Bernstein: Stablecoins Poised as Backbone for Agentic Payments; USDC Dominates Liquidity Race
Context and Chronology
Bernstein’s research note reframes stablecoins as infrastructural money for a new class of programmatic, agent-initiated payments while stressing that agentic flows remain a small, experimental subset of overall stablecoin activity. The note positions two parallel demand drivers: (1) broad, incumbent-led payment and remittance flows that already produce large, routable volume and near-term revenue for issuers and custodians, and (2) optional upside from machine-initiated microflows if deterministic settlement primitives, custody patterns and reconciliation tooling scale. That framing is reinforced by contemporaneous product previews and pilot work — Stripe’s guarded x402 preview on Base, CoinGecko’s 0.01 USDC-per-request experiments, Alchemy’s HTTP-to-onchain gateway, and Coinbase’s guarded agent-wallet trials and x402 activity on Polygon — which close functional gaps between discovery, attestation and settlement but have not yet produced durable, dollar-weighted scale.
Measured Activity, Methodology and Signal-to-Noise
Public measurement of early machine-payment pilots diverges sharply by methodology. A headline 30-day x402 tally near $24M reduces to roughly $1.6M after applying wash-trade filters and dollar-weighting, because many pilot flows are high-frequency, cent-sized events that inflate transaction counts without representing comparable economic value. Independent recalibrations and notes from analytics firms confirm this tension: counting messages (many micro-transactions) produces large impressions of usage, while dollar-weighted analyses show agent-originated commerce remains nascent. Bernstein and other analysts therefore caution against reading raw message totals as proof of routable settlement demand.
Liquidity Leaders and Competitive Positioning
On-chain liquidity and observable settlement usability create a default routing choice for developers building agentic stacks. Bernstein highlights a material advantage for regulated, deep-liquidity issuers: adjusted transaction throughput YTD centers around USDC (reported near $2.4T YTD in some adjusted tallies) versus USDT (near $1.4T). Complementary market research (eg, Macquarie) emphasizes the broader scale of stablecoin movement — combined capitalization and adjusted on-chain dollar movement estimates range across datasets (Macquarie cited ~$312B capitalization and ~$11T of adjusted on-chain flows for 2025) — a spread that stems from differing token scopes, inclusion rules and whether custodial cushions or active balances are counted. Those methodological differences matter because issuer economics, reserve management and routeability depend on where liquidity actually sits and how quickly it can be accessed for settlement.
Ecosystem Signals: Product, Incumbents and M&A
Product previews and integrations are moving from toy demos toward guarded, merchant-friendly stacks: Stripe’s SDK samples and CLI tooling (purl) for x402 on Base preserve taxation, refunds and reporting while enabling developer evaluation; CoinGecko’s API pricing tests show how metered data endpoints can monetize via small USDC charges; Mantle and other projects pair identity/reputation registries with liquidity plumbing; and middleware vendors (Alchemy, Alchemy-style gateways) map HTTP payment flows into on-chain USDC settlements. At the same time, incumbent strategy includes acquisition and issuance plays — Mastercard’s purchase of BVNK and bank pilots around tokenized deposits illustrate that large networks are building issuance and settlement capabilities rather than ceding rails to open L2 plays.
Risks, Policy Frictions and Operational Constraints
Delivery risks are both technical and regulatory: custody and key-management choices (guarded versus pure non-custodial keys), off-chain reconciliation, MEV/ordering and oracle reliability, dispute mechanics, Sybil attacks on reputations, and cross-jurisdictional regulatory divergence (eg, Europe’s MiCA frameworks versus evolving U.S. statute proposals) all shape where liquidity concentrates and which custody models are commercially viable. Supervisory focus on reserve transparency, freezing capabilities and AML/CTF duties increases the commercial advantage of regulated issuers that can pair liquidity with compliance-grade failovers.
Near-Term Implications
For executives and investors, the pragmatic takeaway is to treat early agent telemetry as directional rather than dispositive. Allocate exposure toward firms that combine issuance, custody and developer ergonomics (the profile Bernstein and others flag in Circle and Coinbase), track dollar-weighted throughput rather than raw message counts, and watch integrations that reduce per-transaction cost (L2s, state channels, transport-layer payment logic). Expect a segmented short-term equilibrium: sub-cent, high-frequency agent flows gravitate to low-cost L2 rails and guarded SDKs, while mainstream merchant and consumer spend remains mediated by issuer-provided integrations that preserve familiar compliance and reconciliation features. Material enterprise migration hinges on solving off-chain reconciliation, routing economics and regulatory clarity — a 6–18 month practical window for meaningful proofs of production rather than instantaneous disruption.
Read Our Expert Analysis
Create an account or login for free to unlock our expert analysis and key takeaways for this development.
By continuing, you agree to receive marketing communications and our weekly newsletter. You can opt-out at any time.
Recommended for you

