CRCL · Circle Internet Group

Financial - Capital Markets · mkt cap $27.0B · calls: Q1 FY2026 vs Q4 FY2025

49.0 conviction · conf-adj 49

conf 4/10 partial

enthusiasm:30.0 · trend:8 · quantifies:0 · impact:0 · under_radar:0 · credibility:0 · business_impact:8 · disruption:0 · commitment:0 · confirmation:3

Enthusiasm latest 10 / prev 9 (rising)

Circle’s AI thesis is that agentic software will need programmable, low-cost, trusted payment rails, making USDC, Arc, Circle Gateway and the Circle Agent stack core infrastructure for machine-to-machine economic activity. Enthusiasm rose because Q1 moved from Q4’s testnet and vision language to launched products, market share claims in X402 payments, and quantified internal AI adoption. The credibility is strongest on product readiness and usage/adoption metrics, but management still does not quantify AI-driven revenue, margin expansion or cost savings directly.

GROUNDED NEXT-FY IMPACT vs CONSENSUS

Grounded on actual base — revenue $2.7B · net income $-0.1B · net margin -2.5% · diluted EPS -0.29

These are next-fiscal-year annual uplift estimates, not next-quarter numbers.

Aggregate next-FY est. rev uplift: % · next-FY EPS uplift: % · vs analysts: unclear · priced in: high · confidence: 4/10

Claim	Figure	Arithmetic
99.8% of X402 agentic payments settle in USDC engagement · soft	99.8%	Market-share/settlement metric, not a $ figure. Revenue math would be X402 payment volume * Circle take-rate / $2.746B, but volume and take-rate are undisclosed. No anchor -> null.
Agent marketplace with 500+ endpoints engagement · soft	over 500 endpoints	Endpoint count is product/coverage traction with no transaction volume, fee, or take-rate. Cannot size against $2.746B. null.
~85% of employees weekly-active AI-tool users productivity · soft	approximately 85%	Adoption rate, not a dollar saving or FTE-reduction. after_tax_saving = saving_$*(1-0.21) cannot be computed (no saving_$). Net income negative so eps_uplift meaningless anyway. null.
600+ AI-native apps deployed internally productivity · soft	over 600 apps	App count, no quantified opex saving or capacity gain. Cannot translate to $; EPS uplift also meaningless on negative net income (-$69.5M). null.
Agent Nano payments down to 1-millionth of a penny other · soft	1 millionth of a penny	Min-transaction spec for micropayments; no transaction count, average value, spread, or take-rate. Cannot compute revenue. null.
Circle Gateway agent tx cost $0.00001 other · soft	$0.00001	Unit-cost product spec (Testnet), not incremental revenue or saving; no transaction count. null.
Circle Gateway moves value in <1 second other · soft	less than a second	Latency/product capability, not a quantified financial claim. null.
Tens/hundreds of billions of AI agents will transact other · soft	tens-hundreds of billions of agents	Unanchored TAM vision. No dollar value, timeframe, Circle penetration, or take. Method says set null, do not invent. null.

Assumptions: Tax rate 21%; incremental margin moot (no revenue figure to flow). Revenue base $2.746B (>$1B). Net income is negative (-$69.5M, EPS -$0.29), so per the loss-making guardrail every per-claim and aggregate eps_uplift_pct is set to null. Phasing: N/A — no multi-year dollar claim to phase into next FY. All claims tagged adopter (Circle applying AI to its own USDC network + internal engineering; agentic-payment share is adoption of Circle's own product, not sale of AI compute), so there is no supplier-side revenue to separate.

Top line: Un-sizable from disclosure. Every topline AI claim is an operational or product-spec metric — 99.8% USDC settlement share of X402 agentic payments, 500+ marketplace endpoints, $0.00001 Gateway tx cost, sub-second settlement, 1-millionth-of-a-penny nano payments, and a 'tens/hundreds of billions of agents' TAM vision. None is tied to a dollar of incremental revenue, monetized volume, or take-rate, so none can be multiplied against the $2.746B base. The agentic-payments narrative is strategically real (USDC is the default settlement rail) but management supplied no figure to convert it into a revenue uplift %.

Bottom line: Also un-sizable, and doubly so. The two efficiency claims — ~85% weekly-active AI-tool adoption and 600+ internal AI-native apps — describe adoption breadth, not a dollar opex saving or FTE reduction, so after_tax_saving = saving_$*(1-0.21) cannot be computed (no saving_$). The opex base is ~$329M but no productivity rate or saving was given. Separately the company is loss-making (net income -$69.5M), so any EPS-uplift % is meaningless against a negative base — est_eps_uplift_pct is null per guardrail.

