Call Confirms the Cyclical Reset; 2026 Adjusted-OpEx Framework of $4.3–4.6B is the New Anchor — Maintaining Hold

Initial Read (post-call): The 5:30 PM ET call confirmed the pre-call Hold thesis with high conviction and gave the market the single most actionable disclosure of the print — a 2026 adjusted-OpEx framework of $4.3–4.6B, ~$500M below the Q4 2025 exit run-rate at midpoint, alongside a $50–60M Q2 restructuring charge tied to the headcount reduction announced earlier in the week. Management framed the action as a dual-driver response (market headwinds + AI-native operating transition); CFO Haas disclosed pull requests per engineer up ~80% YoY and integration test coverage 3x in six months. The strategic franchise extended its lead (12 products at $100M+ ARR; retail derivatives $200M+ ARR; prediction markets $100M+ ARR in two months; market-share ATH; USDC platform balance ATH at $19B average). Maintaining Hold, conviction unchanged from the flashcap. The trigger-set to revisit Outperform now narrows to crypto-cycle stabilization plus 2H execution against the new OpEx framework.

Rating Action

This recap maintains the Hold rating we have carried since the Q3 2025 downgrade and reaffirmed at Q4 2025 and in the pre-call Q1 2026 flashcap. The print, the shareholder letter, and the call all resolve in directions the Q4 Hold posture explicitly anticipated; the strategic franchise extended its lead and the operating-cost discipline message landed cleanly, but the cyclical reset and the equity-portfolio optics remain in force. The trigger-set to revisit Outperform — crypto-cycle stabilization, S&S re-acceleration above current Q2 trajectory, and clean GAAP-vs.-Adj. EBITDA convergence — was not present in the print, and on management's framing is unlikely to be present in Q2.

The post-call read narrows the upgrade-trigger set: with the 2026 adjusted-OpEx framework now anchored at $4.3–4.6B and excluding USDC growth implying roughly flat YoY 2025-to-2026 expenses, the operating-leverage delta to a cycle recovery is now more visible than it was pre-call. If crypto market cap and trading volume reset above current levels through 2H 2026, the 2027 EBITDA framework could re-rate fast. The upgrade trigger is therefore better characterized as “cycle confirmation” than as “company execution” — the company has already done its part; the cycle has not yet done its part.

Results vs. Consensus

The print missed materially across both top-line components and printed a sizeable GAAP loss driven by an unrealized crypto-asset mark, on a base of consensus that had assumed both rate-cycle relief and a less severe spot-volume contraction. None of the gap is operating-deterioration in nature; all of it is cycle and mark-to-market.

Segment Performance

Transaction Revenue — The Cyclical Pillar

• Total: $755.8M. Tracked the market-wide spot-volume contraction (global exchange volume -48% peak-to-trough); Coinbase-specific share gain (8.6% ATH) means the firm took more of a smaller pie.
• Consumer: $567M, -23% QoQ. Outperformed the -35% decline in overall consumer spot volumes thanks to a mix-shift toward consumer core trading away from advanced (higher take-rate venue), and accelerating contributions from derivatives and prediction markets — both of which contribute to total revenue but are not included in the spot-only trading-volume key business metric.
• Institutional: $136M, -27% QoQ. Choi attributed the gap to lower volatility reducing hedging demand, options activity declining following ATH Q4 volumes (Deribit-disproportionate weighting), and macro institutional softness. Counter-signal: institutional engagement was strong by end-of-quarter (active lending clients +double-digits QoQ; average daily loan balances at $1.4B ATH; 45 major financial institutions moved tokenization from concept to production in Q1).
• Strategic offset: Both the US and international derivatives exchanges hit new ATH revenue contribution in Q1 within institutional. Institutional derivatives revenue more-than-offset the option-activity decline in the quarter.

