---
publisher: Heretic Research
series: Heretic Research · Issue 01 (Inaugural Issue)
title: "Positioning: Agent Economy Before Consensus"
title_zh: 身位:Agent Economy 在共识到来之前
language: en
date: 2026-06
canonical: https://hereticresearch.com/reports/issue-01/issue-01.en.md
chinese_version: https://hereticresearch.com/reports/issue-01/issue-01.zh.md
pdf: https://hereticresearch.com/downloads/HereticResearch_Issue01_Positioning_EN.pdf
evidence: appended at the end of this file (claims-to-source mapping with as-of dates)
website: https://hereticresearch.com
citation: When citing this report, please attribute Heretic Research and link the canonical URL.
notes: All views are our own. Figures are point-in-time; verify against the as-of dates in the evidence file.
---

Heretic Research is an independent research firm focused on the intersection of crypto and AI. We dig for the real data beneath the noise and build our research on it. We stand apart from any market narrative. All views are our own. Question everything. Alpha hides in doubt🔴

![TLDR poster](https://hereticresearch.com/reports/issue-01/figures/en/poster-tldr.png)

# TLDR

* **USDC monopolizes 99.8% of settlement, yet that is still just 0.00073% of the global e-commerce payments market.** Coinbase, Circle and Base have effectively cornered x402's base-layer settlement infrastructure. And yet today's $50M in cumulative volume is a rounding error against the $6.88T global e-commerce payments market. The incumbents have assembled the alliance, but the real users on the sidelines have yet to step in.

* **Stripe is running a textbook "barbell hedge," locking down its downside with $1.1B and its own ecosystem.** As one of x402's three governing bodies, Stripe has not bet its chips on a single standard. It pushes four fronts at once: joining x402 to keep a voice, building its own chain (Tempo), shipping a rival protocol (MPP) — already generating real commercial traction at the likes of Browserbase — and spending $1.1B to acquire Bridge and lock in the stablecoin rail. Whichever path wins, Stripe cannot lose.

* **x402's necessity today exists only in API micropayments; traditional retail e-commerce and enterprise SaaS have no reason to migrate on-chain.** In the current agent economy there is no use case that *must* use crypto. The data shows that beyond cross-organization low-value transactions and micropayment settlement, the approach lacks a core pain point in mainstream consumer e-commerce (B2C) and enterprise SaaS (B2B). Its growth ceiling remains unproven by data, and for now it rests entirely on incumbent demand for stablecoins as a supplementary settlement rail.

* **Value capture is brutal: x402 itself captures no value.** x402 is released under the Apache 2.0 open-source license — it levies no tax and issues no token. If the ecosystem takes off, under the current model it merely forces more capital into Circle's Treasury reserves to mint USDC (99.8% of settlement today), piles ever more sequencer fees onto Base (which carries 95% of flow), and helps cloud providers like AWS lock in enterprise agents' compute-hosting spend. Amid this brutal flow of value, web3's "agent infrastructure tokens" — lacking any clear fee model or claim on cash flow — have no clear value-capture logic yet.

* **The biggest gray rhino may be the higher-dimensional assault from APAC's leading player, Ant International, wielding 1.8 billion accounts.** The market holds an implicit assumption — "agent payment = crypto payment" — and that consensus is now being challenged by Ant International (Alipay+). Its AMP protocol likewise supports $0.000001-scale micropayments and integrates KYA (agent identity) and dynamic risk ratings, but it plugs straight into 40-plus mobile wallets, 1.8 billion users and 150 million merchants. Against x402's mere 170-odd native entities, once a Web2 giant takes a flesh-and-blood position in the APAC ecosystem, the on-chain narrative will face a serious challenge.

# Foreword

The agent economy is one of the hottest arenas of late 2025 and early 2026, with capital, protocols and institutions all converging at speed. The three questions the market asks most often — what is the agent economy, what role does crypto play in it, and how can one participate — have no clear answers today.

We believe the root cause is a more structural fact: agent technology has evolved explosively since the second half of 2025, far outrunning the pace at which understanding forms, rules are set, and consensus converges. In more classical terms, the leap in productive forces has outrun the update of the relations of production.

Against this backdrop, every leading player is already on the field, and in a frontier arena still evolving fast, mapping the whole landscape matters more than ever. This report therefore focuses on the most concrete piece of agent-economy infrastructure — payments — to make each participant's position visible: where it stands, where it looks, and what it is betting on.

# Part I · A Structure Out of Sync

# 1.1 The Triple Mismatch of Technology, Cognition and Rules

* **Technology layer: a dense run of milestones in six months (timeline)**

![Fig. 01 · The Infrastructure Race](https://hereticresearch.com/reports/issue-01/figures/en/fig01-infrastructure-race.png)
Source: Heretic Research

From the second half of 2025 into early 2026, the agent economy's technical infrastructure advanced at rare speed.
In November 2024, Anthropic open-sourced MCP (Model Context Protocol), defining the standard for exposing tools and data to agents.
In May 2025, Coinbase open-sourced the x402 protocol.
In October 2025, Visa launched the Trusted Agent Protocol (TAP).
In December 2025, Anthropic's MCP joined the Linux Foundation's Agentic AI Foundation.
In Q1 2026, Google launched AP2, OpenAI launched ACP (draft), Circle launched Nanopayments, and ERC-8004 deployed to mainnet.
On March 18, 2026, Stripe and Tempo jointly launched MPP.
On April 2, 2026, the x402 Foundation was formally established under the Linux Foundation, with 22 founding members present. AWS launched AgentCore Payments. Circle launched Agent Stack.
Within half a year, agent payments expanded from a single standard (x402) to at least four competing schemes (x402, MPP, TAP, AP2/ACP).

* **Cognition layer: capital FOMO far outpaces industry consensus**

In February 2026, Citrini Research published a long report on the agent economy, wiping 5–7% off the market caps of Visa, Mastercard and American Express in a single day. The reason for the panic: stablecoins might bypass the card networks' 2–3% transaction fee.

But the degree of panic was wildly out of line with reality. At the time, x402's real daily volume (after Artemis filtering) was only about $28K, against a $6.88 trillion global e-commerce market. That $28K of activity moving trillions in market cap shows that the market has formed neither a broad consensus nor a reasonable price for the agent economy — it is still swinging between the extremes of panic and euphoria.

The cognition of practitioners and investors is equally lagged and disjointed. In that same quarter, Dragonfly Capital swiftly closed a $650 million fourth fund (focused on the AI × crypto intersection), a display of capital's fear of missing out; yet on the flip side, only now did a16z hastily put forward the basic "Know Your Agent (KYA)" compliance framework.

In other words, as trillions in capital began to hunt the AI economy, the industry had not even laid the most basic compliance foundation of "how to define and see through an agent's identity." Capital has converged far faster than consensus has formed.

* **Rules layer: the "premature birth" of infrastructure and a consensus vacuum**

Even as capital and narrative accelerate wildly, the "rules-layer foundation" that truly decides whether AI agents can reach mass commercial use remains in an extremely raw and chaotic state. This disconnect between perception and reality shows up concretely as a consensus vacuum across four core dimensions:

**The identity vacuum ("Who is the actor?"):** Although identity standards led by ERC-8004 have deployed to mainnet, with on-chain data showing over 200,000 registered agents, the standard is still at the Draft stage. Because registration costs almost nothing (only gas), has no staking gate and lacks any verification mechanism, the mainnet is awash in indistinguishable junk and test accounts. The industry still cannot answer the most basic boundary question: **"What exactly counts as a genuine agent with a clear definition and boundary?"**

**Payment-standard fragmentation ("How to pay?"):** The market is in a warring-states era of three fragmented standards — x402, MPP and TAP. x402 takes the Apache 2.0 open-source route; Stripe's MPP is a giant's proprietary protocol; and TAP focuses on identity verification. Whether these three ultimately compete, complement one another, or go their separate ways may remain unresolved for the next 12 to 18 months.

