why I had written off Coinbase

I owned Coinbase stock once, in 2022. Snagged it at $340 during the IPO hype, watched it crater to $42 during the bear market, and sold at $85 in early 2023 with a loss that I used to offset capital gains tax on other positions. I told myself the crypto exchange business model was fundamentally broken. Revenue was tied to trading volume, which tanked during bear markets. Regulatory risk was constant and unpredictable. The company had nah moat because users could switch to any of a dozen competing exchanges with a few clicks. Coinbase was a proxy for crypto sentiment, an I wasn't interested in riding sentiment up and down.

I hadnt looked at the stock in over a year when Coinbase released its Q1 2026 earnings on May 7, 2026. A colleague at the hedge fund where I work as a quantitative analyst cited the report during our morning meeting, noting that revenue had come in at $2.34 billion, crushing the consensus estimate of $1.78 billion. I pulled up the SEC filing during my lunch break and read the whole thing from start to finish. It changed my mind about Coinbase altogether.

the morning the report dropped

I remember the exact moment the earnings crossed the wire because my Bloomberg terminal lit up with alerts. Revenue of $2.34 billion. Earnings per share of $4.12, crushing the consensus of $2.89. The stock was up 12% in pre-market within minutes. My colleague Sarah, who covers fintech equities for a different fund, messaged me on Slack saying the numbers looked like a different company from the one id written off two years ago. She wasn't wrong.

the numbers that shocked me

The headline revenue number was impressive on its own. But the composition of that revenue was what honestly caught my attention. Trading revenue was $892 million, up 34% year-over-year, which I anticipated given the crypto rally in Q1 2026. Subscription and services revenue, however, was $1.45 billion, up 92% year-over-year. That was the real story. Coinbase had transformed itself from a transaction-dependent exchange into a diversified financial platform where subscription revenue actually exceeded trading revenue for the first time.

The subscription revenue came from three main streams. First, USDC stablecoin custody fees, where Coinbase earns a spread on the interest earned by Circle's reserves. With USDC market cap sitting at $48 billion an Coinbase gripping a significant custody share, this alone contributed approximately $340 million in Q1. Second, blockchain rewards from staking services across multiple proof-of-stake networks including Ethereum, Solana, and Aptos, totaling about $520 million. Third, Base, their Layer 2 network on Ethereum, which generated $590 million in sequencer fees and settlement revenue during the quarter.

I sat with those numbers for a while, chewing on the implications. The subscription revenue wasnt just growing. It was recurring, predictable, and largely uncorrelated with crypto price volatility. When Bitcoin crashed in January 2026 and trading volume dried up, the subscription revenue maintained flowing. The staking rewards persisted accruing. Base persisted processing transactions. Coinbase had built something that looked more like a SaaS company than a crypto exchange.

comparing the margins to traditional finance

I pulled the financials from the five largest U.S. banks and compared them to Coinbase's Q1 2026 margin profile. JPMorgan's net interest margin was round 2.1%. Goldman Sachs' return on equity was running at 11%. Coinbase's gross margin on subscription revenue was above 70%, which placed it in rarefied air ordinarily reserved for software companies, not financial intermediaries. The comparison wasn't perfect because the business models differ fundamentally, but the direction of travel was undeniable. Coinbase was becoming a high-margin infrastructure play disguised as a crypto exchange.

the Base network was the wild card

I had largely ignored Base when it launched in August 2023, dismissing it as another Layer 2 in an increasingly crowded field that included Arbitrum, Optimism, an a dozen others. By Q1 2026, Base was the second-largest Layer 2 on Ethereum by total value locked, behind only Arbitrum, with $4.8 billion in TVL. But the real metric that mattered was transaction volume. Base processed 2.1 billion transactions in Q1 2026, more than the Ethereum mainnet itself, cuz transaction fees averaged $0.003 compared to Ethereum's $2.40 average during the same period.

Coinbase's role as the sequencer for Base meant they earned the transaction fees from every single transaction on the network. At Base's scale, those fractions of cents added up to $590 million in quarterly revenue. I dug into the on-chain data on Dune Analytics and found that Base's growth was driven primarily by social applications, decentralized social media platforms, and consumer fintech apps that couldn't afford Ethereum mainnet fees but needed Ethereum-level security. The network had developed an ecosystem that was entirely different from the DeFi-focused ecosystems of Arbitrum and Optimism.

The strategic brilliance of Base, which I had fully missed, was that it turned Coinbase's retail user base of 120 million verified accounts into a built-in demand pipeline for an entire blockchain network. Every Coinbase user who wanted to try a dApp on Base could do so with one click from their existing account. Nah bridging, nah fresh wallet, no gas token to acquire. The friction was near zero.

why I decided to buy despite the risks

The bear case on Coinbase is obvious. Regulatory uncertainty persists. The SEC has an ongoing enforcement action from 2023 that could result in penalties or operational restrictions. Competition from decentralized exchanges continues to erode market share for spot trading. And the subscription revenue, while growing, depends heavily on persisted growth in USDC market cap and staking adoption, neither of which is guaranteed. I weighed all of these risks and decided that the margin profile an revenue diversification justified the position size I was planning.

what I did after reading the filing

I snagged COIN on May 8, 2026, the day after earnings. My average entry price was $412 per share. I allocated $24,000, representing approximately 58 shares. My rationale was simple: the market had been pricing Coinbase as a cyclical trading platform when the data displayed it had become a diversified financial infrastructure company with growing recurring revenue. The earnings multiple, which looked expensive at 32x forward earnings at $412, was actually reasonable if you believed the subscription revenue would continue growing at 50%+ annually.

My personal loan balance at the time was $14,000 from a kitchen renovation I'd financed through SoFi at 9.4% APR. I had been planning to use a bonus from work to pay it down. Rather, I decided to keep the cash invested in COIN and let the stock's appreciation, if my thesis played out, cover the loan payoff within 12 months. The risk was higher but the potential reward was significantly larger.

I've watched COIN trade between $380 and $465 over the past ten days. Volatile, as anticipated. But my conviction on the fundamental thesis hasn't wavered. Coinbase built somethin real during the bear market years when everyone including me was laughing at their declining trading volumes. They built the infrastructure for the next phase of crypto adoption, and the earnings report was the proof.

The net income figure was another number that surprised me. Coinbase reported $612 million in net income for Q1 2026, compared to a loss of $79 million in the same quarter of 2025. The swing from loss to profit was driven not just by higher revenue but by dramatically improved unit economics on the subscription side. Staking rewards are pure margin since the infrastructure costs are essentially fixed. Base sequencer fees are nearly pure margin since the operational costs of running a sequencer are minimal relative to the revenue generated at scale. I've been building discounted cash flow models at the hedge fund for six years an the margin structure of Coinbase's subscription business is honestly impressive once you break it apart.

I shared my analysis with three colleagues at work and two of em ended up buying COIN within a week. The third, a guy named Raj who focuses on traditional fintech companies, told me he'd wait for the Q2 report to confirm that the subscription revenue wasn't a one-quarter anomaly. He's presumably right to be cautious. But ive seen enough Q1 earnings reports in my career to know when a company's revenue mix has crossed a structural tipping point, and Coinbase's transition from trading-dependent to subscription-dominated feels real and durable to me.