a16z and Artemis question hype around AI agent payments and x402 transaction data

Artemis and a16z analysts have put out data to temper the enthusiasm around x402 payments and agentic AI commerce, calling attention to data from stakeholders in the AI agent space that appear to inflate activity and growth rate. 

Data analysis compiled and shared by a16z crypto partner, Noah Levine, turned the spotlight on the apparent inaccuracy in the x402.org data that Bloomberg featured in a March 7 report about stablecoin firms “building payments systems for a future with autonomous AI agents that transact in stablecoins.”

Artemis supported the sentiment by recalling an earlier analysis by its in-house analyst, Lucas, who “built a wash trading filter for x402 volume, flagging wallets that repeatedly transacted with themselves or cycled funds between addresses.”

Artemis: Most of the x402 numbers circulating are noise

Artemis Analytics’ Lucas was not buying the hype when headlines claimed that x402 transactions accounted for 19% of all the transactions on the Base network, the Ethereum L2 chain developed by the Coinbase exchange. 

In the December 2025 report, he noted how x402 activity started picking up steam at the end of October, while pointing out how Base processed about 89% of the over $335 million volume and 86% of 92 million transactions. 

Even though Solana and BNB Chain have also started to get in on the agentic AI payment action, analysts continue to insist that the actual numbers, taking away “gamed volume,” which is artificial stats pumped up by wash trading and self-dealing, the transaction stats aren’t just that impressive. 

Lucas claimed that, as of December, about 48% of transactions and 81% of transaction volume were directly linked to gamed volume, and the numbers could even be higher. 

There’s also the problem of maintaining consistency across different monitoring platforms. While x402.org reported 30-day transaction volume at $24 million, Allium Labs barely had it over $3 million over the same timeframe. That same metric comes in under $2 million on Artemis, which Levine attributes to “how early-stage even the measurement infrastructure is.”

It’s not all gamed volume though

While the inflation of trading volumes grabs headlines, not everything is an overt attempt to game volume. Some of it is actually volume from legitimate product testing by development teams. 

As Lucas highlighted in December, “the gradual decline in transaction size” is a positive sign of x402 payments delivering in the long-term direction.

Levine agrees, pointing to the “infrastructure being built around” the x402 payments ecosystem. He pointed to integrations by Stripe, Cloudflare, Google, and Vercel as signs of progress. 

Erik Reppel, the head of engineering at Coinbase Developer Platform, which is responsible for developing x402, belongs to the same school of thought. Citing Artemis data, Reppel insisted that the 78,000 transactions from the day before represent about “108% growth compounding monthly.” 

The positive sentiment ignores the drop-off from the highs of December, when daily volumes consistently surpassed two million, with real volume staying above one million for most of the month. Instead, he’s comparing it to one year ago, when x402 transactions combined for a grand total of 158 transactions for the entire week.

As for Lucas, his optimism hinges on x402 ticking off four critical boxes to reach the adoption rate that tech giants like Stripe, Cloudflare, Google, and Vercel are betting on agentic commerce to eventually achieve. 

The ERC-8004 agent identity standard is already a step in the right direction to achieving trust, which is critical in verifying a server’s reliability. 

For now, the other three criteria, “discovery,” “merchant adoption,” and “more standards,” remain growth areas that will contribute to building out “one of the clearest examples of how crypto, TradFi, and tech are merging” into one complete payment ecosystem.

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.