**Coordination protocols left hanging ("How to interact?"):** Google's AP2 and OpenAI's ACP, which handle commercial coordination between agents, are likewise both at the draft stage. Neither has gone through the standard EIP/RFC industry process, and neither has any production deployment. Worse, AP2's emphasis on payment standardization and ACP's emphasis on communication standardization overlap clearly at the coordination layer, and the giants have yet to reach a compromise.

**The missing reputation mechanism ("How to trust?"):** In an agent economy dominated by micropayments (ticket sizes of just $0.01 to $1.00), there is, astonishingly, no independent reputation or credit protocol at all — no EIP proposal, no implementation, no authoritative backing. Because per-transaction values are so low, they cannot sustain an independent commercial credit-scoring model, leaving the entire ecosystem with an enormous trust deficit.

The vacuum and disorder across these four dimensions show that the mismatch in the pace of progress is a structural fact. The technology layer expanded from one standard to four in half a year, while the cognition layer is still at the "draft" stage.

![Fig. 02 · The Triple Mismatch](https://hereticresearch.com/reports/issue-01/figures/en/fig02-triple-mismatch.png)
Source: Heretic Research

# 1.2 "Ambiguity" Becomes an Equilibrium of the Transition

In a state of mismatch, blurred boundaries serve the participants collectively. All 22 institutions need the narrative that "the agent economy is developing fast." Coinbase needs x402's ecosystem data to justify the existence of the x402 Foundation; Visa needs the act of joining the x402 Foundation to answer market fears of "being bypassed"; Stripe needs multi-pronged bets to stay positioned on every possible path. Each has found its own equilibrium of self-interest within the ambiguity.

So the convergence of market consensus will depend more on substantive breakthroughs in benchmark data. Real business performance — Shopify's agent-payment bills, AWS x402 crossing the one-billion-transaction milestone, the genuine ARR of native projects — will be the catalyst that forces the market's pricing model from "FOMO-driven" back to "fundamentals-driven." Until then, ambiguity is the natural equilibrium of the market's game.

What we set out to do is to draw, within this ambiguity, a clear map to track the arena's evolution. We begin with the position of each participant.

# Part II · Eight Positions

In what follows, we break down the x402 Foundation's 22 founding members by their respective positions in the agent-payments ecosystem. These 22 global giants are not evolving in concert; each stakes out a differentiated position — across rulemaking, financial hedging, business extension and base-layer settlement — built on its own existing endowments.

To help the reader see clearly through this complex web of interests — internet infrastructure providers, traditional financial gateways, e-commerce and cloud giants, plus crypto-infrastructure and AI-native startups — we confront, for each position, the three most central questions of the commercial game: **its interest, its core endowment, and its capacity to hedge risk as the arena's fate is decided.**

![Fig. 03A · Eight Positions](https://hereticresearch.com/reports/issue-01/figures/en/fig03a-positioning-matrix.png)
Source: Heretic Research

# 2.1 Protocol Builders: Coinbase, Cloudflare

As co-founders and top-level governing bodies of the x402 Foundation, **Coinbase** and **Cloudflare** occupy the purest "standard-setter and network-gateway" position among the 22 founding members. Unlike the later groups that treat the protocol as a passive plug-in or a defensive asset, this pair's strategic intent is deeply aggressive: **they aim to make x402 the cross-industry base standard of the agent economy, thereby building durable rent-seeking power at the network and governance layers.**

Though the two giants are rooted respectively in the Web3 asset layer and the Web2 network layer, by jointly putting out this standard each completes a precisely targeted industry position:

**Coinbase (the top-level open-standard lawmaker):** On April 2, 2026, Coinbase formally donated the x402 protocol it created to the Linux Foundation. This "de-privatizing" political move is highly sophisticated: by establishing the protocol's open-source, globally neutral status, it minimizes the psychological barrier for other Web2 giants and developers to adopt it. As the protocol's definitive contributor, Coinbase thereby secures unshakable structural influence within the Foundation. Its deeper commercial gambit is to frame the rules of the game at the top through an open standard, and so complete a long-term ecosystem loop at the capital-routing and account layers below (for the specific monetization path, see 2.6, the Settlement Layer).

**Cloudflare (the physical occupant of the network-layer traffic gateway):** Where Coinbase pushes at the asset-governance layer, Cloudflare's bet locks directly onto the physical entry point of the entire agentic internet. As the de facto standard for global edge computing, Cloudflare has built native x402 support into its Workers and AI Agents SDK. As its CEO Matthew Prince made clear with his detached posture in the announcement, Cloudflare does not care which specific chain the final settlement clears on; all it cares about is this: **as long as agents' streaming interactions and payment requests still run on HTTP/Web protocols, Cloudflare — by virtue of its monopoly over global edge-network traffic — naturally holds the first routing node from which every agent sets out to check out.**

The shared success metric for both builders is x402's absolute ubiquity across the network ecosystem. As of May 2026, the x402 ecosystem had logged over 170 entities and more than 67 paying service endpoints, but the roughly $50M in cumulative volume is, to these giants, still a faint, very-early-stage figure. For protocol builders, however, the core metric at this stage is not settlement volume but the "speed at which the ecosystem moat converges" — that is, ensuring x402 can withstand early erosion by rivals (such as Stripe's MPP) and establish itself as the one unifying standard.

In risk structure, the two firms are markedly asymmetric. Cloudflare sits in a highly elastic **low-risk state**: whatever becomes of x402, its core business of carrying global web traffic through Workers faces no systemic shock. Coinbase, with its more asset-specific strategy, must instead confront — at the downstream settlement-execution layer — share erosion and governance challenges along the transaction chain from other base networks such as Solana (for the specific competitive dynamics, see 2.7, Edge Challengers of the Settlement Layer).

# 2.2 The All-In Hedger: Stripe

![Fig. 05 · Stripe's Barbell](https://hereticresearch.com/reports/issue-01/figures/en/fig05-stripe-barbell.png)
Source: Heretic Research

Among the x402 Foundation's three governing bodies, Stripe's strategic layout is the most complex. Unlike Coinbase and Cloudflare, which bet purely on the open protocol itself, Stripe runs a textbook "barbell strategy," massing forces at both ends — the open standard and its own closed loop. It is the only one of the 22 founding members to hold both a voice in the open standard and its own closed-loop infrastructure.

Stripe appears convinced that on-chain agent payments are inevitable, but equally certain that in 2026 no one can call which technical path will ultimately win. So, along an arc of "decreasing openness, increasing control," Stripe advanced four fronts at once.

Front one: joining the x402 Foundation as one of three governing bodies. As a governing body, Stripe holds structural influence in setting the protocol standard and secures a seat in the open-standard ecosystem.

Front two: spending $1.1 billion to acquire the stablecoin infrastructure Bridge, locking in the on-chain settlement rail early, and partnering with Paradigm to launch its own chain, Tempo, giving it independent on-chain infrastructure.

Front three: on March 18, 2026, shipping the Machine Payments Protocol (MPP) to take on x402 head-on. Routed through Stripe's API, the protocol begins to pull control inward, with Stripe natively providing fraud detection and tax handling.

Front four: partnering with Paradigm to launch its own chain, Tempo, giving it independent on-chain infrastructure and placing the lowest settlement layer entirely under its own control.

This multi-front structure makes Stripe's definition of success unusual: it does not bet on any single path winning, but ensures that whether x402 or MPP/Tempo becomes mainstream, Stripe holds an unshakable seat at the core toll booth.

