highlight the difference between Coinbase’s reported profitability and the pressure still affecting its traditional trading operations.

Prediction Markets Become a New Revenue Driver

One of the most notable developments in Coinbase’s Q3 earnings was the expansion of its prediction markets business.

The company reported that annualized revenue from prediction markets surpassed $200 million, approximately doubling from the previous quarter. CEO Brian Armstrong described the segment as one of the fastest-growing products in Coinbase’s history.

Bernstein analysts noted that prediction markets and derivatives helped offset weaker activity in spot cryptocurrency trading. Together, these businesses accounted for approximately 25% of Coinbase’s transaction revenue.

The development supports Coinbase’s broader ambition to become an everything exchange, bringing multiple financial products under one platform. Bitcoinist previously covered Coinbase’s expansion into prediction markets and stock trading, which forms part of this wider strategy.

Trading Revenue Remains Under Pressure

Despite the growth of newer products, Coinbase’s traditional trading business continued to face challenges.

Transaction revenue fell 43% year over year to approximately $590 million. Spot trading volume also declined 8% quarter over quarter, reaching $135 billion. Derivatives trading volume decreased 7% from the previous quarter to approximately $986 billion, although the decline was smaller than the broader industry’s 11% contraction, according to Bernstein.

Meanwhile, subscription and services revenue contributed $570 million, accounting for nearly 49% of net revenue. Stablecoin-related revenue reached $328 million, increasing 3% quarter over quarter.

This growing contribution from subscriptions and stablecoins reflects Coinbase’s efforts to reduce its dependence on market-driven trading fees. The exchange has also continued expanding its USDC ecosystem, including through its growing stablecoin integration strategy.

Coinbase’s Diversification Strategy Faces Its Next Test

The Coinbase Q3 earnings and latest financial results demonstrate progress in building revenue streams beyond conventional cryptocurrency trading.

The company’s prediction markets business, derivatives offerings, and subscription services are becoming increasingly important as trading activity fluctuates with broader market conditions. However, the $715 million investment gain behind its quarterly profit also highlights how closely Coinbase’s reported earnings remain connected to cryptocurrency valuations.

The exchange will need to demonstrate that its newer businesses can sustain growth even when digital asset prices and trading volumes weaken.

For now, Coinbase’s Q3 results offer evidence that its diversification strategy is gaining traction, although its core trading business remains exposed to the broader crypto market cycle.

Exchange Tokens Are Being Redesigned Around One Question: How Do You Stop People Leaving?

The airdrop cycle of 2024 and 2025 produced a repeatable pattern, and everyone in the market can recite it. A protocol announces a points programme. Volume climbs. Wallets multiply. The token generates, the airdrop lands, and within weeks a large share of the supply has been sold and a large share of the activity has moved to whichever protocol announced points next.

The mechanism worked exactly as designed. It bought attention and volume for a period, at a price paid in supply. What it did not buy was anything that stayed after the payment stopped.

Perp DEXs have been the most aggressive users of this playbook, and they now have the most reason to move past it. The sector is no longer small enough for growth-at-any-cost to be the only objective: on-chain perpetual volume has climbed from roughly 2% of the total perpetuals market in January 2024 to over 10% by early 2026, with monthly volumes crossing a trillion dollars, and the centralised share of open interest falling from 96.4% at the start of 2025 to 86.5% by April 2026 according to CoinGecko’s research. A sector at that scale is competing for retention, not for first attention.

The token designs now being published reflect that. Three shifts are worth tracking.

The first design change is separating the tradeable asset from the governance asset.

In the standard model, one token does both jobs: it trades on the open market and it votes. The consequence is that voting power is purchasable at any moment, including immediately before a contentious vote, and including by someone with no interest in the protocol beyond the outcome of that vote. Governance attacks of this shape have been executed repeatedly.

The alternative is a non-transferable staking derivative. Stake the liquid token, receive a governance token that cannot be sold, lent or borrowed. Voting weight then has to be accumulated by locking, which means it can only be held by parties who have accepted the opportunity cost of holding.

EVEDEX’s published design for $EVE follows this shape. Staked EVE converts one-to-one into stEVE, which is non-transferable and carries the voting rights in the EVEDEX DAO. It is a straightforward mechanism, and its main effect is to make governance capture expensive in time rather than merely expensive in capital.

The second change is in how unstaking works, and it is more consequential than it sounds.

Lockups have traditionally been binary. Funds are locked for a term, then they are not. This produces the unlock-cliff dynamic familiar from every vesting chart: nothing happens for months, then a large tranche becomes liquid on a known date, and the market prices it in advance.

The newer approach makes exit continuously available but progressively priced. Under the $EVE design, unstaking runs linearly across a year, updating continuously, and holders can withdraw the unlocked portion at any point. Exiting early carries a penalty of up to 50%, declining linearly to zero as the period completes.

The third shift is the most important and the least visible in marketing material.

Emissions-funded staking rewards are circular. The protocol pays stakers in newly issued tokens, the yield looks attractive, more tokens are staked, and the source of the yield is dilution of the same holders receiving it. The advertised APY is a number the protocol chooses, not a number it earns.

Revenue-indexed rewards work differently. The $EVE design explicitly does not express staking rewards as a fixed annual rate; returns are tied to protocol fee generation and to each staker’s share of total staked supply. That produces a lower and more variable headline number, which is commercially unhelpful and analytically honest. It also means the token’s return profile is a function of whether the exchange actually does business — which, for a venue reporting more than 74,000 registered traders and monthly volume above $15B, is at least a testable proposition.

None of this addresses the central asymmetry, which is allocation.

$EVE has a fixed supply of 100 million, with 31% to treasury and DAO, 18% to the core team, 15% to partnerships, 12% to liquidity provision and vaults, 10% to seed and private rounds, 8% to network incentives and 6% distributed retroactively to early users based on accumulated points and badges. That 6% figure is roughly in line with what comparable programmes have distributed, and it is worth stating plainly: retroactive community allocations in this sector are small relative to insider allocations, and elegant staking mechanics do not change that arithmetic.

The other unresolved question is timing. The $EVE token generation event is scheduled for a window of approximately six to eighteen months after mainnet launch — a range wide enough that the market conditions at TGE are unknowable, and unlock schedules for team and investor tranches are the detail that will determine how the design performs in practice.

The direction of travel is real, though. Exchange tokens are being built to reward staying rather than to reward arriving. Whether the mechanisms hold when the first serious drawdown tests them is the part nobody can model in advance.

Visit https://evedex.com/, and follow EVEDEX on X, Telegram and Discord.