[impact n/m (all claims soft/unanchored); EPS uplift n/m (loss-making base)] No hard AI figure points above consensus, and consensus already embeds aggressive AI-flavored optimism: revenue $2.723B (FY25e) -> $3.074B (FY26e, +12.9%) -> $4.104B (FY27e, +33.5%), with net income swinging from -$368M to +$218M to +$430M and EPS from -$1.35 to +$0.90 to +$1.70. That ~$287M FY26 net-income swing on only +12.9% revenue growth implies large operating-leverage/efficiency gains — exactly the AI-driven productivity story management is telling, but already in the numbers. With zero incremental-revenue or cost-saving dollars disclosed, no math lands above this trajectory, so the AI optimism is best treated as already priced in.

MODEL CONSENSUS (impact)

partial

Both agree all 8 claims are unsizable and EPS uplift is null on a negative base; differences only in priced_in and confidence, resolved toward the more conservative read.

Conflicts reconciled

priced_in: X=medium vs Y=high -> used high because Y shows the FY26 net-income swing already embeds the AI efficiency story and no disclosed figure exceeds consensus
confidence: X=6 vs Y=3 -> used 4, below the 4.5 average, because all claims are unanchored/null and the verdict conflict warrants conservatism

Field	Opus 4.8	GPT-5.5
Rev uplift %	–	–
EPS uplift %	–	–
Priced in	high	medium
vs analysts	unclear	unclear
Confidence	3	6
Top line	Un-sizable from disclosure. Every topline AI claim is an operational or product-spec metric — 99.8% USDC settlement share of X402 agentic payments, 500+ marketplace endpoints, $0.00001 Gateway tx cost, sub-second settlement, 1-millionth-of-a-penny nano payments, and a 'tens/hundreds of billions of agents' TAM vision. None is tied to a dollar of incremental revenue, transaction volume Circle monetizes, or a take-rate, so none can be multiplied against the $2.746B base. The agentic-payments narrative is strategically real (USDC is the default settlement rail) but management has supplied no figure to convert it into a revenue uplift %.	Topline claims show strong positioning signals, including 99.8% X402 settlement share, 500+ marketplace endpoints, $0.00001 Gateway transaction cost, and sub-second movement, but none provides payment volume, take-rate, spread, or next-FY phasing. Therefore no defensible revenue uplift can be calculated against the $2.747B revenue base.
Bottom line	Also un-sizable, and doubly so. The two efficiency claims — ~85% weekly-active AI-tool adoption and 600+ internal AI-native apps — describe adoption breadth, not a dollar opex saving or FTE reduction, so after_tax_saving = saving_$*(1-0.21) cannot be computed (no saving_$). Separately, the company is loss-making (net income −$69.5M, EPS −$0.29), so any EPS-uplift % would be meaningless against a negative base — est_eps_uplift_pct is null per guardrail.	Bottomline claims show broad internal adoption, including 85% weekly active employee use and 600+ AI native apps. The opex base is $329.255M, but management gave no productivity rate or dollar saving; EPS uplift is also not meaningful because current net income is negative at -$69.508M.
Reasoning	No hard AI figure points above consensus, and consensus already embeds an aggressive AI-flavored optimism: revenue $2.723B (FY25e) -> $3.074B (FY26e, +12.9%) -> $4.104B (FY27e, +33.5%), with EPS swinging from −$1.35 to +$0.90 to +$1.70 and net income from −$368M to +$218M to +$430M. That ~$287M FY26 net-income swing on only +12.9% revenue growth implies large operating-leverage/efficiency gains — exactly the kind of AI-driven productivity story management is telling, but already in the numbers. With zero incremental-revenue or cost-saving dollars disclosed, there is no math that lands above this trajectory, so the AI optimism is best treated as already priced in rather than an unrecognized upside.	Consensus revenue rises from $2.723B in 2025 to $3.074B in 2026, a $351.374M increase or 12.90%, then to $4.104B in 2027, another $1.030B or 33.51%. Consensus net income moves from -$368.220M in 2025 to $218.145M in 2026 and $430.431M in 2027. The AI claims contain no computable next-FY incremental revenue or savings, so there is no hard math showing upside beyond that trajectory.

Rows highlighted where the two models disagreed.

