Coinbase Traded One Risk for Another

Less dependent on Bitcoin spot but not less macro-sensitive

Total revenue nonetheless fell 14% sequentially and nearly 19% year-on-year, to $1.22bn. Transaction revenue dropped 21% quarter on quarter. Subscription and services revenue, which is supposed to provide a recurring base, itself declined 5%.

What we observe is that Coinbase has reduced its product concentration without reducing its macro sensitivity. It has diversified its sources. Transaction revenue depends on crypto asset prices, volumes and volatility. A growing share of Coinbase's revenue now depends on the level of US short-term interest rates, through stablecoin revenue, corporate interest income and, to a lesser extent, financing activities. Both are macro variables outside management's control.

The second beta: 29–35% of revenue is rate-exposed

In finance, a business's "beta" measures how sensitive it is to a given external factor. Coinbase's first beta has always been crypto prices, the main driver of its trading revenue. The results show a second one emerging: three revenue lines now carry varying degrees of exposure to US short-term interest rates instead, stablecoin revenue, corporate interest income and interest and finance fee income.

(Source : Coinbase earnings)

The distinction matters: stablecoin revenue and corporate interest income move with Fed Funds, through the yield on USDC reserves and the company's treasury holdings. Interest and finance fee income also depends on lending spreads, loan book size and borrower demand. Including this third line produces a 35% share, mathematically correct, but best presented as a broader definition rather than 35% of revenue directly and exclusively indexed to benchmark rates. On a narrower definition, the broader rate-exposed share of Q2 total revenue was $424.04mn, or 34.8%.

(Source : Coinbase earnings)

The proof: at least $90.8mn of annual rate-driven pressure

One of the most revealing numbers in the quarter comes from the analytical bridges in the 10-Q. Coinbase details two rate-related impacts relative to Q2 2025:

Stablecoin revenue: Coinbase explicitly identified a $55.9mn headwind on Q2 revenue compared to Q2 2025, driven by lower average rates and partially offset by higher USDC balances.

Corporate interest and other income: this line declined $34.9mn year-on-year, a drop that Coinbase attributes to a 73-basis-point decrease in the average rate earned.

Combined, lower rates created a $90.8mn year-on-year revenue headwind in Q2, even as average USDC held in Coinbase products and average DeFi borrow/lend balances reached record levels. This is the clearest evidence that Coinbase's diversification has introduced a meaningful exposure to the rate cycle.

Interest and finance fee income adds another layer. This line grew 11% year-on-year to $66.1mn, driven by record average borrow/lend balances, which reached $1.49bn in Q2 2026 compared to $199mn a year earlier, but the company notes that rates partially offset this growth. Record balances were partly offset by lower rates: the negative rate impact across all three lines is likely larger than $90.8mn, but the available disclosures do not allow a precise quantification.

(Source : Coinbase earnings)

This confirms that lower rates already had a material effect on Q2 revenue. However, the available information does not support the construction of a reliable annualized run rate or a precise sensitivity per 100 basis points. Producing either would require assumptions that could mislead.

USDC rewards complicate the economics

USDC rewards, the yield paid to customers holding USDC on Coinbase, totalled $119.1mn in Q2, up 16% year-on-year. This expense is classified within sales and marketing, but economically it functions as a revenue share with customers on the float business. It is the most directly identifiable cost of attracting and retaining the USDC balances that support stablecoin revenue.

At $119.1mn, USDC rewards represented approximately 41% of reported stablecoin revenue ($292.1mn). This comparison is indicative rather than a genuine margin calculation because the two lines do not cover the same balance perimeter. Stablecoin revenue also includes revenue generated on USDC balances held off-platform, for which Coinbase does not necessarily pay a direct reward to holders.

An important nuance: the 10-Q indicates that Coinbase has already lowered the reward rate alongside the decline in benchmark rates. The risk is therefore not one of reward rigidity, but of elasticity, meaning the extent to which the reward rate can be reduced before higher-yielding alternatives begin to pressure customer retention and balance growth. This competitive dynamic is the true constraint on float business margins, and it is not quantifiable from public data.

Two macro betas, a possible offset

Coinbase now carries two distinct macro exposures:

Beta 1 (crypto activity): transaction revenue, blockchain rewards and certain asset-related fees are driven by crypto asset prices, volumes and volatility. Transaction revenue alone still represented 52% of net revenue in Q2, with the broader crypto-sensitive perimeter extending beyond that figure.

Beta 2 (interest rates): stablecoin revenue, corporate interest income and financing income are driven by the yield environment. This beta is newer, growing and already large enough to move quarterly results.

In theory, the two can offset each other. A rate-cutting cycle compresses the yield on USDC reserves but could simultaneously support crypto asset prices, increase risk appetite, boost trading volumes and grow assets on platform. If this offset materializes, Coinbase's revenue stability would improve through the cycle.

Q2 illustrates the offset's imperfect nature. The decline in average rates pressured stablecoin revenue and corporate interest income year-on-year, while crypto activity deteriorated separately over the quarter, with falling prices, volumes and volatility. This weakness cannot, however, be attributed to rates alone: it also depends on the position within the crypto cycle, sentiment, liquidity, institutional flows, regulation and retail investor behaviour. Coinbase's diversification therefore creates channels that can offset one another in certain configurations, but it guarantees neither an immediate offset nor linear revenue stability.

Based on the published revenue lines, approximately 85% of Q2 total revenue can be attributed to the two main macro channels identified in this analysis: crypto market activity and short-term rates. Under a broader definition that includes all interest and financing income, this proportion reaches approximately 91%.