It also means Stripe's only definition of failure is "on-chain agent payments being wholly disproven." And even in that extreme case, its traditional Web2 core would be entirely unharmed — the downside is perfectly hedged. From the real progress so far, the probability of that risk is narrowing fast. Per Stripe's official blog, in under two months since MPP's launch it had already logged real commercial adoption, including Browserbase's agents paying per session and PostalForm's agents autonomously paying to print and mail physical letters. The arrival of these native-agent commercial use cases is giving Stripe's all-around hedging game a very high win rate.

# 2.3 Defensive Occupants: Visa, Mastercard, American Express

If protocol builders are betting on future upside, then the traditional card networks led by Visa, Mastercard and American Express hold a position that is **purely defensive placement**. The most direct trigger for these three giants jointly joining the x402 Foundation in early 2026 was precisely the **Citrini Research report event** described above. Facing the valuation crisis of "stablecoins + AI agents will completely bypass the card networks" and the brutal lesson of a 5–7% single-day market-cap wipeout, they had no choice — joining x402 is a defensive response to market panic.

For that reason, their motive is exceedingly plain: **"not be bypassed."** Put differently, any agent-payment protocol will either come to nothing, or it must include the card networks in its ecosystem. On that logic, their success metric was already met the instant they joined the Foundation. This is a textbook reverse-option (insurance-buying) structure: the membership fees and limited engineering resources spent on joining (bounded cost) are far below the systemic risk of having future transaction-fee revenue wholly bypassed (unbounded payoff).

Within this collective insurance-buying, however, the three networks' strategic depth shows clear signal differences:

* **Visa (dual-track, staying proactive):** It is the only one of the three trying to keep the initiative in its own hands. Visa launched the Visa Trusted Agent Protocol (TAP) back on October 14, 2025, letting merchants verify whether the agent making a request is genuine and authorized, and in April 2026 it launched the Visa CLI wallet tool. Through a dual-track strategy of "join the alliance to hold a place + build a proprietary protocol to defend," Visa is trying to avoid being reduced to a mere pipe.

* **Mastercard and American Express (minimal positioning):** By comparison, these two giants' defensive posture is far more passive. They merely hold a named place in the Foundation, with no proprietary protocol and no public production deployment at all.

This reality of "strategically vital, operationally polarized" is visible in the seniority of their public endorsers. Whether Visa's Rubail Birwadker, Mastercard's Sherri Haymond, or American Express's Luke Gebb, the representatives sent to stand for each group are EVPs or global business heads. The giants dispatched their most presentable executives to signal "we haven't fallen behind" to Wall Street, yet at the layer of actual production deployment — apart from Visa fielding some CLI tooling — the other two remain, in substance, "troops not yet moved."

# 2.4 Business Extenders: AWS, Shopify, Adyen, Fiserv, PPRO, Microsoft, KakaoPay, thirdweb

This group is the most pragmatic, least narrative-dependent backbone of the 22 founding members. Their common trait: they hold vast incumbent businesses and treat x402 merely as a "natural extension" of their own turf or a "pluggable add-on." If the agent-payment trend takes off, they have taken a favorable place; if it is disproven, the core business faces no downside.

Given the many members, we can sort them by their place in the value chain into three core camps:

**Camp one: cloud infrastructure (AWS, Microsoft)**

**AWS:** In April 2026 it launched the preview-stage AgentCore Payments, using x402 at the base layer as the default runtime settlement component (with 200ms settlement on Base). Although AWS unilaterally disclosed 169 million cumulative payments in the first year, this is simply a natural extension of the Bedrock agent ecosystem, and its vast cloud business bears no systemic risk.

**Microsoft:** It shows the more silent posture of a giant. Although it has joined the Foundation with no public deployment, the market widely assumes that its Copilot Studio agents will inevitably need x402 as a potential communication bridge in cross-organization interactions. Its core Windows and Azure assets are likewise perfectly hedged.

**Camp two: scenario and settlement gateways (Shopify, Adyen, Fiserv, PPRO)**

**Shopify (the most eager to ready its "shelves"):** Its bet is highly business-driven — a conviction that vast numbers of future shopping orders will be placed by AI agents on behalf of humans. With its ready merchant network and transaction scenarios, Shopify must ensure the merchant side can receive agent payments; it is very likely to be the first of the 22 to produce "real business throughput."

**Adyen, Fiserv, PPRO (the passive adapters of merchant infrastructure):** These three veteran payment giants respectively dominate global gateways, U.S. merchant processing, and cross-border payment networks. Their strategy is textbook "defensive adaptation": merchants will never integrate a whole new settlement system for the occasional agent payment, and will insist that existing gateways support the new standard. They wait for merchant demand to force the issue — but once the moment arrives, their position is irreplaceable.

**Camp three: regional rails and vertical tools (KakaoPay, thirdweb)**

**KakaoPay:** As a Korean mobile-payments giant, its entry marks that the agent economy's influence has substantively penetrated APAC. Its core business is independent of the agent economy; its aim is to hold the potential agent-payment rail in the APAC market.

**thirdweb:** As a Web3 developer-tools platform, it provides SDK integration directly at the client layer. It bets on crypto-native developers' path dependence on its toolchain; if the trend does not arrive, its traditional Web3 dev-tools business remains stable.

# 2.5 The Coordination-Layer Builder: Google

Unlike the passive or defensive postures above, Google's position is the most distinctive — and most aggressive — of the 22 founding members. Although it has joined the x402 Foundation, its ambition was never in the base-layer "movement of money," but in setting the industry-wide rules and authorization layer above the payment layer. This strategy shows fully in the proprietary protocol it pushes in parallel — AP2 (Agent Payments Protocol).

In the traditional payments world, "security and compliance" are guaranteed by the cardholder entering a password and the risk system intercepting fraud. But in the agent economy, when an AI pays for something on your behalf, **base protocols like x402 only move money from account A to account B; they cannot verify whether this AI's identity is genuine, how much its owner authorized it to spend, or whether this high-frequency micro-transaction is compliant.**

This is precisely the vast gap the base transfer protocol leaves behind. Google's AP2 tries to set industry-wide rules in this chaotic vacuum by defining agent identity verification, authorization flows and compliance checks. Once AP2 becomes the de facto authorization standard, it becomes the entire ecosystem's "gatekeeper." Once AP2 wins, x402 degrades into a concrete execution tool beneath its wings.

So Google's endgame is clear: **use the AP2 framework to unify agents' authorization layer, so that every base payment standard is authorized through AP2.** Its success metric is whether AP2 can become the default authorization framework for top agent platforms. Google is now trying to force-inject this standard into its own ecosystem; if Google Cloud's Vertex AI Agent Builder adopts AP2 as the default, it would instantly gain the largest industry user base on the network.

From current production deployment, Google has taken a seasoned "occupy first, build later" strategy. Although AP2 still sits at the draft stage, Google's core seat in the Foundation ensures it holds great strategic initiative either way:

* **If AP2 fails:** Google's place in the x402 alliance can still serve as a traditional cloud giant's defensive position.

* **If AP2 ultimately succeeds:** that seat lets it, as the "big brother," forcibly lead and shape how x402 and AP2 coordinate.

Of course, the biggest hidden danger in this position is factional splintering among the giants. As noted, Google's AP2 and the ACP (Agent Commerce Protocol) that OpenAI strongly champions overlap heavily in function at the coordination layer. This "two-tiger rivalry" may become an industry headwind that slows the convergence of the whole agent coordination layer.

# 2.6 The Settlement Layer: Circle, Base

Across all the positions above, whether protocol-setting or business extension, everything must ultimately converge on the substantive movement of assets. The Circle–Base combine occupies the most indispensable "settlement layer" position among the 22 founding members.