QUANTIFICATIONS

X402-Agentic payments settled using USDC: 99.8% (today / Q1 FY2026 call, topline)
“USDC already has an enormous lead in Angentic payments today, with 99.8% of all X402-Agentic payments being settled using USDC.”

agent marketplace endpoints: over 500 end points (Q1 FY2026 call, topline)
“We also launched the first version of our agent marketplace, an open hub for users and agents to discover, pay for and invoke agent services that transact in USDC programmatically with over 500 end points already available for agents.”

AI tool weekly active employee adoption: approximately 85% of employees (Q1 FY2026 call, bottomline)
“Notably, we have seen rapid uptake of AI coding tools, weekly active users of AI tools, building automations at Circle have rapidly grown to approximately 85% of employees.”

AI native apps deployed internally: over 600 AI native apps (this year / Q1 FY2026 call, bottomline)
“Our teams have already deployed over 600 AI native apps this year, as we have capitalized on major breakthroughs in agentic and AI development.”

Agent Nano payments minimum transaction size: as small as 1 million of a penny (Q1 FY2026 call, topline)
“We also brought Agent Nano payments online, where USDC transactions as small as 1 million of a penny enable high-frequency machine-to-machine payments.”

Circle Gateway agent transaction cost: [ $0.00001 ] (Q4 FY2025 call, topline)
“In fact, we just went into Testnet release of a new capability with Circle Gateway that allows for agents to autonomously and programmatically automate cross chain USDC transactions with a transaction cost of [ $0.00001 ].”

Circle Gateway agent value movement speed: less than a second (Q4 FY2025 call, topline)
“Circle Gateway, which is a feature that would allow autonomous agents to hold a balance and spend not just on Arc, but on other networks and have a transaction cost of [ $0.00001 ] and get that value moved in less than a second to all these other apps and services that are out on these networks.”

potential AI agent population: tens or hundreds of billions of AI agents (future / Q4 FY2025 call, topline)
“We are entering a world where, in my view, likely tens or hundreds of billions of AI agents will interact and perform economic functions over the Internet.”

PAST (realized)

We have seen rapid uptake of AI coding tools, weekly active users of AI tools, building automations at Circle have rapidly grown to approximately 85% of employees.
Our teams have already deployed over 600 AI native apps this year, as we have capitalized on major breakthroughs in agentic and AI development.
we just went into Testnet release of a new capability with Circle Gateway that allows for agents to autonomously and programmatically automate cross chain USDC transactions with a transaction cost of [ $0.00001 ].

CURRENT (now)

USDC already has an enormous lead in Angentic payments today, with 99.8% of all X402-Agentic payments being settled using USDC.
Our adoption of AI is not just outward looking. We are undertaking a significant internal transformation and AI adoption inside Circle is accelerating.
Now inside Circle, AI is also becoming foundational infrastructure across all of our functions.

FORWARD (guidance)

We are rapidly moving into a world in which AI-powered software machines, coordinating on blockchain computers, deliver an increasing share of global economic activity.
We believe that the dominant form of transactions in the identic economy will be conducted on these new economic operating systems and networks and that well-regulated digital cash instruments like USDC will be the preferred form of payment.
We are entering a world where, in my view, likely tens or hundreds of billions of AI agents will interact and perform economic functions over the Internet.
Our product velocity is accelerating and I anticipate that to continue alongside the exponential improvements we're seeing from AI coding agents.

TRACK RECORD — PROMISE vs DELIVERY

—/100 (no quantified promises) no-quantified-promises 3 calls reviewed

Across these three calls Circle's AI content is genuinely substantive (agentic payments via Circle Gateway, the Circle Agent stack of agent wallets/nano-payments/marketplace, the X402 standard, ~85% internal AI-tool adoption, 600+ internal AI apps, AI-aligned Arc infrastructure), but it is all delivered-and-reported current-state metrics or forward vision — earlier calls never set a prior quantified AI target (a number plus a timeframe/milestone) that a later call could audit. The one dated milestone, Arc mainnet in 2026, is blockchain infrastructure rather than an AI capability and remains too-early, so there is no judgeable quantified AI promise-vs-delivery record yet.