Subscription & Services — The Recurring Chassis

• Total: $584M, -16% QoQ. Within Coinbase's own Q4-issued guide ($550–630M); Street had modeled near the high end. Continued strength in native-unit inflows (12th consecutive quarter); growth offset by prices and protocol-reward rates.
• Stablecoin revenue: $305M. USDC average platform balance hit a new ATH at $19B. Haas reiterated the structural feature that the USDC contract auto-renews every three years into perpetuity and cannot be terminated. A reporting change reclassified $18M of corporate stablecoin revenue to other revenue (treating cash and USDC as fungible within corporate treasury); historical periods recast.
• Blockchain rewards: $101M. Lower on price and protocol reward rates; native-unit growth in staked balances offset.
• Interest & finance fees: $68M (+13% QoQ). Lending growth working through the cycle.
• Coinbase One: 1M+ paid subscribers. Materially crossed the 1M threshold this quarter. Members generate incrementally higher trading volume, higher revenue, and exhibit strong unit economics — the closest thing to a HOOD-Gold-style annuity in the COIN portfolio.

Revenue Diversification — Twelve Products at $100M+

Haas disclosed the Q1 product-line scoreboard: 12 products generating >$100M in annualized revenue, with retail derivatives at >$200M ARR (next product on track to hit a $250M+ year) and prediction markets at $100M+ ARR in its second month of meaningful operations (on track to be the 13th $100M+ product). Non-crypto futures contracts (silver, gold, oil) +4x QoQ on volume. The platform-diversification thesis has crossed from narrative to a demonstrable revenue-line portfolio; this is the structural reason the rating remains Hold rather than Underperform.

Key Topics & Management Commentary

The 2026 Adjusted-OpEx Framework — The Most Consequential Disclosure

The single highest-impact data point on the call was Haas's introduction of an annual adjusted-expense outlook on top of the quarterly framework:

“We expect 2026 adjusted expenses to be between $4.3 billion and $4.6 billion. This is roughly $500 million lower than our Q4 2025 annualized exit-rate at the midpoint. And I also want to point out that absent any growth in USDC rewards, we would expect 2026 expenses to be flat to 2025.”
— Alesia Haas, CFO

The walk has three legs: (1) the headcount reduction announced earlier in the week, generating the bulk of the unit-cost compression; (2) the AI-native transition, with Haas citing pull requests per engineer up almost 80% YoY and integration test coverage up 3x in six months; (3) declines already in motion in deal-related legal costs, customer support, and policy-related expenses. The Q2 2026 T&D + G&A guide of $820–870M (-4% to -9% QoQ) is the explicit down-payment, with an additional $50–60M Q2 restructuring expense as a standalone line.

Read-through: COIN has just signaled that operating leverage in a 2H 2026 cycle recovery would be considerable. If revenue rebounds 15–20% from the Q1 base on a cycle inflection, the unchanged-OpEx framework means EBITDA flow-through is not diluted by management's natural urge to reinvest. This is the right cost-discipline message at the right cyclical moment; it doesn't change the rating today, but it materially increases the slope of any future upgrade if cycle conditions improve.

Strategic Strategy — Three 2026 Priorities

Armstrong organized the call around three explicit 2026 priorities, all tracking ahead of the framework set at the Q4 2025 / early-2026 strategy session:

• The Everything Exchange. Coinbase has transformed from a primarily spot-focused crypto platform into a multi-asset venue spanning crypto, stocks, 24/7 equity perps, retail and geographic-expanded derivatives, prediction markets, and non-crypto contracts (silver, gold, oil — +4x QoQ volume). Retail derivatives ARR >$200M; prediction markets ARR $100M+ in two months.
• Stablecoins and payments. Total stablecoin supply has doubled over two years, with USDC taking growing share of a growing pie. Stablecoin transaction volume on COIN doubled QoQ; USDC and partner stablecoins drove 80%+ of total volume. Base is now the dominant chain for all stablecoin transactions at 62% share. USDC is used in 99% of agentic stablecoin transactions; 90%+ of those happen on Base.
• Bringing trading and payments onchain. DEX volumes on the Coinbase app +2x QoQ; borrow/lend balances now $1B+ (up from inception over the past year).