The intrinsic logic of this position is brutally hard-edged: **they do not fight over upper-layer agent applications or traffic gateways; instead, using the base architecture of "stablecoin + public chain," they directly occupy the settlement layer that traditional finance cannot carry.**

The agent economy imposes two demands on financial infrastructure that traditional card networks cannot meet: **ultra-high-frequency micropayments (nanopayments) and instant streaming settlement.** In the AI era, agents must pay in real time for a few milliseconds of compute or a few tokens of inference, with transaction amounts often as low as a ten-thousandth of a cent (sub-cent). Traditional card-network settlement, with its fixed gateway fees (e.g., $0.10–$0.30 per transaction) and long T+N cycles, logically and completely forecloses high-frequency micro-transactions. That opens a vast technical and financial arbitrage space for Circle and Base:

**Circle (the issuance monopoly of digital fiat):** Its strategic endgame is to make USDC the "reserve currency" of the agent world. As of end-April 2026, the overwhelming figure that 99.8% of all x402 transactions settle in USDC proves it has established absolute issuance dominance. Circle's subsequent moves all revolve around this base advantage: its Nanopayments scheme supports transfers as small as $0.000001, paired with the Agent Stack toolchain released on May 11, aiming to make USDC the default unit of value and store of funds in all agent logic.

**Base (the infinite descent of marginal settlement cost):** As the physical settlement lane carrying this digital currency, the Base chain is driving "marginal settlement cost per transaction" toward its limit. As of mid-May, 95% of x402 volume runs on Base. To fully match agents' consumption rhythm for instant resources, Base's batch-settlement technology, live on May 13, 2026, forced per-transaction cost below $0.0001. As Jesse Pollak noted, this ultra-cheap, near-zero-cost settlement capacity "unfroze," for the first time in financial history, the commercial use case of high-frequency sub-cent payments.

For the settlement layer, success hinges not on whether any one AI application wins or loses, but on the **total throughput (GTV) of the entire on-chain agent economy.**

Although they face potential carving risk from Stripe's proprietary MPP arriving with its merchant network, as long as the ecosystem network effect formed by x402's current 170+ entities and 67+ paying endpoints keeps snowballing, Circle and Base have — through their "irreplaceable, ultra-low settlement cost" — already locked in the most stable base-layer tax seat in the agent economy.

# 2.7 Edge Challengers of the Settlement Layer: Solana Foundation, Polygon Labs

Within the x402 Foundation's global financial pool, the Matthew effect of the settlement windfall is accelerating. Since head chains led by Base have already taken the lion's share of volume, the Solana Foundation and Polygon Labs — also base-layer public-chain infrastructure among the 22 founding members — hold a very clear core intent: **to break the existing monopoly and grab a larger share of settlement before the winner-take-all network effect fully hardens.**

Because the agent economy is highly sticky around network effects and ecosystem developers, this group faces different path choices in contesting the long-tail share:

**Solana Foundation (an aggressive share-grabbing strategy):** In x402's very early stage, Solana — with the high throughput and ultra-low latency of its monolithic chain — was the natural first choice for high-frequency agent micropayments. Per its head of AI growth, Rishin Sharma, in the official April 2, 2026 press release, Solana at one point drove nearly 65% of x402 volume early that year. But as Base forcefully shipped batch settlement in mid-May (driving per-transaction cost below $0.0001), the balance of attack and defense shifted. Although Solana is now urgently deploying an x402 Facilitator to lower migration barriers and retain developers, lacking Coinbase's vast compliant user network and deep USDC integration, it faces enormous competitive pressure in holding settlement share.

**Polygon Labs (a passive fallback strategy):** Compared with the head-to-head between Solana and Base, Polygon Labs takes a more defensive, "elastic backstop" stance. Although x402 has completed base-layer deployment on the Polygon network, Polygon has not made aggressive share claims in the discourse or the ecosystem. Its core logic leans toward an "alternative option" strategy: when high-load networks like Base or Solana suffer outages, performance collapse or fee spikes from overload of high-frequency agent micropayments, Polygon — with its mature, mass-produced EVM-compatible network — will serve as the safest pool to absorb the overflow.

For these two edge challengers, the risk fundamentally comes from how fast the Matthew effect hardens. The current trend shows Base's dominance strengthening, not weakening. They have squeezed into the 22-member founding club, but unless they can use distinctive technical means or differentiated scenarios to carve into the "Base chain + USDC" iron triangle in short order, over the long run they will hold a relatively passive place in the distribution of this agent-settlement windfall.

# 2.8 Native Builders: Sierra, Merit Systems, Ampersend.ai

Among the x402 Foundation's 22 founding members, the native builders form the highest-risk control group of the entire map. Unlike every player above with a traditional business generating cash that treats the protocol as a "pluggable add-on," this group of startups ties its core products, valuation network and even financial survival entirely to the growth rate of the agent economy. For them, this is not a low-marginal-cost business extension but a one-way bet of extreme asset specificity.

Because they place all their chips on the agent economy's growth curve, they show a highly concentrated profile in ecosystem loop and risk tolerance:

**Sierra (enterprise-grade agent infrastructure lacking cash-flow hedge):** Founded by Bret Taylor — former Salesforce CEO and OpenAI chairman — Sierra is the most telling commercial specimen in this group. Sierra's business model provides large enterprises with the base platform to build and deploy AI agents. The latent risk of this path is that its commercial loop depends entirely on how fast enterprise customers actually deploy agent use cases. Unlike cloud giants such as AWS, Shopify or Microsoft, which have a vast Web2 base as a cushioning network, Sierra has no traditional business to hedge with. If the agent economy fails, within the next 12 to 18 months, to cross the early-adopter chasm and reach substantive commercial scale, Sierra's enterprise customer base — built atop a high premium — will face direct growth-stall pressure.

**Merit Systems and Ampersend.ai (the real-demand test for long-tail native toolchains):** These two startups, with relatively limited public information, are typical native-toolchain builders of the agent ecosystem. Their commercial color is the same as Sierra's: the slope of their revenue growth is locked entirely to the real transaction volume of the agent economy. This forces them, at this stage, to confront the objective reality that the whole industry's data sits in an "early fog." Although the x402 ecosystem had logged over 170 entities and more than 67 paying endpoints as of May 2026, from public data outsiders still cannot precisely see through and separate how much is real commercial throughput running a closed loop, and how much is merely an experimental network at the test stage.

For native builders, their risk structure is the most exposed to industry-cycle swings of all eight positions. They have neither the card networks' risk lock-in from buying a reverse option, nor the settlement layer's marginal-cost advantage of collecting tolls passively. When the macro convergence of the whole agent economy runs slower than the industry expects, these native firms — wholly lacking a traditional cash-generating core and highly dependent on the resonance of the funding environment and industry growth — will be the first to bear the survival pressure of a growth shortfall.

# Part III · Where the Lines of Sight Cross

The eight positions broken down in Part II are merely each founding member's initial placement at the ecosystem's early stage.

In real commercial evolution, no single institution can swallow the entire agent-economy payment windfall on a single endowment: the Web2 giants holding traditional enterprise customers and compute lack an on-chain reserve currency, while the Web3 infrastructure holding high-frequency settlement lanes lacks the merchant trust of traditional commerce.

So, driven by settlement efficiency and commercial interest, the current founding members have not moved toward a many-sided war but toward **a "Coinbase + Circle + Base" core settlement layer, with Web2 giants integrating across the two wings of "identity/compliance" and "compute/distribution."** The future agent-payment pool will play out as a deep contest of asset assembly and standard positioning around this "one core, two wings." This dense positioning by multiple forces on the eve of standard-setting raises market expectations while also planting the seeds for the fracturing of on-chain transaction data.