PRICED-IN (REFINED)

HIGH (already in)

Est. revisions rising · Fwd P/E -74.6 · EV/Sales 8.9x

AI claim maps to Subscription and Services, Transaction Revenue, Other Services

Analyst ratings show upward migration since January with buys rising and sells lower, while last-month price targets are above last-quarter and last-year averages. Forward estimates also bake in strong revenue growth and a swing from negative EPS in FY2025 to profitability in FY2026-FY2027. The forward P/E is not meaningful because next-FY EPS is negative, but EV/Sales near 8.9x and P/S near 9.4x are rich for a capital-markets business, so rising estimates plus a premium sales multiple indicate AI-related upside is already substantially reflected.

COVERAGE — ENTHUSIASM TRAJECTORY + CATALYSTS

2Q3 FY20254Q4 FY20257Q1 FY2026

AI enthusiasm across 3 calls — trend ↗ rising

AI moved from vague agentic-economy framing to a named Circle Agent stack tied to platform strategy and product launches.

RECENT AI CATALYSTS & NEWS

2026-05-11 · My Top 2 Recent Tech IPOs to Buy for May 2026
2026-05-11 · CRCL's USDC: The Currency of the AI Economy
2026-05-11 · Circle Internet Group Q1 Earnings Call Highlights
2026-05-11 · Circle Chases Agentic Growth to Scale Stablecoin Infrastructure
2026-05-11 · Circle Surges On Results, Stablecoin Growth. These Crypto, AI Stocks Fall.
2026-05-11 · Circle Launches AI Infrastructure to Power the Agentic Economy

BUSINESS IMPACT - QUALITATIVE MATERIALITY

7/10 qualitative impact material medium-term · mixed evidence

Where AI matters: USDC agentic payments and internal productivity

Circle has real AI-linked product deployment: agent wallets, nano-payments, Gateway, marketplace endpoints, and claimed 99.8% USDC share of X402 agentic payments. The upside could expand USDC usage if machine-to-machine payments scale, but disclosed metrics are mostly product/adoption signals with no AI-driven revenue, balances, take-rate, or cost-savings dollars.

Caveats: Agentic payment volumes may be tiny despite high share metrics; No disclosed AI revenue uplift, take-rate, reserve-balance impact, or opex savings; Stablecoin competition from banks, fintechs, and other issuers could compress economics; Regulatory or compliance failures around autonomous wallets could slow adoption

AI DISRUPTION / CANNIBALIZATION RISK tailwind · 2/10

AI does not directly automate away Circle's core model of regulated stablecoin issuance, settlement infrastructure, and reserve-linked economics; it more plausibly increases transaction demand. The main threat is indirect: AI-native payment standards or competing stablecoins could commoditize rails if Circle fails to preserve trust, liquidity, and developer adoption.

OPTIONS / MARKET STRUCTURE

option liquidity: good

ATM IV

—

TYPICAL BID-ASK

—

OPEN INTEREST

—

proxy inputs — dollar-ADV $1.7B · beta -0.462728 · px $100.85

source: proxy (no options chain on FMP)
FMP /stable/ exposes no options-chain endpoint on this key, so ATM IV, bid-ask spread and open interest are unavailable. Liquidity below is a PROXY from dollar-ADV, beta and price level (a stand-in for option depth), not measured option-market data.

CONFIRMATION — INSIDERS · 13F · LANGUAGE

Mixed — insiders selling, institutions adding, management language 5/10 measured.

INSIDERS selling 34 open-market sell(s) vs 0 buy(s) — net distribution

INSTITUTIONS (13F) adding as of 2026-03-31: 172 new / 99 closed positions; 379 increased / 111 reduced; institutional ownership +3.69pp; +81 net 13F holders

MGMT LANGUAGE 5/10 measured AI is strategically prominent, but language is broad and future-oriented; only limited concrete ownership via Agent stack launch.

commit “including today's launch of our new Circle Agent stack, which I'll discuss in greater detail shortly.”

commit “we are layering across this entire platform, the deep product and technology capabilities to compound alongside AI and Agentic economic growth.”

hedge “we believe will transform the global economic system over the coming decade.”