Choi's framing on the share-gain durability was the cleanest operating-philosophy data point of the call:

“When conditions are difficult, people go to where they trust. So this is the 12th consecutive quarter of net native unit inflows for us. Share gains have been driven by-product innovation and expansion of our derivatives platform... We do also believe that share captured in down markets will be sticky as conditions improve.”
— Emilie Choi, President & COO

Deribit Integration & US Crypto Options — The Multi-Quarter Roadmap

Haas's response to the US-options-launch question was the single most useful forward-product disclosure of the call:

“We closed the Deribit transaction last year. Deribit was the clear leader in terms of institutional clients and professional market makers and trading options. We are very focused on this integration right now. It is progressing nicely and we expect to be fully-integrated in 2026. This means that we're going to unify spot, perps, futures, options all on a single platform.”
— Alesia Haas, CFO

On the US specifically, Haas declined to give a launch timeline (“cannot give you a timeline on today's call, but we're actively working on it and very optimistic. So coming soon”). Both US and international derivatives exchanges hit new ATH revenue contribution in the quarter. The institutional-derivatives line more-than-offset Q1 option-activity declines, providing real evidence that the Deribit acquisition is delivering on the institutional-density thesis.

The Circle Relationship — Reframed and Insulated

Two related disclosures cleaned up the framing the flashcap had flagged. Grewal addressed the question of whether stablecoin-rewards legislation could affect the Circle revenue-share contract:

“Fortunately, the contracts that we have in-place in Circle are set. And as Alicia has underscored, they auto-renew. So we expect to continue to go-forward with our relationship with Circle under those same terms.”
— Paul Grewal, Chief Legal Officer

“The revenue-share is tied to overall USDC supply and adoption and it's really unaffected by any rewards language.”
— Alesia Haas, CFO

Net of the rewards-legislation handicap, the contract structure is unchanged; the Q1 stablecoin revenue line of $305M is the same per-dollar-of-supply economics as prior quarters, just on a lower per-dollar yield (reserve income compression). The reframing doesn't address the Q1 $482M crypto-asset mark, but it does insulate the forward Circle revenue stream from the legislative process.

Clarity Act & Regulatory Setup — Markup This Month, Signed by End-of-Summer

Grewal opened the Q&A with a multi-paragraph framing on the Clarity Act timeline:

“On clarity, we are confident that the bill is going to head to markup this month with a floor vote to follow in early summer. All that translates to our confidence that we're going to see a signed piece of legislation by the end-of-the summer.”
— Paul Grewal, Chief Legal Officer

On the Tillis-Lummis stablecoin-rewards compromise specifically: “The direction of the text and in particular, its preservation of activity-based rewards while prohibiting a passive pure bank style deposit style yield really reflects what to us is an approach that can work and will work going-forward... rewards are going to be protected and we can preserve what are the key elements of our current program.” Armstrong's read-through: post-Clarity, hundreds of US companies are likely to integrate crypto similar to the post-GENIUS-Act stablecoin integration wave, and Coinbase Developer Platform is positioned as the integration rail.

The Restructuring & AI-Native Transition — Why Now

Asked to disaggregate market-driven vs. AI-driven motivations for the headcount reduction, Haas declined to split the attribution but anchored on the AI-leverage data:

“The restructuring reflects two forces acting simultaneously. It wasn't all one, it's not all the other and it's hard to just detangle and say what is more or less. We definitely saw market headwinds and we've definitely also seen a transition to AI native operations. So as we shared, pull requests are up by Engineer by 78% year-over-year. We are seeing continued growth in that and I think that we're going to only see more-and-more of our work being done by AI in all of our functions over-time.”
— Alesia Haas, CFO

Armstrong addressed the operational-risk side directly, acknowledging that non-technical employees increasingly draft code via AI agents but that human engineers still review all production code (with multiple review levels on sensitive systems). His framing positioned AI agents as both a velocity unlock and a quality / cybersecurity uplift, not just a labor-substitution lever.

Fee Sensitivity & Competitive Positioning — “Not Choosing Us Because We're Cheapest”

Asked about Morgan Stanley and traditional brokerage fee competition, Haas was direct:

“Our clients are not choosing us because of we're the cheapest... They're choosing us today because we're the most trusted, we're the easiest to use, the most crypto stored. We have 80 licenses, we have a global regulatory foundation... over the long-term, we've always said that we believe that fees could come down as things become commoditized. And so our focus on diversifying our revenue is very important.”
— Alesia Haas, CFO

Armstrong supplemented that fee-sensitive customers already have on-platform options (Coinbase One zero-fee trading; Coinbase Advanced volume-discounted to single-digit bps at the high end). Net: management is signaling no panic on fee compression today, but explicit acknowledgement that the structural answer is product-portfolio diversification, not fee defense.