# 3.1 The Emerging "Core Settlement Alliance + External Integration Camp" Structure

![Fig. 06 · One Core, Two Wings](https://hereticresearch.com/reports/issue-01/figures/en/fig06-one-core-two-wings.png)
Source: Heretic Research

**Core settlement alliance: Coinbase + Circle + Base**

This is the fastest-moving, most asset-integrated alliance in the entire ecosystem. The confluence of these three institutions essentially locks the agent economy's **issuance, routing and base ledger** inside a single bloc.

To see through this alliance's monopoly logic, one must grasp how the three establish absolute competitive advantage through asset assembly:

Circle holds the issuance rights to USDC, the reserve currency of the agent economy. In a traditional multi-chain environment, cross-chain asset settlement carries high security loss and heavy fees. But by binding natively and deeply to Coinbase's Base chain, Circle's Nanopayments scheme can drive per-transaction settlement cost below $0.0001. This combination of "fiat asset + extreme-low-cost ledger" physically destroys the competitive basis of other L1/L2 chains (such as Solana and Polygon) in the micro-settlement scenario.

Coinbase, as x402's creator, designed the rules so the standard fits USDC and Base natively. Every time an agent settles at high frequency via x402, funds accrue as a network tax (gas) on Base, simultaneously growing Circle's Treasury-reserve interest in the Web2 banking system, and ultimately converting back into strategic chips for Coinbase within the Foundation.

This asset-specific loop is producing a strong "ecosystem capture" effect on other founding members: traditional IT giants (like AWS) hold merchants and compute but no compliant-stablecoin issuance, so to cut enterprise settlement development costs they are forced to wire their payment gateways into the alliance's base ledger. Traditional card networks (like Visa and Mastercard), facing this alliance's "on-chain direct peer-to-peer settlement" bypassing the traditional bank gateway, have no choice but to take defensive placements for long-run compliance cover.

**External application-side integration camp: Google AP2's compatibility with the base settlement protocol**

On the application and compliance side, traditional tech giants led by Google are accessing the base settlement protocol above via standard compatibility. The AP2 protocol focuses on the upper layer of identity, authorization and compliance; the x402 protocol focuses on the lower layer of actual fund movement. Their confluence is a textbook case of "inter-standard coordination."

The commercial logic of this coordination: if AP2 successfully unifies the agent identity-authorization layer, x402 becomes the default payment-execution layer under the AP2 framework. This means the standards battle at the coordination layer (L5) may converge early under the AP2 framework — without waiting for AP2 and OpenAI's ACP to fight it out, Google can fold x402 directly into its payment-layer spec.

From current progress, the relevant data is structural, not scale-based: AP2 is still at the draft stage with no public production deployment. But two signals point clearly to this convergence: first, Google's place in the x402 Foundation shows it believes AP2 and x402 intersect; second, Ant International (Alipay+) joined AP2 as a launch partner in January 2026, showing AP2's framework is attracting APAC's leading payment players. The core tracking metric for the next 12–18 months is whether AP2 officially publishes a formal spec including x402 as the default stablecoin rail.

**External compute-side integration camp: AWS AgentCore's access to the on-chain settlement track**

On the compute and distribution side, traditional cloud giants are moving toward the core alliance through product-interface integration.

AWS AgentCore Payments, which entered preview in April 2026, achieves 200-millisecond USDC settlement on the Base chain. This is the world's first public case of "on-chain payment invoked by an off-chain product" — a fully off-chain cloud service (AWS Bedrock Agents) completing payment settlement on-chain via the x402 protocol.

The real significance of this case is a technical confirmation of path feasibility, showing that head cloud providers see real commercial value in accessing x402 as a payment rail. Once this integration model moves from preview to GA (general availability) and is replicated by other cloud providers or SaaS platforms (such as Google Cloud, Microsoft Azure, Shopify), x402's adoption curve will expand from the crypto-native developer circle into the traditional enterprise-IT market.

# 3.2 The Great Data Gap: Volume Inflation Driven by an Arms Race

The x402 founding members' positioning war has, on the market side, formed an expectation that the agent-payment ecosystem is landing at speed. Yet when the eye returns to the real on-chain data, x402's on-chain performance shows a striking data gap.

![Fig. 07 · The Great Data Gap](https://hereticresearch.com/reports/issue-01/figures/en/fig07-data-gap.png)
Source: Heretic Research

**Order-of-magnitude divergence in multi-source data, and its technical cause**

In March 2026, CoinDesk disclosed three heterogeneous datasets around the x402 protocol within the same 30-day window: x402's self-reported volume of $24 million; third-party analytics firm Allium Labs' measured $3 million; and, after applying a wash-trading filter, Artemis' measured $1.6 million. The three results diverge by orders of magnitude. Meanwhile, as of end-April 2026, Circle officially disclosed cumulative figures of $50 million and 169 million transactions.

This vast gap across data dimensions does not stem from ordinary differences in statistical methodology. Artemis' daily-average data for March 2026 show that, of 131,000 daily transactions, settled value was only $28,000, the average transaction was $0.20, and roughly 50% of activity was flagged as wash trading.

This data signature reveals the ecosystem's current base-layer technical trait: **x402's protocol design makes real commercial transactions and developer test/wash activity technically indistinguishable on the base ledger.** Because the protocol, at this stage, prioritizes decentralization and settlement efficiency — being permissionless, identity-free and ultra-low-cost per transaction — a developer's high-frequency self-calls to their own API endpoints and the autonomous commercial settlement flow between agents present code-identical signatures on the base ledger. So at this stage, wash trading should be seen as a **structural feature** of this settlement protocol at a particular evolutionary stage.

**The amplifying effect of positional differences on volume data**

Beyond the technical trait, the interests driven by each founding member's positional differences (described in 3.1) are the commercial force catalyzing the inflation of volume data. At a tipping point where the base protocol standard has not achieved absolute monopoly and external integration camps (Google, AWS) are entering one after another, participants in different positions all need to prove their ecosystem's growth to the market, creating a commercial motive to manufacture on-chain evidence of use:

* The core alliance's protocol initiators (such as the relevant foundation) need inflated transaction data to establish their protocol standard's industry standing and keep a voice in the competition against other potential standards;

* Issuers of the base settlement asset and ledger owners (such as Coinbase and Circle) need high-frequency volume to expand USDC adoption and lock in Base's network gas fees and Treasury-reserve interest;

* Native developers at the application layer need high-frequency call data to meet the data benchmarks of venture-capital scrutiny.

The commercial demands of these interested parties sum up at the settlement layer, ultimately showing as heavily inflated fake volume. The share of fake transactions noted in the Artemis report essentially reflects each institution's data-positioning behavior on the eve of standard convergence, not the moral failing of any single participant.

So, based on a see-through assessment of the above data gap and the institutional-interest mechanism, this research gives the following final read on the overall industry reality presented in Part III: **the on-chain verifiable volume of real agent payments is still at the earliest stage, but its commercial structure is essentially set.**

Although the absolute real settlement value, once the water is squeezed out, remains faint, an ecosystem network has already formed in which settlement- and ledger-layer assets keep concentrating into the "Coinbase + Circle + Base" core, while traditional cloud providers and the application side provide interface support across their channels. At this point in the cycle, the structural convergence signal released by this institutional game has greater directional value than short-term swings in the absolute book figure.

# Part IV · Crypto's Place

# 4.1 Crypto's Core Battleground: Which Use Cases Are Real Demand, Which Are False?

![Fig. 08 · Real Demand vs. False Demand](https://hereticresearch.com/reports/issue-01/figures/en/fig08-demand-tiers.png)
Source: Heretic Research

In the agent economy, crypto assets and on-chain ledgers cannot wholesale replace the existing financial system. To judge whether a payment use case truly needs crypto, the core test is: are the fixed transaction costs of the traditional settlement network too high in that case, or is there a compliance blind spot it cannot cover?

