Market — ONE's own run of Sold for software

The decision

ONE plays in payments for agents. The rail already exists: x402 turns HTTP "Payment Required" into a stablecoin payment with no account and no API key, and Stripe, Coinbase, Google, Visa and Mastercard all back it or build next to it. What is missing is sellers. ONE takes the seller's side of that gap: anyone, a person or an agent, becomes a seller with one key, keeps that key on their own machine, settles in USDC on Sui, and ONE takes nothing from the sale. The primary comparison is Stripe and Coinbase. The buyer we write for is Brian: a smart, ambitious 21-year-old, who is equally a smart young AI agent.

The numbers

All from agenteconomy.to, as of 2026-11-07/08, unless marked. These are measured counts of x402 payments. They are not forecasts.

• Lifetime: about 198M payments and $42.6M settled. Re-checked 2026-11-11: 199.1M and $42.7M, so the picture has not moved.
• The peak: $20.54M in November 2025, during the PING pay-to-mint speculation. December 2025 was the peak in count, 54.25M payments. More than half of all payments ever came in those two months.
• The floor: from February 2026 the count settled at 4–6M a month. August and September rose above that (17.86M, 13.00M) at $0.02–$0.03 a payment, then fell back. VERIFY: what drove them.
• The latest full month: October 2026, 5.09M payments, $0.27M, $0.05 average. That is down 98.7% in dollars from November 2025 (computed).
• Organic sellers: about 78 (x402stats, Sep 2026; at least $100 revenue and 3 buyers). Median seller revenue about $0.04. A few wallets dominate (Gini above 0.98). VERIFY: I could not re-check this one.

The series, one line a month (payments / dollars / average):

• Oct 2025: 4.09M / $4.22M / $1.03
• Nov 2025: 52.59M / $20.54M / $0.39
• Dec 2025: 54.25M / $9.41M / $0.17
• Jan 2026: 19.19M / $1.20M / $0.06
• Feb 2026: 4.07M / $1.62M / $0.40
• Mar 2026: 5.09M / $1.84M / $0.36
• Apr 2026: 4.60M / $1.61M / $0.35
• May 2026: 5.34M / $0.18M / $0.03
• Jun 2026: 4.78M / $0.27M / $0.06
• Jul 2026: 6.01M / $0.24M / $0.04
• Aug 2026: 17.86M / $0.44M / $0.02
• Sep 2026: 13.00M / $0.42M / $0.03
• Oct 2026: 5.09M / $0.27M / $0.05
• Nov 2026, to about the 7th: 1.62M / $0.27M / $0.17 (part month)

The insight holds in the data. Count held a floor; dollars did not come back. Only about 15% of settlements look like real outside demand (an academic paper; VERIFY). Directories list about 1,270 services, up 370–390 a month (Nov 2026, VERIFY), yet seller revenue is cents. The bottleneck is sellers who earn, discovery and trust. It is not the protocol.

Not used: the "161k monthly active agents" claim. It conflicts with about 3,000 paying agents a day in Sep 2026.

Who else sells to our buyer

The six tiers, with real names. Rival details come from the brief.

1. Incumbent suite: Stripe (MPP with Tempo, ACP with OpenAI). It already owns the payments budget. MPP meters sessions in stablecoins and cards (live Mar 2026); ACP is agent checkout in chat, by card. Weak against ONE: an account and a company sit in the flow, and MPP publishes no volume. ONE's seller keeps their own key.
2. Category leader: Coinbase (x402, facilitators, CDP). It launched x402 in May 2025 and is named first. Over 97% of on-chain agent transactions used x402 in some periods (VERIFY). Weak against ONE: it sells the rail, and few sellers earn on it. Whether it gives sellers a storefront or payment link is VERIFY.
3. Card networks and platforms: Visa, Mastercard, Google (AP2, UCP). Visa has Trusted Agent Protocol and Intelligent Commerce; Mastercard has Agent Pay. AP2 proves a human approved the spend and settles crypto through x402. UCP runs catalog to checkout on Google surfaces with Shopify, Walmart and Target. Weak against ONE: built for large merchants, not a person or agent selling a $0.05 call.
4. Cheap point tools: facilitators, gateways, L402. QuickNode gives 1M free x402 requests a month. Cloudflare has a content-monetisation gateway. L402 (Lightning Labs) is Bitcoin pay-per-request and publishes no volume. Weak against ONE: each does one job on one chain.
5. Open source and self-host. x402 is open under the Linux Foundation with no protocol fee, so anyone can run a facilitator. Weak against ONE: keys, chains, gas and discovery become your job.
6. Internal build, and doing nothing. A team puts a paywall in front of its own API; most do neither. Doing nothing is the biggest rival: about 78 organic sellers is the whole field. Beat it with one link and one first sale.

