preferred on
Points of Focus
- Solana handled 76% of x402 transactions over a four-week window.
- TRM Labs estimates only 0.6%-7.5% of x402 volume is agentic.
- Artemis, Solana’s own cited source, found roughly half of activity is gamed.
Solana’s official account said the network handled 76% of all x402 transactions between Aug. 22 and Sept. 19, citing Artemis data showing 23.2 million transactions against 3.39 million for the next-largest network, Ripple.
JUST IN: Solana handles 76% of all @x402 transactions.
23.2M in four weeks. The next-largest network did 3.39M. pic.twitter.com/OJv35U4kXF
— Solana (@solana) September 22, 2026
The claim aligns with a documented trend rather than an isolated data point, but it comes alongside new research questioning how much of x402’s overall activity represents genuine commerce.
X402 is a payment standard built on the long-dormant HTTP 402 status code, co-founded by Coinbase and Cloudflare in September 2025 and hosted by the Linux Foundation, letting a server request payment before delivering a resource so software, including AI agents, can pay for APIs and data without accounts or checkout flows.
Solana’s x402 share has been rising most of 2026
Solana’s share of x402 activity has swung considerably over the year, from as low as roughly 49.7% to as high as 88%-90% at various points before it surpassed Base in daily transaction counts on Aug. 25.
That shift is notable, given Base had a structural head start: Coinbase’s developer platform team built x402, and Coinbase’s layer-2 network was the protocol’s original home before Solana began closing the gap.
TRM Labs finds most x402 activity is not from AI agents
In a report published on its blog, TRM Labs examined $52.7 million across 198.9 million x402 settlement transactions on Base, Solana, and Polygon since May 2025. After screening out self-payments, large flows dominated by one or two payers, and sellers with fewer than 10 distinct buyers, that volume narrowed to $25.62 million of likely genuine commerce. Applied to that screened total, TRM’s two behavioral tests put the share that appears agentic between 0.6% and 7.5%.
TRM’s reasoning centers on a distinction any script can defeat by accident: “The assumption is that a true agent explores across multiple services and products, while an address repeating the same price behaves more like a script hitting one service over and over,” the report said. The company’s stricter screen additionally required facilitator-broadcast payments with varying amounts averaging under $1, sustained across months, plus either public agent registration or payments spread across multiple sellers, a bar most of the settlement data examined did not clear.
TRM describes x402 as the most widely adopted of several competing agent-payment standards, alongside Google’s AP2 and OpenAI and Stripe’s ACP, and notes it is also the one whose full volume is visible onchain. This transparency exposes it to more scrutiny than standards whose activity isn’t equally auditable.
Artemis, Solana’s cited source, has flagged the same problem
TRM’s findings are not an isolated challenge to how x402 volume gets read.
Artemis’ research, the same data provider Solana cited for its 76% figure, separately found that roughly half of observed x402 transactions network-wide reflect artificial or gamed activity, including test transactions and incentivized volume rather than genuine commerce.
That context changes what a transaction-count leaderboard actually measures. Solana’s 76% share likely reflects where x402 activity concentrates today.
Still, the underlying protocol’s data providers estimate that most of that activity, on every chain it runs on, isn’t the AI agent commerce the metric is generally read to represent.
Whether Solana’s lead holds once genuine agentic use overtakes scripted and incentivized traffic is a separate question the current numbers cannot answer.
Share
Most read articles