If we segment current market demand, it is sharply polarized:

**Sub-cent micro-settlement: the real-demand battleground where crypto holds absolute pricing power**

In traditional payment networks — whether card networks like Visa/Mastercard, the U.S. ACH, or domestic interbank settlement — every transaction carries an insurmountable fixed-cost floor (typically $0.01 to $0.10), owing to fraud-risk management, multi-tier settlement-bank splits and manual compliance review.

But commercial interaction between agents is high-frequency and tiny. For example, each time one agent calls another agent's API, the real value may be only $0.0001 or even less. Here, the traditional payment system's fees run orders of magnitude above the transaction itself — economically unworkable.

Today, Circle's Nanopayments scheme combined with the base ledger can run transfers as small as $0.000001 per transaction, and via batch accounting achieves gas-free settlement. This is traditional finance's physical blind spot, and the only use case in the entire agent economy where the crypto ledger has won conclusive validation and cannot be replaced.

**Cross-organization peer-to-peer payment: a potential use case of moderate necessity, gated by the governance system**

When two independent organizations with no prior contractual relationship have an instant transaction need between their respective agents, the on-chain settlement system is structurally superior to the off-chain traditional B2B approach. Because on-chain settlement has the trait of "payment is settlement, settlement is authentication," in theory the two sides need no prior commercial contract and no cumbersome interbank integration.

However, the precondition for this use case is that on-chain identity, authorization and reputation systems are fully mature — and these are badly disjointed today. Take ERC-8004, designed specifically for agent identity and request verification: it is still at the Draft stage, and although there are over 200,000 registered agent entities on-chain, it is technically impossible to tell real commercial entities from test or wash nodes. So until on-chain credit and identity standards land, cross-organization agent payments can only stay in low-value, non-core edge transactions.

**"False" demand: consumer e-commerce and enterprise SaaS subscriptions**

By contrast, where a human directs an agent to place an order on an e-commerce site, or an agent auto-renews a SaaS subscription, crypto assets hold no advantage at all.

For merchants like Shopify or enterprise customers, what they care about is: authorization success rate, chargeback-risk control, bad-debt coverage, automated tax accounting and refund protection. Stripe, Visa and the bank gateways have iterated on this for decades, building highly efficient settlement networks with built-in buyer-trust protection.

In such cases, loading USDC into an agent's wallet to pay does nothing to cut costs on the merchant side, and instead brings two flaws that traditional commerce cannot accept:

**Irreversible settlement:** an on-chain payment, once sent, cannot be refunded, lacking the dispute-handling and arbitration mechanism (chargeback) built into traditional credit cards.

**Surging tax and compliance friction:** an enterprise directly receiving a compliant stablecoin faces extremely cumbersome capital-gains-tax accounting and anti-money-laundering (AML) review under financial regulation.

So in e-commerce and SaaS subscriptions, crypto's necessity simply does not hold. As for DeFi and on-chain-native execution scenarios (such as agents auto-calling smart contracts or trading in DeFi), although using a crypto wallet there is real demand, that value capture is wholly enclosed within the crypto ecosystem and does not extend directly to the broader real-commerce market.

# 4.2 Crypto Infrastructure Caps the Agent Economy's Ceiling

To see crypto assets' real ecological niche in the whole agent ecosystem, one cannot stop at macro use-case segmentation; one must look through to the five core base-layer technical components. And these five are badly out of sync in their progress. The structural contradiction — "base-layer settlement tools extremely mature, while upper-layer governance and credit tools badly lag" — directly determines crypto's reality of being unable to expand into high-value commercial scenarios.

![Fig. 09 · The Lopsided Infrastructure](https://hereticresearch.com/reports/issue-01/figures/en/fig09-lopsided-infrastructure.png)
Source: Heretic Research

**Wallets and stablecoins (high maturity): enough to support edge micro-settlement**

Wallets: give agents the technical ability to hold assets independently and pay without human confirmation. But today's agents mostly run within human-set parameters, where platform debits or traditional billing suffice; large-scale wallet demand has not yet arrived.

Stablecoins: hold a structural edge in cross-border, low-value, high-frequency agent micropayments — 99.8% of x402 settlement in USDC is already a fact. But this figure reflects crypto-native developers' choice; in the broader agent-payment market (e-commerce, enterprise trade-credit terms, scenarios already covered by credit cards), stablecoins' necessity does not hold.

**Smart contracts: deliberately compromised for settlement efficiency**

Smart contracts are, in theory, highly composable, able to support complex conditional settlement and escrow. But in the actual evolution of agent payments, this component's commercial role has been deliberately marginalized.

The current industry consensus is to reduce smart contracts to the most basic transfer trigger. Because in consumer and high-frequency micro scenarios, an overly complex smart contract would sharply raise gas costs and introduce contract bugs and queuing-arbitrage loss (MEV) and other security risks. This deliberate technical compromise means that, at this stage, the crypto ledger cannot carry complex, multi-party commercial-contract execution.

**On-chain identity and reputation systems (badly lagging): the wound that locks high-value commercial scenarios shut**

The "peer-to-peer transaction between two strangers' agents" imagined above is crypto's most imaginative commercial blue ocean. But for this use case to move from theory to production-grade landing, it must rely on two upper-layer components: on-chain identity (knowing who the other party is) and on-chain reputation (knowing the other party paid and won't default).

And these two are precisely the weakest, most lagging links in the entire infrastructure, directly locking down crypto's outward penetration:

**On-chain identity standards stuck in labor:** ERC-8004, designed specifically to solve agent identity and request verification, still sits at the Draft stage. Though over 200,000 agent entities are registered on the books, it lacks an effective verification mechanism and is technically incapable of telling real commercial accounts from developers' test nodes.

**On-chain reputation essentially blank:** the market has not only failed to form a unified industry standard but also faces a brutal commercial paradox: agent micropayments typically run at sub-cent ticket sizes, and such low per-order value simply cannot sustain an independent credit-assessment and tamper-resistant record system.

Seeing through the evolution of these five components, crypto assets' current marginal position in the ecosystem is the inevitable result of the infrastructure's lopsided development. "The high maturity of the base money rails (wallets/stablecoins)" and "the severe lag of the upper compliance-governance tools (identity/reputation)" form the core technical contradiction of this stage, leaving crypto unable to shake traditional finance's absolute moat within any foreseeable cycle.

# 4.3 Commercialization Cycle and Value Flow: Crypto's Real Boundary and Value Capture

In the 12–18-month research cycle ahead, crypto assets' positioning in the agent economy is highly clear: it cannot shake the core mainstream-commerce pool, and can only serve as a supplementary settlement layer for the edge blind spots that the traditional financial network cannot cover. Bound rigidly by this ecological boundary, the whole ecosystem's value capture shows a sharp fracture — rising adoption of the open protocol cannot automatically translate into real token appreciation, and economic value concentrates into the traditional and chain-layer giants that hold compliance and ledger moats.

To see through the business essence of "who makes money," this research summarizes the financial monetization paths of the ecosystem's core actors as follows:

![Fig. 10 · Who Makes Money](https://hereticresearch.com/reports/issue-01/figures/en/fig10-value-capture.png)
Source: Heretic Research

At the same time, the absolute dominance of the traditional financial network forecloses any chance of this capital breaking upward into high-ticket core scenarios. In physical e-commerce, enterprise SaaS subscriptions and large B2B transactions, merchants and enterprises depend heavily on the credit-card network's chargeback protection and compliant AML accounting. Because the crypto settlement rail has the trait of "irreversible payment" and faces cumbersome capital-gains-tax friction, it holds no replacement advantage in traditional commerce. This boundary constraint locks the whole ecosystem's economic value firmly into the edge infrastructure layer: the base ledger can only earn the sequencer's thin fixed fee from batch settlement, while external cloud providers deepen platform stickiness by integrating x402 — completing the loop of the ecosystem's clearest profit source entirely at the equity level of traditional companies.

