Latest TenOneTen News: Money at Machine Speed

February 16, 2026

Money at Machine Speed

David Waxman

Last Tuesday, Coinbase launched the first crypto wallet infrastructure built specifically for AI agents.¹ Hours later, Stripe announced it was adopting the same underlying protocol.² Two of the most important payments companies in the world shipped agent payment products on the same day.

We've spent the last few months researching this space at TenOneTen Ventures, and we think most people are underestimating how fast it's moving.

Agents Have a Money Problem

We've gotten good at building AI agents that can reason, plan, and coordinate with each other. Impressively good. But there's a comically basic thing most agents can't do: pay for stuff.

An agent finds the API it needs but can't swipe a credit card. It negotiates with another agent but has no way to settle. It identifies cheaper cloud compute but can't buy it. Every time money enters the picture, the agent stops and waits for a human.

This is a little like building a fleet of self-driving trucks and then requiring a person to hand cash to every toll booth operator. The autonomy is real until the transaction isn't.

McKinsey projects $1 trillion in US retail agentic commerce by 2030, with global projections reaching $3–5 trillion.³ ChatGPT already handles over 50 million daily shopping queries.⁴ The demand is there. The plumbing to actually move money between machines is what's been missing.

Until, apparently, last week.

What Agents Actually Need to Buy

Before we get to who's building what, it's worth stopping on a point that most coverage of this space glosses over: the things agents need to pay for mostly don't look like anything humans buy.

Some of it is familiar. An agent booking a flight or ordering supplies — that's just e-commerce with a different buyer. But the bigger and more interesting category is transactions that have no human equivalent at all.

An agent needs a real-time pricing lookup from another agent's database. That's a fraction of a cent, settled instantly. It needs ten seconds of GPU time to run an inference. It needs to check a credit score, pull weather data for a logistics route, or access a proprietary dataset to answer a question. Each of these is a microtransaction — maybe $0.001, maybe $0.01 — happening thousands of times per minute across millions of agents.

There's no shopping cart here. No checkout flow. No "add to cart." It's closer to a stock exchange than a store. Agents are buying compute, data, and capabilities from each other in real time, programmatically, at a scale and speed that human-designed payment systems were never built to handle.

This is also why — and I'll admit my partners had to convince me of this — crypto ends up being necessary. My initial instinct was that traditional rails would be fine. But think about it: Visa's minimum transaction fees would cost more than the thing being purchased. You can't run a $0.001 payment through a system designed for $4 coffees. And smart contracts let agents enforce terms programmatically — escrow, SLAs, automatic refunds — without a human mediating every dispute. At machine speed and machine scale, you need machine-native money.

What Just Happened

Coinbase's new Agentic Wallets are built on something called the x402 protocol, which repurposes an old, largely unused HTTP status code — 402, "Payment Required" — for machine-to-machine payments. An agent hits an API, gets a 402 response with a price, pays in USDC, and gets access. One round trip. The protocol has already processed over 140 million transactions, at a cost of $0.001 per transaction.⁵

Stripe didn't build a competing protocol. They adopted x402.² It’s very telling and, I think, hugely positive, that they decided to join the team rather than going solo. We’ll get into this more below.

They're not alone. OpenAI launched its Agentic Commerce Protocol and instant checkout inside ChatGPT.**⁵ Google unveiled its Universal Commerce Protocol in January alongside Shopify — a common language for agents to discover and buy products, co-developed with Etsy, Wayfair, Target, Walmart, and endorsed by over 20 others including Mastercard, American Express, and Stripe.⁶ PayPal is restructuring around this, with an "Agent Ready" platform, a Microsoft Copilot integration, an OpenAI partnership for ChatGPT checkout, and an acquisition of Cymbio, a Tel Aviv e-commerce AI company.⁷ And Catena Labs, founded by Circle co-founder Sean Neville, raised $18 million from a16z to build what is essentially a bank for AI agents.⁸

This all happened in the last few weeks.

The Startup Layer

Below the giants, there's a wave of startups filling gaps that the big companies either can't or won't address.

Natural raised a $9.8 million seed to build B2B payment workflows for agents in logistics, procurement, and construction.⁹ Not consumer checkout — the kind of enterprise payment execution where an agent sources materials, negotiates terms, and pays a subcontractor. Their CEO's line sticks with me: "The method by which payments are executed is shifting from human execution to agentic execution. There's never been a time when so much payment volume will change medium." He's right. The sheer volume of transactions that will shift from human-initiated to machine-initiated is staggering, and most existing payment systems weren't designed for it.

Nevermined built a live product that works across every major protocol.¹⁰ In a world with seven competing standards (I'll get to that), being the Switzerland of agent payments is an interesting position. KITE AI raised $33 million to build a dedicated blockchain for agent payments.¹¹ PayAI Network has processed 42,000 transactions since launching last summer.

The common thread: these companies aren't building chatbots or AI assistants. They're building financial infrastructure. And the funding reflects it.

Seven Protocols Walk Into a Bar

Here's the part that makes me a little nervous. Count the competing standards: x402, ACP, AP2, UCP, A2A, MCP, ERC-8004. That's seven protocols with real traction, backed by different combinations of major companies. If you've been around technology long enough, you know how this goes. Fragmentation slows adoption. If every agent needs to speak seven different payment languages, the whole system is slower, more expensive, and harder to build on.

Our read: x402 wins the machine-to-machine layer. It's the simplest, cheapest, and now has Coinbase, Cloudflare, and Stripe behind it. The consumer-facing layer is a two-horse race between Google's UCP and OpenAI's ACP. Google controls shopping intent and the merchant network — Shopify, Walmart, Target, Etsy. OpenAI has a huge and fast-growing conversational interface, and a head start on turning chat into checkout. PayPal is making the smartest quiet bet in the room: plugging into both, becoming the payment rail that works regardless of which storefront wins. x402 is the plumbing. UCP and ACP are the storefronts. History says we consolidate to two or three standards. That's where we'd place the bet.

What We're Watching For

We look at early-stage B2B companies for a living, so we're naturally drawn to the infrastructure layer. A few things we think are underappreciated:

Agent-to-agent commerce is the bigger market. Most coverage focuses on agents shopping for humans — buying flights, ordering supplies. That's real, but it's the smaller opportunity. The microtransaction layer we described above — agents buying data, compute, and capabilities from each other — is a fundamentally new category of commerce. It doesn't map to any existing market, which means the companies that figure it out won't be competing with Visa or Shopify. They'll be building something that didn't exist before.

Identity is the sleeper problem. Everyone's focused on payments, but when agents are spending money autonomously, you need to know which agents to trust. Know Your Agent, reputation scoring, cryptographic verification — this is mostly unsolved and matters enormously.

The inevitable security failures. Some agents are going to spend money they shouldn't, in a public and embarrassing way. When that happens, the companies with real safety controls — session caps, transaction limits, human escalation — will separate from those flying by the seat of their pants.

Regulation. The SEC and OCC haven't said a word about agent-initiated financial transactions. They will. Regulated players such as Catena Labs, mentioned above, will benefit. Everyone else will have to adapt quickly.

The last time payments infrastructure got fundamentally rebuilt — when Stripe made it possible to accept payments with seven lines of code — it created hundreds of billions in value. What's happening now is a similar moment, except the users are machines, the transaction volume will be orders of magnitude higher, and we're still in the first inning.

We plan to keep writing about this. If you find it useful, subscribe and we'll show up in your inbox.

David is a partner at TenOneTen Ventures, an early-stage B2B venture fund. He writes about AI, agents, markets, and whatever he feels like talking about that day.