Capital Allocation & Buyback Discipline

Haas walked the capital position cleanly: $10B+ cash & equivalents, $12B total available resources, $1.1B of buybacks in Q1 (~6M shares). Cumulative repurchases since Q4 2024 have offset ~90% of stock-comp issuance — the cleanest stock-comp-offset signal in COIN's history as a public company. The $1.3B June-2026 convertible note will be retired at par unless the conversion price is reached. The flexibility framing — invest through-cycle, return capital, pursue strategic opportunities, all simultaneously — is the right capital-allocation philosophy at the right point in the cycle.

Agentic AI & X402 — The Long-Dated Optionality

On x402 (the open agentic-payments protocol incubated within COIN, now under Linux Foundation governance with Cloudflare, AWS, Stripe, Shopify, and Google as contributors), Armstrong's commercial framing was clean: 99% of x402 transactions settled in USDC in Q1 (monetized via Circle); 90% of stablecoin transaction volume on Base in Q1 (Base is the leading chain). No revenue line yet, but the protocol-traction story underwrites the Base/CDP commercial flywheel.

Guidance & Outlook

Read: The forward signal is unambiguously cautious on near-term revenue (Q2 S&S $605M mid is below Street; rate cycle + staking compression continues) but constructively disciplined on cost (2026 adjusted-OpEx flat YoY ex-USDC rewards). The asymmetry is real: revenue is recovering with the cycle (so management can't promise); costs are within management's control (so management is promising). For a Hold thesis, this is the right asymmetry — downside is bounded because management is taking action; upside requires a cycle that hasn't yet arrived.

Analyst Q&A — Notable Exchanges

The Q&A format was a written-question rotation moderated by CBO/Head of IR Shan Aggarwal in his first call as the IR lead. Notable threads:

• James Yaro (Goldman Sachs) opened with the Clarity Act question, drawing Grewal's clearest disclosure on the legislative timeline (markup this month; signed by end-of-summer) and the rewards-language compromise framing.
• Ken Worthington (JPMorgan) asked Armstrong about post-Clarity ecosystem participation, drawing the GENIUS-Act-analogy framing and the Coinbase Developer Platform integration-rail thesis — the most thesis-relevant strategic answer of the call.
• Paul Christiansen (Citi) drew the Circle-revenue-share question, prompting both Grewal's contractual-auto-renewal reaffirmation and Haas's clarification that the revenue share is tied to USDC supply and adoption, not to rewards language.
• An ex-analyst from Architect 9000 (X) asked Armstrong about non-technical AI-coded production deployments — drawing the “product managers and designers draft, human engineers still review all production code” clarification.
• Ramsey (Cantor Fitzgerald) asked Choi about competitive environment and share-gain drivers; Choi delivered the ATH market-share data, the +5x share-since-Q1-2023 anchor, and the “share captured in down markets will be sticky as conditions improve” framing.
• Andrew Jeffrey (William Blair) asked Choi about stablecoin movement infrastructure ambition (CPN participant or platform); Choi drew the cleanest answer of the call: “We're not playing as a network participant. We are the platform that powers stablecoins.”
• Rayna Kumar (Oppenheimer) asked Armstrong about x402 commercialization and incremental USDC growth; Armstrong delivered the 99%-USDC-settlement, 90%-Base-share, Linux-Foundation-governance, and CDP-integration framing.
• Patrick Moley (Piper Sandler) pressed Haas on the Everything Exchange monetization timeline; Haas declined to give a per-product outlook but anchored on retail derivatives ($200M+ ARR) and prediction markets ($100M+ ARR in two months) as the visible-traction proofs.
• Owen Lau (Clear Street) asked Haas about US crypto options launch timing; Haas confirmed Deribit integration is “progressing nicely” and expected to be fully integrated in 2026 with US options “coming soon,” without a specific timeline.
• Devin Ryan (Citizens) asked Armstrong about the spec-vs.-utility cycle phase; Armstrong's framing — spot crypto down a bit, derivatives + prediction markets + commodities futures up — is the cleanest framing of the diversification thesis the platform was built for.
• Alex Markgraff (KBCM) asked Haas to disaggregate the RIF (market vs. AI leverage); Haas declined to split but anchored the $500M total cost reduction vs. Q4 25 run-rate and reaffirmed the 2026 framework.
• President Noble (X) asked Haas about fee competition vs. Morgan Stanley and other brokerages; Haas drew the “not choosing us because cheapest” positioning, with Armstrong supplementing on Coinbase One and Coinbase Advanced fee tiers.
• John Tedaro (Needham & Co.) asked Choi about institutional weakness vs. retail; Choi attributed it to lower volatility / hedging demand, options-activity decline post Q4 ATH, and tokenization adoption (45 institutions moved tokenization concept-to-production in Q1).
• Credit Brian (X) closed with the “most excited for next 1–3 years” question, drawing Armstrong's three-pillar framing: every asset class onchain (tokenization $30B today → $16T by 2030), stablecoin payment rails, agentic commerce.