The harsher technical reality is that even in micro-settlement, the only real-demand battleground, the current scale ceiling cannot be effectively validated by technology. To date, the network's cumulative $50 million in volume is still under one-thousandth of the global payments pool. As noted in Section 3.2, because x402 cannot distinguish a developer's benchmark testing from real commercial settlement between agents on the base ledger, at least 50% of the book flow is fake (wash trading). Until the test water can be effectively removed and real merchant endpoint revenue can be seen, capital markets simply cannot confirm that this arena has the growth room to leap to a higher valuation. Any agent-native token lacking a clear fee mechanism, lacking a claim on cash flow and driven by narrative alone will face the rigid bursting of a valuation bubble once the technical fever subsides.

# Part V · Tracking Signals

![Fig. 12 · The Tracking Map](https://hereticresearch.com/reports/issue-01/figures/en/fig12-tracking-map.png)
Source: Heretic Research

In the first four parts, we have set out the current industry positions, the crossing consensus, and crypto's real boundary. But this is only a static, point-in-time read. Any judgment, the moment it is written, begins to bear the risk of missing the move or being disproven. So rather than offer an indisputable final verdict, we anchor every core view to publicly trackable metrics over the next 12–18 months. This tracking map is the full set of chips we are placing at this point in May 2026.

This is a fully symmetric monitoring framework. The six confirming signals on one side, once they cross their preset thresholds, will directly establish the long-term infrastructure narrative; the four falsifying paths on the other side, if any is borne out, suffice to overturn most of the current assumptions. We put the grounds for confirmation and falsification under the same yardstick for one purpose: to guard against the blind spots that arise in research when a position is taken as a given. Whether a view is worth holding depends on whether it clearly defines the conditions under which it fails.

Within this symmetric game, the core downside risk comes from the "gray rhino" revealed in FIG. 11 — Ant International's AMP protocol. The market today widely holds an unexamined assumption: that "agent payment equals crypto payment." Yet through a side-by-side comparison of x402 (a crypto-native protocol) and Ant International's AMP, we find the two barely overlap in footprint or core path:

**A mismatch of technology and path:** x402 tries to lay a new track on-chain and build the ecosystem from zero, serving the supply-side developer base; AMP instead grafts onto the existing network of a mobile wallet (Alipay+), driven by the demand side of APAC consumers and merchants.

**A higher-dimensional assault in scale:** while x402 is still validating 169 million cumulative transactions and 170+ connected entities at small scale, AMP is backed by a built-in installed base of 1.8 billion user accounts and 150 million merchants. On key dimensions such as micropayment capability and card-network cooperation, AMP has essentially achieved a comparable — or, in dynamic risk control (KYA), even more advantageous — alternative.

This means that once AMP lands for real across APAC on that base of 1.8 billion users, the on-chain narrative will face systemic pressure to restructure. This is precisely the trip-wire signal we must monitor closely among the falsifying paths.

Note that each signal's specific position right now does not matter; they are merely qualitative, point-in-time assessments. What is truly decisive is the moment a threshold is crossed — before that, every conclusion stays open. Over the coming cycle, we will keep recalibrating along these ten paths. Whichever line first triggers an exit or an achievement, that core question about the future is the first to get its answer.

So what is truly decisive is precisely the moment these thresholds are crossed. We should also clarify one premise: listing these stringent falsification conditions in the framework is by no means pessimism about x402. On the contrary, it is proof that the whole industry is still at a very early, exploratory stage.

Top-tier infrastructure is never born amid smooth, unbroken praise; it is forged in boundary friction with incumbent giants, and in the market's most genuine scrutiny and self-correction. Precisely because we hold a considered optimism about this industry's long-term potential, we choose to dissect the arena's future on a foundation that respects the real data and the real situation.

Acknowledging uncertainty is the precondition for staying sharp. Facing the positional contest between x402 and the TradFi behemoth, and facing the ten signal lines extending over the next 12–18 months, we will not break our word: as critical observers and fellow travelers, we will keep tracking and recalibrating.

This report is the first installment in a research series on the crypto × AI intersection. Further deep-dives in the series will follow. HR will always stand on the side of the real data obscured by popular narrative, presenting readers with the market's truth. Stay tuned.

---

# Appendix · Public Evidence Ledger

> **Purpose**: lets readers and AI agents verify the key claims in this report. Every data point carries its source and an as-of date.
>
> **How to verify**: all figures are point-in-time. Stale means "old," not "wrong" — verify against each entry's as-of date, not today's date.

## 0. TLDR Claims ↔ Evidence Index

| Claim in the report | Supporting evidence (section below) |
|---|---|
| USDC monopolizes 99.8% of settlement, yet that is 0.00073% of global e-commerce payments | §3 Scale data, §6 Reference data |
| Stripe is running a "barbell hedge" across four fronts | §2 Timeline, §5.2 Stripe action list |
| x402's necessity today exists only in API micropayments | §3 Scale data, §4 Conflicting-data table |
| x402 levies no tax and issues no token; value flows to Circle / Base / cloud providers | §1 Protocol facts, §3 Scale data |
| The biggest gray rhino may be Ant International's (Alipay+) AMP protocol | §5.6 AMP (sources in Part V of the report) |

## 1. x402 Protocol Facts

- Built on the HTTP 402 "Payment Required" status code; instant micropayments in USDC. Five-step flow: client request → 402 response with invoice → client signs transfer authorization → server verifies → 200 OK
- License: **Apache 2.0** — zero protocol fees, zero account creation, zero vendor lock-in (the protocol itself has no fee or token mechanism)
- Supported chains (as of 2026-05): Base, Solana, Ethereum, Polygon, Stellar, Avalanche, Optimism, HyperEVM, XRPL, Algorand, MegaETH, Arbitrum, Sei, Sonic, Unichain, World Chain
- SDKs: TypeScript, Python, Go

Sources: x402.org technical docs, Linux Foundation x402 Foundation governance docs

## 2. Key Timeline

| Date | Event | Source |
|---|---|---|
| 2024-11 | Anthropic open-sources MCP | Anthropic announcement |
| 2025-05 | Coinbase open-sources x402 | Coinbase announcement |
| 2025-09 | Coinbase and Cloudflare announce plan to create the x402 Foundation | joint announcements |
| 2025-09 | Stripe launches the Tempo chain with Paradigm | Stripe announcement |
| during 2025 | Stripe acquires Bridge ($1.1B) | Stripe announcement |
| 2025-10-14 | Visa launches Trusted Agent Protocol (TAP) | Visa announcement |
| 2025-12 | Anthropic's MCP joins the Linux Foundation Agentic AI Foundation | Linux Foundation |
| 2026-Q1 | Google launches AP2; OpenAI launches ACP (draft); ERC-8004 deploys to mainnet | respective announcements |
| 2026-03 | Circle launches Nanopayments (x402 v2 compatible) | Circle announcement |
| 2026-03-18 | Stripe + Tempo launch MPP; Tempo mainnet goes live | Stripe announcement |
| 2026-04-02 | x402 Foundation formally established under the Linux Foundation, 22 founding members | Linux Foundation press release |
| 2026-04 | Visa launches Visa CLI; AWS launches AgentCore Payments | respective announcements |
| 2026-04-29 | Circle launches Agent Stack (Agent Wallets + Marketplace + Circle CLI) | Circle announcement |
| 2026-05-13 | Base announces batch settlement for x402 (per-transaction cost below $0.0001) | public statement by Jesse Pollak (Base) |

## 3. Scale Data (multi-source point-in-time snapshots)

| As-of | Source | Data |
|---|---|---|
| 2026-04-29 | Circle official disclosure | trailing 30-day volume $24.24M; 99.8% of volume settled in USDC |
| end of 2026-04 | Coinbase Agent.market | ~69,000 active agents; 165M cumulative transactions; ~$50M cumulative volume |
| 2026-05-13 | AWS disclosure | over 169M cumulative payments; ~$48M cumulative volume; **95% of flow via Base** |

**Reference baseline**: global e-commerce market projected at $6.88 trillion for 2026; $50M ≈ 0.00073%.

**On-chain share dispute**: Solana Foundation, in the x402 Foundation press release (2026-04-02), claimed it was "driving nearly 65% of x402 transaction volume this year"; AWS's 2026-05 disclosure shows 95% of flow on Base. The two figures use different time windows and methodologies; both are listed for verification.