New growth engines still need to prove their economics

Prediction markets

Prediction markets were one of the highlights of Q2, with contract volume and revenue both more than doubling compared to the prior quarter. Management indicated that annualizing Q2 revenue levels would place the business on a run rate of approximately $100mn, without disclosing the exact quarterly figure. Fees charged to customers are recorded within transaction revenue, primarily within consumer transaction revenue, while amounts paid to prediction-markets partner Kalshi are recorded as transaction expenses. Management indicated a product margin of approximately 50%, likely after these exchange fees rather than after all allocated operating costs.

At this level, the economics appear less attractive than those traditionally associated with retail crypto spot trading, although Coinbase does not publish a directly comparable product-level margin. The company also plans to introduce Combos, which will allow multiple predictions to be bundled into a single trade, and to expand its Prosumer Toolkit, a set of advanced order types and analytics tools aimed at high-frequency traders, to deepen liquidity and engagement. Their impact on take rates and contribution margins is unknown.

Deribit

The acquisition of Deribit, announced in 2025 and completed in August 2025 for $4.3bn, represents a major transaction for Coinbase. The purchase price allocation shows $2.82bn of goodwill and $1.39bn of acquired intangible assets, primarily related to customer relationships, technology and the brand. Institutional transaction revenue rose 65% year on year to $100.1mn, primarily attributed to Deribit's integration. Coinbase does not, however, disclose Deribit's standalone revenue, costs or profitability, which still prevents a precise measurement of the return on invested capital. The strategic logic is nonetheless visible. The next announced milestones include providing US customers with access to global perpetuals (leveraged derivatives with no expiry date) liquidity through a CFTC-regulated pathway, as well as establishing, via Deribit, unified global liquidity for US and international customers. Their execution timeline and financial contribution are unknown.

Base, x402 and Agentic Finance

The Q2 deck presents striking statistics on "agentic" activity, meaning transactions executed autonomously by AI agents rather than by a human placing a trade. According to Artemis data cited by Coinbase, more than 99% of measured onchain agentic commerce was conducted in USDC, more than 90% of agentic stablecoin transaction volumes ran through Base and more than 97% of onchain agentic transactions used x402 during the quarter. Base is Coinbase's own Ethereum layer-2 network, and x402 is an emerging payments protocol that lets software agents pay for resources automatically over the internet. Coinbase thereby positions itself across the full stack of onchain agentic finance, from stablecoin payments through to the settlement network and the communication protocol.

The key caveat concerns the size of the denominator. Artemis defines the scope as autonomous machine-executed transactions via smart contracts on monitored public blockchains, but Coinbase discloses neither their absolute volume nor their current economic significance. The market remains at an early stage and x402 does not yet constitute a distinct revenue line. Coinbase instead plans to monetize this activity through its existing products, notably Base-related and USDC-related revenue. We therefore assign no explicit value to this optionality in our near-term analytical framework, although this positioning could become strategically important if the market reaches meaningful scale.

What investors should watch

The Q3 guidance provides a first partial test of the analytical framework presented above. Management is guiding subscription and services revenue of $500–580mn, or approximately $540mn at midpoint, compared to $555mn in Q2. The increase in average USDC market capitalization and average balances held in Coinbase products is expected to be a tailwind, offset by the non-recurrence of the performance earn-outs recorded in Q2 and by lower average crypto asset prices since the start of the quarter.

On expenses, the $980–1,080mn adjusted expense guidance is broadly flat compared to Q2 at midpoint. The full-quarter effect of the headcount reductions and other efficiency measures taken in Q2 is expected to be offset by higher USDC rewards and marketing efforts behind recent product introductions. The restructuring announced in May reduced headcount by approximately 14%, or roughly 700 positions.

Three issues will determine Coinbase's trajectory over the next twelve months.

Q2 confirms that the two betas do not offset each other automatically or over the same time horizon. Year-on-year, the decline in average rates exerted direct pressure on stablecoin revenue and corporate interest income. At the same time, crypto activity further deteriorated over the course of the quarter, with falling prices, volumes and volatility. This weakness cannot, however, be attributed to rates alone: it is part of a broader cyclical dynamic, visible since the second half of 2025, and also depends on sentiment, liquidity, institutional flows, regulation and retail investor behaviour. The question for the next twelve months is therefore less whether the two betas offset each other mechanically, and more whether a new crypto activity cycle can materialize with sufficient speed and scale to absorb the gradual compression of rate-linked revenue.

How elastic are USDC balances to the reward rate? Coinbase has already lowered the rate paid to users, even though total reward spend has continued to rise due to balance growth. How customers respond to further reductions will determine the extent to which the business can preserve its margins without weakening deposit retention.

Can new products demonstrate their profitability? Prediction markets are scaling at an indicated product margin of approximately 50%. If these economics remain unchanged, the growing share of this business in the revenue mix could dilute consolidated transaction margins relative to the historical crypto spot activity. Coinbase does not disclose Deribit's standalone revenue, costs or profitability, which still prevents an assessment of its contribution to group margins and a precise measurement of its return on invested capital.

Put together, these three questions describe the same underlying tension. Coinbase has traded one concentrated risk for several smaller, less correlated ones, but less correlated is not the same as hedged. Q2 showed both betas moving against the company at once rather than offsetting, and two of the three new growth engines still report through disclosures too thin to let an outside observer judge whether they are actually profitable. The diversification is real. Whether it produces a materially more stable business, rather than simply a more complicated one, is the question the next few quarters will have to answer.