Visa Faces Disruption From Stablecoins Powering AI Agents
Programmable stablecoins paired with autonomous software agents are creating credible low-cost settlement paths that can bypass legacy card rails; recent product previews (Stripe, Coinbase) and research (Citrini) prompted traders to reprice incumbents and redirected capital toward custody and settlement middleware. Near-term adoption will be segmented: sub‑dollar, high‑frequency agent flows favor L2 tokenized rails while regulated merchant flows will likely stay with card rails augmented by custody and compliance wrappers.

Stanley Druckenmiller: Stablecoins Poised to Become Core Payment Layer
Billionaire investor Stanley Druckenmiller told Morgan Stanley that stablecoins could form the primary payments backbone within a decade to 15 years while acknowledging bitcoin’s maturing role as a store‑like asset. Market data and industry pilots show rapid growth and corridor‑level adoption now, but regulatory divergence, reserve practices and engineering limits make a bifurcated outcome — bank‑backed tokenized deposits alongside private stablecoin rails — the most likely path.
Circle and Coinbase Positioned as Primary Stablecoin Proxies
Bernstein identifies Circle and Coinbase as the clearest public plays for stablecoin upside, while product previews from Stripe and middleware experiments show early infrastructure for machine-initiated USDC billing. Measurement methods differ — dollar‑weighted tallies and message‑count tallies produce divergent 30‑day figures — so near‑term economic scale remains pilot‑stage despite visible engineering momentum.

Stripe Signals Stablecoin-Led Surge in Agent Commerce
Stripe has opened a guarded preview that lets developers accept USDC from autonomous agents via an x402 path on Base, pairing its orchestration APIs with web-native on‑chain settlement and developer tooling. Parallel experiments (CoinGecko pricing, Coinbase and Mantle pilots, and emerging ERC-8004 registries) plus regulatory divergence mean the technical feasibility is real but the commercial topology — who captures routing, custody and compliance fees — remains unsettled.

Grayscale: Blockchains Poised to Become Payment Rails for AI Agents
Grayscale argues that autonomous AI agents with digital wallets could shift many tiny, automated payments onto public ledgers, with rolling, low‑value stablecoin flows the earliest operational signal. This thesis comes amid a broader repricing of software equities (estimates vary from ≈$1T to ≈$2T wiped) and rising demand for compute and observability, which together favour regulated stablecoin issuers, custody providers and layer‑2 scaling vendors — even as tools like OpenAI’s EVMbench expose new contract and privacy risks.

Macquarie: Stablecoins Reach $312B as Banks, Card Networks Adopt Onchain Dollars
Macquarie finds combined stablecoin market capitalization near $312 billion and estimates adjusted onchain dollar transfers at about $11 trillion in 2025. Independent datasets and surveys add nuance — circulating supply is frequently reported near $300B , Messari reported a recent weekly inflow spike of $1.7B , and user research shows rising payments and payroll use alongside persistent liquidity and AML concerns.
Coinbase Expands x402 to Polygon While Agent Payments Lag
Coinbase extended its x402 facilitator to Polygon as independent analysis puts monthly AI-agent payments near $1.6M , far below some published figures. Complementary industry moves — Stripe's guarded preview for USDC via x402 on Base , CoinGecko's 0.01 USDC per-request experiment, and Alchemy's payments gateway — show multi-rail momentum but also clarify why high transaction counts can translate to modest dollar flows.

Banks Adopt Multi-Provider Stablecoin Payment Rails
Banks are moving from single-vendor pilots to modular, multi-provider stablecoin rails to reduce vendor lock-in and improve cross-border payout resilience. This shift sits alongside competing trends — large vendors are also bundling integrated stacks — producing a bifurcated market where orchestration platforms and vertically integrated players each offer distinct trade-offs.