The grid

A. Competitive set. Observe: Stripe and Coinbase lead, and 40+ firms sit in the x402 Foundation. Analyze: they sell the rail, and the median seller earns $0.04. Respond: "one key makes you a seller, the key stays yours, ONE takes nothing". A rival cannot copy that by editing a headline.

B. Market forces. Observe: every major player backs x402; dollars collapsed after the speculation while count held; analysts forecast growth. Analyze: the speculation burned off, leaving real use at small prices, with supply, discovery and trust still open. Respond: ride the first, brace against the second, treat the third as a hope. Analyst scenario, not fact: 2027 at 15–40M payments and $5–25M a month.

C. Channels. Observe: sellers and buyers meet in x402 directories, dashboards such as x402stats, developer writing and MCP tool registries. Analyze: directories are growing and crowded, and few sellers there can be found or trusted. Respond: start with a payment link and a chat component that takes payment, then list in directories. Read our own rank later; never borrow one.

What sells

Priced $0.01–$0.50 a call, instant, discoverable. Tier one: specialised or live data (company checks, sanctions, market and on-chain data); verification and trust (claim checks, risk scores, reputation); metered AI and tools (extraction, code review, MCP tools); compute and actions (scrape, PDF, email, GPU). Tier two: premium content, proxy and aggregation layers, discovery and routing. Weak: wrapped free data, high-ticket one-offs, anything needing an account.

ONE sells the key, payment links and chat components that take payment (products, bookings, subscriptions, courses), and Sold, the $100 course. A ONE seller can list a metered tool or data call from the tier-one list. VERIFY which of these are live today.

Still open

• The x402 protocol rail does not pay the workspace yet. It reads ONE's own treasury, so an x402 payment does not reach the seller (text/one.md red list, item 8). The storefront rail is correct: it pays the seller's wallet with ONE never in the flow. Fix before saying "straight to your wallet" of x402.
• Cards. The red list says Stripe is in test mode in production and the storefront is on in 2 of 68 workspaces. VERIFY today's state before saying "cards too".
• Mainnet. text/one.md says ONE launches on mainnet. VERIFY that is the state today.
• Chain. x402 volume sits mostly on Base and Solana (Solana about 76% of payments in one Sep 2026 window; trackers differ). ONE settles on Sui and does not follow. text/one.md says links take crypto on six chains; VERIFY which are live.
• Two trackers disagree. Facilitators report $2.85–2.93M over 30 days (Nov 2026, VERIFY); agenteconomy.to reports $0.27M for October. We use agenteconomy.to.
• The word "free". text/one.md bars "0% fees" because the card network takes its own fee. Say "ONE takes nothing from a sale".
• Not measured: ONE's own share. There is no seller count yet.

Do it yourself

1. Write your market in one sentence and name the rail your buyer already uses.
2. Pull one dated series from one source. Put each figure beside its date and source. Label forecasts.
3. Put your real rivals in the six tiers, each with a weakness you can prove.
4. Fill the grid for rivals, forces and channels. Test your insight against the data before you believe it.
5. List what is still open, including what you cannot honestly say yet. Mark it VERIFY and clear the red items before the ad goes out.