---
title: "The Four Pillars x402 Report Maps the Agent Market Solana Already Runs"
description: "Four Pillars' 131-page x402 report models annual agent payment volume reaching as high as $1.9T by 2028, with roughly 96% of it landing in purchases an agent makes without a human approving anything. Solana already handles about 70% of x402 protocol volume, so the report's headline scenario is running today on rails that shipped this year, from Pay.sh to Ramp's 70,000 business wallets."
author: "Mike Hale"
published: 2026-09-15T16:02:00.000Z
updated: 2026-09-16T18:43:39.181Z
category: "Payments"
tags: ["Four Pillars"]
canonical: https://stacksmith.xyz/articles/solana-x402-four-pillars-report
source: "Stacksmith"
---

# The Four Pillars x402 Report Maps the Agent Market Solana Already Runs

Four Pillars researcher Junhyeok Yoo has published [All about x402](https://research.4pillars.io/en/research/all-about-x402), a 131-page report that puts annual x402 payment volume as high as $1.9T inside AI-intermediated B2B commerce by 2028. Roughly 96% of that volume is agents buying APIs, data, inference, and compute per request from sellers they never signed a contract with.

That is the market Solana is already winning, taking [roughly 70% of x402 protocol volume over the 30 days to August 24, 2026](https://cryptobriefing.com/solana-x402-market-share-dominance/).

## What the report models

The report runs three adoption scenarios that land almost 10x apart, from $186.45B under gradual adoption to $1.9161T under high adoption, with $686.47B as the baseline case.

Yoo splits agent commerce into three tiers by how much purchasing authority a person hands over:

- **T1**: The agent finds and compares products and a human approves the payment
- **T2**: The user sets a budget and purchase conditions and the agent buys within them
- **T3**: The user gives a goal and a budget, and the agent decides what to buy, from whom, and in what order

He then takes Gartner's forecast of $15 trillion in AI-intermediated B2B spending in 2028 as a fixed total and works down from it. T3's share of that total is set first, at 20%, 25%, or 35% depending on the scenario, and what remains is split between T1 and T2. Each tier then gets its own x402 processing share. T1 gets 0%, on the reasoning that a human clicking a checkout button will keep using a card. T2 gets a share of stablecoin payments, then a share of x402 within that. T3 gets the largest share by a wide margin.

| Scenario | T3 share of B2B spend | x402 payment volume | Share of 2025 global B2B payments |
| --- | --- | --- | --- |
| Gradual | 20% | $186.45B | 0.10% |
| Rapid | 25% | $686.47B | 0.37% |
| High | 35% | $1.9161T | 1.03% |

T3 accounts for roughly 96% of the total in every scenario.

## Headless merchant endpoints are the first real market

Agentic commerce gets proven out on cheap, high-frequency digital resources long before it touches anything expensive. That is the report's sharpest call, and Solana already has the services to test it against.

Yoo calls those sellers headless merchant endpoints. They publish machine-readable prices, input specs, and terms, and they skip the product pages and checkout flows built for humans. His argument for why they come first is economic. Individual failures are cheap, the transactions repeat often enough to generate real operating data, and the cost of a human approving each one would exceed the value of the thing being bought.

He backs the cheap-and-frequent part with a fixed Allium snapshot taken on August 4, 2026, covering $52.47M of cumulative payment volume across 233M transactions. Payments under $0.10 are 86.78% of x402 transaction count but only 9.39% of dollar volume. Payments of $10 or more are 0.15% of count and 35.72% of volume. The weighted average payment works out to $0.2252.

On Solana, that supply side arrived this year.

[Pay.sh](https://solana.com/news/solana-foundation-launches-pay-sh-in-collaboration-with-google-cloud), launched by the Solana Foundation with Google Cloud in May 2026, is the clearest example. It puts BigQuery, Gemini, and Cloud Run behind per-request stablecoin payments, alongside 50+ community API facilitators including PayAI, Crossmint, Corbits, and Merit Systems. An agent pays per call with no account, no API key, and no subscription. That is the report's headless merchant endpoint described almost word for word, running on Solana.

QuickNode puts its own RPC endpoints behind [x402 paywalls that take USDC on Solana](https://www.quicknode.com/docs/build-with-ai/x402-payments), so an agent can query chain data with nothing but a funded wallet. Its [`@quicknode/x402-solana` library](https://github.com/quiknode-labs/x402-solana) does the payment negotiation, and its nanopayment tier skips per-request settlement entirely by batching through Circle's Gateway. Nobody signs up for anything. The agent shows up with a wallet and buys one RPC call.

On the demand side, [Solana Compass names BlockRunAI](https://solanacompass.com/news/solana-claims-top-spot-on-x402-as-ai-agents-settle-33m-usdc-in-a-single-week) as a leading volume contributor, an AI-native trading service whose agents buy market data and analytics as they go.

