Wintermute argues that the core infrastructure debate in crypto is largely over and that the next major build cycle sits outside the sector’s old boundaries. In its latest industry thesis, the trading firm says the more important question is no longer what crypto can still build for itself, but what a world of autonomous systems will require from crypto rails. Its answer is the machine economy: a setting where AI agents, warehouse robots and automated research systems stop acting as software tools and begin operating as economic actors that make decisions, move money and carry out tasks on their own. The report points to three shifts that make that thesis more plausible now than a few years ago: models can take actions rather than simply answer prompts, open standards for payments and machine interaction are taking shape, and agents can run over long periods with persistent context. Wintermute says those changes expose the limits of financial and trust systems designed around humans and firms as accountable counterparties. It highlights three areas it believes are worth watching most closely: an agent economy layer, physical AI, and machine-driven discovery. The firm also notes that Coinbase, Robinhood and Binance have each rolled out agent trading infrastructure in recent months, while Robinhood has even launched a new chain for that purpose. At the same time, Wintermute flags unresolved security and liability risks, citing incidents involving Grok and Moonwell as signs that the stack is still early.

Visa and blockchain analytics firm Artemis have released a joint report arguing that AI agent payments are no longer confined to proof-of-concept experiments. Using on-chain data from the x402 and MPP protocols, the report says agentic payment systems have already processed more than 100 million transactions this year, showing that machine-driven commerce is starting to operate at real scale. The report separates AI commerce into two layers. In larger, human-facing purchases such as travel bookings or subscription management, existing card rails still fit the job. In machine-to-machine activity, though, the payment profile changes sharply: transactions are frequent, often worth less than $1, and in many cases only a few cents. Visa argues that this is where stablecoins and lower-cost blockchain settlement become more practical than traditional card networks. The report also points to growing industry coordination around standards, citing 40 members in the x402 Foundation, including AWS, Google, Visa, Mastercard, Stripe, Coinbase and Circle. At the same time, it identifies liability as a central unresolved issue. If an AI agent pays the wrong party or is manipulated into moving funds, the current refund and dispute systems may not be able to keep up with the speed and complexity of agent-driven transactions.

Galaxy Digital has laid out a broad framework for what it calls the “inference capital markets,” arguing that AI inference is moving from a purely technical service into an asset class that can be priced, hedged, financed and traded. In a research report written by Galaxy Digital Vice President of Research Lucas Tcheyan and circulated in Chinese by TechFlow, the firm links several parallel developments into one market structure: the rise of GPU price indexes, planned GPU futures from Intercontinental Exchange and CME Group, tokenized claims on future AI inference, useful proof-of-work networks that subsidize inference production, and stablecoin-funded lending against GPU hardware. The report’s central claim is that inference has now overtaken training as the main driver of global GPU demand, while autonomous agents are emerging as a new class of machine-native buyers that can pay for model output programmatically. Galaxy argues that the market is still early and fragmented. It sees progress on the off-chain side, where Ornn, Silicon Data and Compute Desk are building reference pricing for compute, and where Kalshi, ICE and CME are already moving toward tradable GPU-linked products. On-chain, the report highlights Venice’s VVV and DIEM system for tokenized inference access, Pearl and Ambient’s different attempts to turn inference production into useful proof-of-work, and USD.AI’s stablecoin-based credit model for financing AI hardware. Even so, the report says the sector has not yet solved its hardest questions: whether real demand for verifiable, censorship-resistant inference will grow beyond a niche, how token value can be tied to actual product usage instead of emissions and speculation, and whether legal enforcement and collateral recovery in GPU-backed lending can hold up in a true stress cycle.

Visa and Artemis have published a joint research report, "Agentic Payments from the Ground Up," outlining how AI agent payments are developing and where blockchain-based settlement is starting to show traction. The report splits the market into two broad categories. One covers larger commercial payments made by agents on behalf of users, such as airline bookings and subscription management. The other focuses on machine-to-machine micropayments, including API calls and purchases of computing resources, where individual transactions are often below $1. Among the protocols highlighted, x402, a machine payments protocol incubated by Coinbase and Cloudflare and later managed by the Linux Foundation, has processed about $15 million in adjusted transaction volume since going live in May 2025. It has recorded roughly 109.6 million transactions, with activity centered on Base, Solana, and Polygon. The report also points to Machine Payments Protocol, or MPP, built by Stripe and Tempo with support from Visa. Since launching in March 2026, MPP has settled about $25,000 across around 115,000 transactions. Visa said the rise of AI agent payments is increasing demand for low-cost, high-frequency, machine-native payment infrastructure. The report adds that stablecoins and blockchain networks could become an important part of micropayment use cases, while the broader payments system is more likely to evolve through a mix of cards and stablecoins rather than one replacing the other.

Binance Research said in a July 8 report that stablecoins are no longer expanding mainly as liquidity instruments for crypto trading. The report argues that the sector is shifting toward a broader role as digital financial infrastructure, spanning value storage, yield generation, payments, settlement and cross-border treasury management. The report points to several data sets inside the Binance ecosystem and across public blockchains. On Binance, 30% of users with at least $10 in assets now keep more than half of their portfolios in stablecoins, up from 4% in 2020. It also highlights persistent fiat-to-stablecoin premiums across multiple economies, stronger savings demand in emerging markets, and growing use of tokenized Treasury products and structured on-chain yield products. Binance Research also links stablecoin growth to merchant payments, non-dollar stablecoins, weekend settlement activity, AI agent micropayments and on-chain foreign exchange. In its view, the category is becoming less tied to crypto bull and bear cycles and more connected to wider global financial demand. The report says regulation, reserve transparency and macro conditions will still shape how quickly this shift unfolds.

NEAR Intents said its payment solution has been integrated into the machine payments protocol, or MPP, used by Stripe and Tempo, making it one of the supported payment methods on the system. With the integration in place, agents can settle payments through MPP using any asset on any supported blockchain. Users are also able to pay with assets they already hold on any supported chain, while merchants can receive the specific asset they want on their preferred network. According to NEAR Intents, cross-chain swaps are handled in the background by its solution network through the 1Click Swap API. The update centers on payment routing and settlement across supported chains rather than requiring users and merchants to transact on the same network or in the same asset.

Tempo Chain, using Simplex BFT consensus, achieves sub-second finality. With Stripe, it introduces the Machine Payments Protocol (MPP), enabling autonomous AI agent payments at $0.001 per transaction, ushering in a new era of AI-driven payment infrastructure.

Coinbase and Stripe are competing to define AI-native payment rails through x402 and MPP, reviving the HTTP 402 concept with stablecoins, Layer 2 networks, and infrastructure built for machine-driven micropayments.