A year ago, machine-to-machine payments were just a concept. Now, four competing payment architectures are live, backed by Coinbase, Stripe, Google, Visa, and American Express. AI agents have settled over $730 million across 176 million transactions, and traditional giants have invested over $8 billion in acquisitions to secure their position in this new payment stack.
This report, developed by Keyrock in collaboration with Coinbase and Tempo, examines how this payment stack is assembled, whether the economic model is viable, and what obstacles it faces.
Protocols are not competing, but stacking.
In September 2024, if you wanted an AI agent to make payments, there was essentially only one insecure option. Twelve months later, four architectures already exist, backed by some of the largest companies in the tech industry.
Coinbase built x402, a crypto-native protocol that turns stablecoin wallets into universal API keys. Stripe and Tempo launched MPP, a payment-agnostic standard that processes credit cards, cryptocurrencies, and Lightning Network payments through a single HTTP flow. Google assembled AP2, an authorization layer that enables users to delegate payment permissions to agents via cryptographic authorization. Visa expanded its existing card network to offer AI-ready tokenized credentials.
What most reports overlook is that these four solutions are not purely competitive. There is indeed overlap at the protocol layer, but the more important dynamic is that they are coming together to form a payment stack. The right question isn’t “Which protocol will win?” but rather, “Which companies control the most layers and therefore capture the most value?”
A $0.30 wall
Among the 176 million x402 payments to date, the median transaction amount ranged between $0.01 and $0.10, with 76% of transactions below the $0.30 minimum card processing fee threshold. This figure nearly explains why traditional payment channels cannot serve this market: a fixed processing fee of around $0.30 per transaction makes micropayments unprofitable. An agent paying 3 cents for a weather API call cannot route it through Visa.
Layer 2 stablecoin settlement cost: $0.0001. For agents, this means blockchain infrastructure is essential.
Single stablecoin dominance
Of those 176 million payments, 98.6% were settled in USDC. Stablecoins have essentially become the default settlement layer for machine-to-machine commerce; they are the only tools capable of handling microtransactions without collapsing the economic model.
This concentration is both a validation and a vulnerability. It affirms Circle’s position as the default settlement asset, but also means the entire Agent payments ecosystem relies on a single stablecoin issuer’s reserve management, regulatory standing, and technological infrastructure. No one in the industry openly discusses this. We believe they should.
Vertical integration race
Coinbase and Stripe each cover five of the six layers of the emerging payments stack. Coinbase controls the settlement layer (Base), wallets (Agentic Wallets), routing (internal infrastructure), payment protocol (x402), and governance (as an AP2 partner). Stripe mirrors this layout through Tempo (settlement), Privy (wallets), Bridge (routing, acquired for $1.1 billion), MPP (protocol), and its compliance infrastructure.
Over the past 12 months, traditional giants have invested over $8 billion in acquisitions to fill gaps in their payment stack coverage. Capital One acquired Brex for $5.15 billion, Mastercard purchased BVNK for $1.8 billion, and Stripe acquired Bridge. These are all infrastructure consolidation moves by companies that view machine payments as a natural expansion of their core business.
From Bot Activities to Agent Commerce
The machine economy has arrived. It just hasn’t started doing business yet. But the signals are clear: AI agents account for 37% of all Safe transactions on the Gnosis Chain, peaking above 75%. Coinbase has deployed tens of thousands of agents with built-in guardrails. Over 104,000 agents are registered across 15 or more directories and registries.
The shift from extractive bot activity to productive agent commerce is underway. The payment infrastructure examined in this report is precisely what enables this transformation.
Regulation is a constraint.
MiCA, the GENIUS Act, and the EU AI Act will all reach enforcement stages within weeks of each other in mid-2026. None of them address autonomous machine-to-machine transactions. This is not a future issue; it is a current one, unfolding in real time with real capital at stake.
What happens next?
The market is moving toward greater agent autonomy, but we believe the pace will not be set by technology—technology is essentially ready. The pace will be set by the trust infrastructure needed to make all of this secure. The vision of complete permissionlessness is theoretically appealing, but it assumes a level of AI reliability that does not yet exist. Until agents stop hallucinating, they should not be granted unsupervised access to user funds.
We believe the bottom-up argument is the most compelling framework for understanding what’s next. The crypto rail has already won the case for micropayments. As transaction volumes grow and trust infrastructure matures, larger and larger transactions will migrate on-chain. The question is not whether machine-native payments can scale, but how quickly the trust layer can catch up to the settlement layer.
This article is a summary of the key findings. The full report delves deeper into the data, including protocol architecture analysis, insights from Coinbase and Tempo interviews, transaction economics modeling, and the regulatory landscape.
Author: Ben Harvey; Translated by Shenchao TechFlow