## Where the money actually goes

x402 has no protocol fee, so there is no pool of revenue to split. Five layers each monetize a different unit instead, and every one of them has a Solana number attached to it today.

| Layer | What it earns on | Report's 2028 estimate | Solana today |
| --- | --- | --- | --- |
| Stablecoin issuers | Net new circulating balance | $25.52M to $262.38M in reserve income | [$11B+ in stablecoins circulating](https://solana.com/x402/what-is-x402), and over 99% of x402 volume settles in USDC |
| Blockchains | Settlement transactions | Modeled on Base at $0.0057 per settlement | [$0.00025 per transaction](https://solana.com/x402/what-is-x402) with 400ms finality |
| Facilitators | Verification and settlement | $0.76M to $7.75M on fees, or $190M to $1.9B managed | [Corbits, PayAI, Coinbase, and ACK](https://solana.com/docs/payments/agentic-payments/intro-to-x402) all support Solana, with PayAI currently absorbing transaction fees |
| Wallets | Funding and settlement routes | $37.29M to $766.43M in ancillary fees | [Coinbase Agentic Wallets are built on x402 and cover Solana](https://www.crossmint.com/learn/agent-wallets-compared), with Crossmint and Privy covering Solana too |
| Discovery and routing | Supplier selection and order allocation | $55.93M to $574.82M | [Coinbase's x402 Bazaar](https://docs.cdp.coinbase.com/x402/bazaar) accepts Solana network identifiers (the size of its Solana catalog is unverified) |

Solana's cost structure changes the math in the chain row. To make the Base figure economic, the report assumes 500 payment events get batched into a single onchain transaction. It cannot apply that formula to Solana, because under the `exact` scheme each accepted payment already maps to one onchain transaction at roughly one twentieth the cost. The report models a separate Solana path instead, covering both that per-payment settlement and a draft batch-settlement spec that runs through prefunded escrow channels.

The facilitator row carries the widest spread in the table, and the two figures describe different businesses built on identical payment volume. Charging per onchain settlement tops out around $7.75M and shrinks as batching improves, while charging per payment event for policy checks, compliance, and audit records reaches $1.9B. The money sits in the operations a seller will not build in-house, not in submitting the settlement.

## Solana runs most x402 payments today

Solana took the daily x402 transaction lead from Base in [late August 2026, for the first time in six months](https://cryptobriefing.com/solana-x402-market-share-dominance/), and has held it since.

Over the 30 days ending August 24, [Solana Compass put x402 at 75.41 million transactions and $24.24 million in volume across all chains, with 94,060 active buyers and 22,000 active sellers](https://solanacompass.com/news/solana-claims-top-spot-on-x402-as-ai-agents-settle-33m-usdc-in-a-single-week), citing the x402.org dashboard. Solana took roughly 70% of that monthly volume, and agents settled $3.3 million in USDC on Solana in the week ending August 24.

Most of that came from infrastructure that shipped this year.

- **May 2026.** The Solana Foundation and Google Cloud launch Pay.sh.
- **June 2026.** AWS adds AI traffic monetization to WAF Bot Control, with Coinbase's facilitator settling USDC on both Base and Solana.
- **July 14, 2026.** The x402 Foundation launches under the Linux Foundation with the Solana Foundation as one of 17 Premier members, alongside Visa, Mastercard, AWS, Google, and Stripe.
- **August 20, 2026.** [Ramp onboards more than 70,000 business customers](https://solanacompass.com/news/solana-claims-top-spot-on-x402-as-ai-agents-settle-33m-usdc-in-a-single-week) to x402 agent wallets, embedding pre-authorized USDC budgets into corporate spend software.
- **August 25, 2026.** Solana passes Base in daily x402 transaction count.

Ramp is the report's T2 configuration arriving in production. A company sets the budget and conditions, the agent spends inside them with an audit trail, and every one of those 70,000 businesses becomes a potential buyer for anyone selling an endpoint. It landed days before Solana's record week.

[Solana payment channels](https://stacksmith.xyz/articles/solana-payment-channels) fit on top of all this. They let an agent escrow a ceiling, meter spending offchain with signed messages, and settle once, which maps onto the batching schemes the report describes and pushes the per-payment cost far below even Solana's base fee.

## How much of the volume is real?

Weekly x402 transactions peaked near 6.8 million in December 2025 and fell to [under 510,000 by the week ending February 9, 2026](https://phemex.com/news/article/solanas-x402-market-share-drops-amidst-rising-polygon-competition-60911). Solana's own share has [swung from 49.7% to 88% inside 2026, and cumulative all-chain transaction estimates sit between 120 million and 187 million depending on the tracker](https://cryptobriefing.com/solana-x402-market-share-dominance/). A spread that wide says more about the measurement than about the market.

Yoo opens the report by saying he deliberately avoided explaining x402's growth through transaction counts, because growing volume does not distinguish sustained demand from what he calls *Gamed Transactions*. The model rests on stacked assumptions instead.

The high-adoption scenario multiplies a 35% T3 share by a 50% open request-level payment share by a 70% x402 share within that, and each of those three numbers is the author's own estimate rather than an observation. The gradual case runs the same chain at 20%, 20%, and 30%, which is most of where the 10x spread comes from. He labels them as such, which is more than most market-sizing work manages.

Jesse Chung of Seal puts the supply-side version of the same problem more bluntly in the report's contributor section. Half the services listed in the x402 directory today, he says, are failing or gone, because micropayment revenue is too thin to justify holding a quality bar on any individual provider.

## What to take from it

The T1/T2/T3 frame is what survives the rest of the report. It gives you a way to ask whether a given agent transaction needs a new payment rail at all, and the honest answer is usually no until the agent is picking suppliers on its own. On that kind of transaction, Solana's $0.00025 fees and 400ms finality decide whether the seller has a business at all.

Two conditions govern whether that market forms, and both sit on the supply side. Sellers have to keep endpoints alive long enough to earn repeat calls, the quality problem Chung names. Discovery has to get good enough that an agent can choose between five providers of the same data on price and reliability, which is where the pricing power settles once payment execution itself is standard. Solana has the volume lead and the distribution deals. Whether that converts into the T3 market the report models comes down to the services behind those endpoints holding their ground.
