Wes Knipe

Wes Knipe: analytics consultant with a background in economics, financial advice and early-stage software. Current research: how AI agents pay for things, and how much of that market is real.

Wes Knipe

I'm Wes. I studied quantitative economics at Cal Poly, started out at a couple of financial advisory firms, and then moved to early-stage software to see how products and companies are actually built. These days I consult on analytics projects and build research tools in my spare time.

About · Résumé (PDF) · wes@wknipe.com

Current research

How AI agents pay for things

A new web standard, x402, lets software pay per request in digital dollars: fractions of a cent, settled on a public ledger. Like early crypto exchange volume, much of the reported activity is fake. I measure the real market: how much is genuine demand, who buys and comes back, what things cost, and whether sellers deliver.

$134k

settled per week

Everything paid through x402 on the Base network in the week of 21 Sep, before any cleaning.

→

$73k

genuine per week

What is left (55%) after removing payments that loop back to the payer, tests, and shops with a single customer.

→

8,176

genuine buyers

Distinct paying wallets that week. A wallet is an account, not necessarily one person or firm.

Data through 2026-09-27 · updates Mondays

x402 in 60 seconds →

What's here

• Market dashboardWeekly volume, before and after cleaning, and where the money goes.→
• Price compsWhat comparable paid APIs charge per call, like comps in a valuation.→
• Endpoint statusA daily check that each listed service answers and charges what it says.→
• AI-access censusWhich of the top 1,000 websites block AI crawlers, and which ask to be paid.→
• Which models buy wellHow sensibly different AI models spend money when asked to buy.→
• Weekly reportThe numbers in one short note each Monday, with RSS.→
• Prediction ledgerForecasts written down and timestamped before the data comes in.→

Everything else

The rest of the data and tools, on the same index.

DataMarketHow much x402 volume is real, where it goes, and who carries it. Start here.Open →DataSellersEvery seller with 5+ genuine buyers, whether those buyers come back, and how much rides on the biggest one.Open →DataBuyersHow many real buyers pay each week, how many come back, and how little most of them spend.Open →DataPricesWhat listings ask per call, how prices move, and every posted-price change.Open →ToolBest executionA demo: say what your agent needs and see three endpoints, the cheapest that answers, the best value and the most used.Open →ToolPrice compsDescribe an API and see what comparable listings charge, and whether anyone pays them.Open →ToolEndpoint statusDoes the endpoint answer, and does it charge what it says? Checked every day, politely.Open →ToolAI-policy checkerPaste a domain and see what it tells AI crawlers: robots.txt per bot, content signals, licences, 402.Open →DataState of AI accessHow much of the top 1,000 websites is closed to AI crawlers, and who is trying to charge instead.Open →BenchmarkAgent benchmarkWhich AI models spend money sensibly, and what 1,000 of their decisions cost.Open →APIAPI & dataFree JSON behind every chart, CSVs, paid x402 endpoints for bulk history and routing, and an MCP server.Open →BriefState of x402, one pageThe numbers people ask for first, on one printable page you can cite.Open →