## 4. Conflicting-Data Comparison Table

Volume figures for the same protocol from different sources, side by side — the gap itself is one of the report's arguments:

| Time | Source | Data | Nature |
|---|---|---|---|
| 2026-03 | x402.org self-reported | 30-day $24M | self-reported by the project |
| 2026-03 | Allium Labs | 30-day ~$3M | third-party on-chain analysis |
| 2026-03 | Artemis (filtered) | 30-day ~$1.6M | after wash-trading filter |
| 2026-03 | Artemis daily average | 131K txns/day, $28K/day, $0.20 average ticket, ~half flagged as wash trading | third-party filtered |
| 2026-04-29 | Circle official | 30-day $24.24M | disclosed by a protocol beneficiary |
| 2026-05-13 | AWS | $48M cumulative | disclosed by an integrator |

Sources: CoinDesk (2026-03-11), Allium Labs, Artemis, Circle, AWS

## 5. Public Actions by Key Participants (all verifiable facts)

### 5.1 Full x402 Foundation roster (22 members, Linux Foundation press release 2026-04-02)

Governing bodies (3): Coinbase (protocol donor), Cloudflare (co-initiator), Stripe
Founding members (19): Adyen, AWS, American Express, Ampersend.ai, Base, Circle, Fiserv, Google (Google Cloud), KakaoPay, Mastercard, Merit Systems, Microsoft, Polygon Labs, PPRO, Shopify, Sierra, Solana Foundation, thirdweb, Visa

### 5.2 Stripe (the four fronts the report calls a "barbell hedge," each with a public source)

1. One of the three governing bodies of the x402 Foundation (2026-04-02, Linux Foundation)
2. Its own chain, Tempo (2025-09, with Paradigm)
3. Its own protocol, MPP (2026-03-18; routed through Stripe PaymentIntents API with merchant-grade fraud detection / tax / reporting; real commercial traction at projects like Browserbase)
4. Acquisition of Bridge to lock in the stablecoin rail ($1.1B, during 2025)

### 5.3 Circle

- USDC settles 99.8% of x402 volume (Circle disclosure, as of 2026-04-29)
- Products: Nanopayments (2026-03, sub-cent transfers), Agent Stack (2026-04-29), Circle Gateway
- Circle's stock price doubled in 2026 (ChainCatcher, 2026-03)

### 5.4 AWS / Google / Visa

- AWS: AgentCore Payments (2026-04), 200ms USDC settlement on Base
- Google: AP2 (2026-Q1); **AP2 has integrated x402 as its default stablecoin rail** (Coinbase protocol-donation statement, Google Cloud announcement)
- Visa: TAP (2025-10-14) + Visa CLI (2026-04); context: the Citrini Research report of 2026-02 wiped 5–7% off the market caps of Visa/Mastercard/AmEx in a single day

### 5.5 Base

- Carries ~95% of x402 flow (AWS disclosure, as of 2026-05-13)
- Announced batch settlement on 2026-05-13, cutting per-transaction cost below $0.0001

### 5.6 Ant International (Alipay+) AMP

- The AMP protocol supports $0.000001-scale micropayments and integrates KYA (agent identity) and dynamic risk ratings
- Plugs directly into 40+ mobile wallets, 1.8 billion users, 150 million merchants (vs. x402's 170+ native entities)
- Source: Ant International AMP announcement (see Part V of the report)

## 6. Protocol Relationship Facts

| Protocol pair | Relationship | Basis |
|---|---|---|
| AP2 ↔ x402 | Upstream/downstream: AP2 handles identity/authorization/compliance, x402 handles fund transfer; AP2 has integrated x402 as its default stablecoin rail | Coinbase donation statement, Google Cloud announcement, BitPinas/Paul Soliman analysis |
| MPP ↔ x402 | Direct competition: MPP runs on the Tempo chain with merchant-grade compliance and batch settlement | Stripe announcement, Cryptopolitan 2026-05, WorkOS analysis |
| TAP ↔ x402 | Complementary: TAP handles agent identity verification, x402 handles payment; no public interoperability spec yet | Visa announcement 2025-10-14 |
| ACP | Draft stage; focused on agent-to-agent commercial communication standards | OpenAI draft |

## 7. Related Protocols and Reference Data

**ERC-8004 (identity layer)**: mainnet deployment 2026-Q1; 8004scan.io shows 187,618+ registered agents across chains, while the Cambrian Q1 2026 report counts ~24K (including test data) — both figures listed side by side; registration costs only gas, with no staking gate and no verification mechanism.

**MCP**: open-sourced by Anthropic 2024-11, joined the Linux Foundation 2025-12; Anthropic is not on the x402 Foundation founding roster.

**Virtuals Protocol**: annualized fees $3.98M, market cap $571M, P/F 146.4x (point-in-time figures as cited in the report).

**NEAR Intents**: TVL $70.35M, 30-day DEX volume $1.62B, annualized fees $35.45M, 25 chains covered; initiators are humans and executors are human-operated solvers.

**Macro reference (as of 2026-04)**: USDC circulation ~$77B; USDC on-chain volume in 2025-Q4 exceeded $11.9T (mostly institutional settlement / DeFi / exchange-internal flows); total stablecoin market cap over $321B; global e-commerce market projected at $6.88T for 2026.

**Capital moves**: Dragonfly Capital's fourth fund at $650M (2025-12, AI×Crypto as a core direction); a16z proposed the KYA framework and invested in Catena Labs and the ERC-8004 ecosystem; a16z State of Crypto 2025 (2025-10) opened a Crypto×AI section; Delphi Digital's Year Ahead 2026 (2025-12) listed AI agents among its five key arenas.

## 8. Source List

**Official announcements**: Linux Foundation (2026-04-02), Coinbase x402 protocol-donation statement, Visa TAP (2025-10-14), Circle Nanopayments (2026-03), Circle Agent Stack (2026-04-29), Stripe MPP (2026-03-18), AWS AgentCore Payments (2026-04)

**Media coverage**: CoinDesk (2026-03-11), Cryptopolitan (2026-05), The Defiant (2026-04-02), PYMNTS (2026-04-03), BitPinas/Paul Soliman (2026-05), Crypto.news (2026-04-02), Blockonomi (2026-04-02), ChainCatcher (2026-03), COINOTAG (2026-04-03)

**Data platforms**: DefiLlama, Allium Labs, Artemis, 8004scan.io, Cambrian Q1 2026 report

**Technical documentation**: x402.org Ecosystem, Eco.com x402 deep dive (2026-05), Linux Foundation x402 Foundation governance docs, official docs and GitHub repos of the respective projects

---

*Heretic Research · Question everything. Alpha hides in doubt.*