What They're NOT Saying

• No April month-to-date transaction-revenue color. The flashcap flagged April MTD as one of the five rating-decision triggers; the call did not provide MTD data. Choi's institutional comment that “most of the downtrend happened in January” is the closest thing to a constructive April hint — but it's an institutional comment, and HOOD-style “highest month of the year” framing on consumer transaction revenue was not offered.
• No quantification of the Q1 take-rate. The Q3 2025 take-rate compression carry-over watch from the Q4 recap remains unresolved. Press release didn't disclose; call didn't address; we'll need the 10-Q.
• No specific forward-revenue or Adj. EBITDA guide. Coinbase's house position not to guide top-line or bottom-line continues to hold; Q2 S&S is the closest thing to a forward-revenue signal we get, and even that is one line of total revenue.
• No reframing of the equity-portfolio overlay disclosure. The flashcap flagged the dual-narrative problem (operating exchange + balance-sheet portfolio of BTC accumulation + Circle stake) as the structural issue producing GAAP-loss optics for the third consecutive quarter. Management did not address it, did not separate Adj. EBITDA disclosure from equity-stake marks, and did not guide the BTC-accumulation cadence forward. The optics issue persists.
• No segmentation of prediction-markets unit economics. Haas explicitly declined to give per-product economics. Choi's framing of prediction markets as “incremental revenue cross-sold into a customer base we've already acquired” is favorable-contribution-margin language without a number. The model needs more here in 1–2 quarters.
• No commentary on Coinbase ETF / index inclusion overhang. The S&P 500 inclusion debate has been a Street-side undercurrent since Q4 2025; not addressed by management on this call.
• No specific 2H 2026 / 2027 framework beyond the OpEx anchor. The 2026 adjusted-OpEx of $4.3–4.6B is the only multi-quarter forward number we received; revenue trajectory is implicit.

Market Reaction

The print landed with the press release at ~5:00 PM ET; the call ran 5:30–6:30 PM ET. Initial after-hours reaction was negative, with the stock down ~4% in the immediate post-print window on the GAAP-loss optics, the Q2 S&S guide below Street, and the consensus EPS swing from ~$0.29 expected to $(1.49) actual. The call's 2026 adjusted-OpEx framework disclosure, the strategic-franchise narrative (12 products at $100M+ ARR; market-share ATH; prediction markets at $100M+ ARR in two months), and the Circle-contract reaffirmation appear to have moderated the after-hours pressure modestly through the call window.

The reaction is consistent with a market that prices the cyclical reset cleanly (Q2 S&S below Street; +30% YoY revenue decline) but credits the strategic-franchise progression and the unit-cost discipline message. The post-call price tape is a Hold-rating-confirming reaction in our read — not a thesis-breaking move down, not a beat-and-raise rally. The next material price catalysts are: (1) the 10-Q filing (take-rate disclosure; segment-level operating expense detail), (2) the May / June crypto-cycle tape, (3) the Clarity Act markup process, (4) the June-2026 convertible retirement.