VERBATIM AI QUOTES

“We are rapidly moving into a world in which AI-powered software machines, coordinating on blockchain computers, deliver an increasing share of global economic activity.”
— Jeremy Allaire, Q1 FY2026

“we are layering across this entire platform, the deep product and technology capabilities to compound alongside AI and Agentic economic growth.”
— Jeremy Allaire, Q1 FY2026

“everything has been designed from day 1 with AI and Agentic flows in mind, from developer tooling to new services on Arc built entirely for AI agents.”
— Jeremy Allaire, Q1 FY2026

“Arc is built as an economic OS that anticipates this convergence with AI operating systems, and we are accelerating our product investments in this space, both on top of Arc and by building on the wide range of applications and integrations that already exist for USDC.”
— Jeremy Allaire, Q1 FY2026

“This morning, we launched agent wallets, a new product that allows agents to permissionlessly build onchain wallets, conduct transactions, on-ramp USDC, and operate within predefined policies and safety guard rails.”
— Jeremy Allaire, Q1 FY2026

“We also brought Agent Nano payments online, where USDC transactions as small as 1 million of a penny enable high-frequency machine-to-machine payments.”
— Jeremy Allaire, Q1 FY2026

“USDC already has an enormous lead in Angentic payments today, with 99.8% of all X402-Agentic payments being settled using USDC.”
— Jeremy Allaire, Q1 FY2026

“Our adoption of AI is not just outward looking. We are undertaking a significant internal transformation and AI adoption inside Circle is accelerating.”
— Jeremy Allaire, Q1 FY2026

“With this, our product velocity is increasing dramatically.”
— Jeremy Allaire, Q1 FY2026

“Our teams have already deployed over 600 AI native apps this year, as we have capitalized on major breakthroughs in agentic and AI development.”
— Jeremy Allaire, Q1 FY2026

“We are entering a world where, in my view, likely tens or hundreds of billions of AI agents will interact and perform economic functions over the Internet.”
— Jeremy Allaire, Q4 FY2025

“I want to talk specifically about AI at Circle. We are seeing an explosion of developer activity around AI and it's becoming an important driver for Circle's platform, and we believe an important and potentially significant driver for USDC adoption.”
— Jeremy Allaire, Q4 FY2025

“We're also building systems to better support Agentic payments.”
— Jeremy Allaire, Q4 FY2025

“we just went into Testnet release of a new capability with Circle Gateway that allows for agents to autonomously and programmatically automate cross chain USDC transactions with a transaction cost of [ $0.00001 ].”
— Jeremy Allaire, Q4 FY2025

“Now inside Circle, AI is also becoming foundational infrastructure across all of our functions.”
— Jeremy Allaire, Q4 FY2025

“Our product velocity is accelerating and I anticipate that to continue alongside the exponential improvements we're seeing from AI coding agents.”
— Jeremy Allaire, Q4 FY2025

ANALYST QUESTIONS ON AI

Q (Q1 FY2026, Unknown Analyst): I wanted to touch on Circle Agent stack and an update on stablecoin use cases for agentic commerce. How much adoption has there been so far? And what do you look to achieve over here?
A: The Circle Agent stack launch is a significant launch for us. It makes available all of the power of USDC of Circle's infrastructure to agents, whether they're agents that developers are building or just autonomous agents that already exist to be able to store value transact, do that safely.

Q (Q1 FY2026, Unknown Analyst): I wanted to touch on Circle Agent stack and an update on stablecoin use cases for agentic commerce. How much adoption has there been so far? And what do you look to achieve over here?
A: So we do actually believe that some meaningful portion of activity is AI-driven already, and we've seen that kind of sharply increasing in Q1 as these Agentic technologies have increased.

Q (Q4 FY2025, Devin Ryan): I want to start on kind of this agentic evolution. I think it's a compelling case. And just want to get a sense of how you think from a timing perspective this plays out? And does it start with trading liquidity and then progressing to payments and borrowing and lending or how do you see that? And then how do you make sure that USDC is in the middle of that? And can Arc perform relative to other Layer 1s technically to support this?
A: AI agents are realizing and the developers of those AI agents are realizing that agent to agent transactions need a reliable, low-cost trusted medium of exchange. And so virtually all of the AI payments infrastructure that we're seeing, the agent type activity is happening with blockchains. It's happening with USDC.

Q (Q4 FY2025, John Todaro): going back to the Agentic AI comments, I would agree with you. Just with the ways of the crypto equities have been trading and then just the crypto token market in general, is agenetic AI and payments and all that within those ecosystems? Do you see the excitement kind of extending beyond stablecoins in Arc? Could this be a general tailwind for the sector?
A: I mean I think this is one of the most exciting -- obviously, I'm biased, but I think this is one of the most exciting kind of points of convergence out there.

Q (Q4 FY2025, Peter Christiansen): regarding the conversation on [ Jeto ] commerce, which looks incredibly compelling, how should investors think about this opportunity transforming Circle's operating/financial model?
A: We think about this in a few ways. I think one is just as a major new demand driver for the utility of our stablecoin network.