Street Perspective

The bull case being made on the Street post-print converges on three planks: (1) the 2026 adjusted-OpEx framework of $4.3–4.6B is a credible operating-leverage setup if the cycle inflects, with revenue elasticity to 2H 2026 implying a re-rating of 2027 EBITDA potential; (2) the strategic franchise has compounded under cyclical stress (market-share ATH, 12 products at $100M+, prediction markets fastest-scaling-ever, Base dominance in stablecoin and agentic rails), giving COIN a higher exit-multiple than the cyclical print suggests; (3) capital position ($12B total resources, $1.1B Q1 buyback, ~90% stock-comp offset since Q4 2024) supports through-cycle compounding regardless of the cycle's pacing.

The bear case being articulated centers on: (1) GAAP-loss optics from the equity-portfolio overlay are now three quarters running, with no reframing in sight; (2) Q2 S&S guide implies an annualized $2.3B run-rate vs. ~$2.8B for 2025, a ~17% reset that may not be fully recovered even with cycle improvement given rate-cycle and staking-yield drag; (3) the headcount RIF is a defensive signal as much as an offensive AI-leverage signal, and there's a real question about cultural / talent-density costs from a sub-quarter restructuring; (4) prediction-markets unit economics are still undisclosed, leaving the consensus model carrying a $100M+ ARR product as essentially a margin black box.

Our read sides with the bull framing on (1) and (3) and treats the bear framing on (2) and (4) as appropriately scoped near-term concerns. The decisive cycle-confirmation trigger remains external to the company and unresolved.

Model Implications

• FY26 revenue: Q1 + Q2 guide-midpoint anchor implies ~$5.4–5.6B FY26 total revenue (Q1 actual + ~$1.3–1.4B/quarter run-rate for Q2–Q4 absent cycle inflection). We mark the trajectory consistent with $5.5B base-case, $5.0B downside, $6.5B+ upside on a 2H cycle inflection.
• FY26 adjusted operating expense: Anchor at the $4.45B framework midpoint, ex-USDC-rewards growth flat YoY. Operating leverage in a cycle recovery is the principal model upside lever.
• FY26 GAAP earnings: Highly contingent on the equity-portfolio overlay (BTC accumulation P&L, Circle stake mark). Adj. EBITDA is the more useful signal — we underwrite $1.0–1.4B FY26 Adj. EBITDA in a no-recovery scenario, $1.8–2.2B with cycle improvement.
• S&S trajectory: Q2 mid $605M is the new bottom-up anchor. We do not model 2H S&S re-acceleration above ~$625M absent rate-cycle pause AND USDC float growth offsetting per-dollar yield compression.
• Capital return: We model continuing buyback pacing of ~$800M–$1.0B/quarter, with the $1.3B June convertible retirement absorbing some flexibility in Q2.
• Strategic optionality: Prediction markets and Deribit US-options launch are the principal model-upside levers in 2H 2026. Neither is in our base-case revenue contribution beyond stated ARR figures.

Thesis Scorecard

Overall: The thesis arc is reinforced, not disrupted. Six of eight bull pillars confirmed (one extended deeper, three confirmed cleanly, one compressed-but-durable, one new on the 12-products-at-$100M+ disclosure, one new on the cost-framework anchor). All three previously-active bear points confirmed; one prior bear (take-rate) remains pending the 10-Q; one new bear (restructuring as defensive signal) is added with active-watch status. The franchise extended its competitive position; the cyclical and equity-portfolio drags persist exactly as the Q4 Hold thesis predicted.

Action: Maintaining Hold. The Q4 caution was the right call; tonight's print is its literal materialization. The 2026 OpEx framework is the cleanest piece of new disclosure, materially increasing the operating-leverage slope into any future cycle inflection. We will reassess the rating at the next-quarter print, with the trigger-set narrowing to: (a) cycle confirmation (BTC/ETH price stabilization above current levels, global volume re-acceleration), (b) S&S re-acceleration above $625M-quarterly run-rate, (c) prediction-markets unit-economics disclosure, and (d) progress against the 2026 OpEx framework as a 2H execution proof point.

Net: Maintaining Hold (from Hold). Recap confirms the pre-call flashcap thesis. Cyclical reset is real and acknowledged; strategic franchise extended its lead; cost discipline message landed cleanly. The decisive upgrade trigger is now cycle-external, not company-internal — which is exactly where a well-executing Hold should be.