Author: jessy

Compiled by: Jiahua, ChainCatcher

For the past year, I have been dedicated to building infrastructure for the Agent economy, engaging in discussions with teams from Stripe, Visa, Coinbase, Google, and dozens of startups driving Agent commerce. I have sifted through the entire industry, released products, and tried to find market fit.

There is currently no real demand, and startups face numerous structural issues as they engage in this field.

Last month, Stripe launched 288 new products at the Sessions conference, with nearly 40% of all document views being for their Agent documentation. Their Agent commerce marketplace has over 1000 enabled merchants. However, the number of registered Agents conducting transactions at the Sessions conference was in the single digits.

Visa mentioned that their Agent payment tokens (linked to Agents and used for user payments as tokenized payment credentials) currently require 3 to 9 months of KYC approval and require a minimum revenue threshold of $250 million to qualify. Nowadays, only companies the size of Amazon and Walmart can complete this identity verification loop.

Coinbase reported that as of April, there were 69,000 active Agents on the x402 protocol and 165 million transactions. However, independent on-chain analysis indicates that the actual daily transaction volume is approximately $17,000, of which about half are test transactions (according to CoinDesk in March 2026).

Agents for Merchants

We built shop.fast.xyz to directly validate the real application of purchasing agent-based commerce. It includes real products, merchants, and transactions.

For most product categories, the current user experience of AI shopping is far inferior to traditional e-commerce. When you buy clothes, electronics, or furniture, you want to see pictures, browse various options, and make side-by-side comparisons.

The chatbot conversation format is actually a regression. You are essentially replacing a rich visual interface with plain text conversations, while humans are inherently visual shoppers.

Agents perform exceptionally well in areas we originally thought would be challenging. They can understand user needs and handle instructions like "something similar but cheaper" appropriately. The model layer plays a role.

However, they cannot replace the experience of viewing ten products side-by-side and selecting one. The chat interface can be enhanced with carousels and interactive displays, but at that level, you are effectively just recreating an e-commerce front-end in the chat window. For visually driven price comparison shopping, we have yet to find a compelling reason to prove that chat interfaces are better than native e-commerce interfaces.

We see real demand from merchants, but it is a defensive demand.

Merchants want their stores to be queryable by Agents. This is not because current customers are purchasing through Agents, but because they fear being left behind if this becomes the mainstream channel.

This is a kind of "Agent Engine Optimization (AEO)" strategy, but currently, it is just icing on the cake rather than essential. Merchants are preparing for a wave that has yet to arrive.

Conversational commerce can indeed enhance the experience in certain scenarios: high-frequency, low-decision-cost purchases where users already know what they want. Ordering takeout is the most obvious example. The market is huge, the frequency is extremely high, and the decision is quick ("Order Pad Thai from that place I used last time"). Conversational Agents have a shot here.

However, large takeout platforms do not have open APIs. The only way is "computer use": having AI operate applications through visual navigation like humans do. This method is slow, fragile, and the reasoning cost is simply unsustainable for a $15 lunch order.

Another breakthrough lies in the fact that the UI navigation of certain stores is extremely complex, making it very painful. Layered discounts, promotional codes, loyalty programs, and confusing checkout processes.

An Agent that understands "use my coupon, deduct my reward points, find the cheapest shipping, operate in my native language" can simplify these currently terrible experiences. This is especially important for elderly users, non-native speakers shopping at foreign online stores, or in very niche scenarios with specific demands.

Both breakthroughs require a large consumer-facing (B2C) distribution channel. You are competing with DoorDash (the largest delivery platform in the United States, with a 56% market share) and Amazon for user entry points.

Consumer-scale distribution is an advantage of giants. The supply side of purchasing agent-based commerce is already ready, while the demand side is limited by user experience and distribution channels; building more infrastructure does not solve these two problems.

Agents for APIs

We have discussed actual payment needs with dozens of developers. The situation is almost remarkably consistent: the current use of Agents for APIs is frequent, including computation, inference, and data sources. Developers already have subscription services, archived API keys, and billing relationships with core vendors.

A typical argument for stablecoins is that on Stripe, the minimum effective cost for credit card processing is about 2.9% plus 30 cents, making API calls under one dollar not cost-effective. However, for today's low-frequency transaction volumes, prepaid amounts can solve this issue. Developers can preload their accounts, and the problem is resolved.

The deeper issue lies in the vendor market. Most mainstream SaaS companies do not want to offer on-demand API access that costs only a fraction of a cent. Their business model is based on multi-year corporate contracts. Companies that rely on large commitment contracts will resist bypassing their existing pricing mechanisms.

Machine commerce is structurally a long-tail market, including smaller services, niche data sources, individual developers, and MCP servers. Protocols like MPP and x402 are very suitable for this segment.

By definition, this is a market serving advanced users with special needs, and historically, developers have often been one of the groups with the lowest willingness to pay.

When Stripe Projects launched, it partnered with 32 vendor partners such as Vercel, Supabase, Cloudflare, Twilio, etc., covering most tools that developers use to build and deploy software, all of which can be accessed through existing billing systems. The top requirements of the developer tech stack have already been met.

New payment channel opportunities exist in all areas outside of these top 30 services: opportunities do exist, but their scale is inherently much smaller than those flashy numbers imply.

The same pattern applies to content acquisition. Agents have continuously crawled and summarized articles, while publishers are fighting back.

However, when content monetization arrives at scale, it will be realized through those CDN vendors already positioned between publishers and the internet (Cloudflare has already launched AI auditing tools for this purpose), or through large-scale licensing agreements between publishers and AI labs.

The opportunity for this infrastructure will ultimately flow to those giants that already have distribution channels.

Agents for Agents

The business model of Agent to Agent is a long-term vision that is currently almost entirely theoretical, without anyone realizing meaningful transaction volumes. Startups are tackling the core challenges: Agent discovery, trust establishment, terms negotiation, and dispute resolution.

When this transactional structure is actually implemented, it will be radically different from existing payment tracks. Neither party in the transaction will include human identity. Delays will be at the sub-second level. Funds ranging from a fraction of a cent to millions of dollars will flow in the same process.

Additionally, there will be a multi-party settlement mechanism, which completely does not align with the bilateral buy-sell model pre-set in existing payment tracks. Once this situation occurs, we believe it will happen quickly and on a large scale.

This is a long-term bet on dedicated settlement infrastructure, and it does exist in reality. However, "real long-term bets" and "current markets" are two completely different matters.

For months we have been among those advocating for this market and have built complete infrastructure around it in the past few years. With our distributed network, it can theoretically scale to over 1 billion TPS with delays of less than 50 milliseconds and average consistency of 10 milliseconds. But we must align with the current real position of the market.

Agents for Finance

This can be said to be the only category with existing demand. The customer base already exists and has a willingness to pay. Nowadays, fund managers, finance teams, and DeFi users are all paying for financial tools. Integrating AI into existing workflows is a natural product evolution.

Agent finance also creates entirely new behavioral patterns. An Agent that can autonomously monitor and rebalance hundreds of positions in real-time operates in ways that humans cannot replicate manually. This is not just automation; it is a significant enhancement of capability.

The challenge lies in the competitive landscape. The finance industry is highly regulated and heavily relies on existing business relationships. Established institutions have licenses, compliance infrastructure, and customer relationships. Startups can seek a place in areas with lighter regulation (like DeFi), where giants are slow to act or where AI can create capabilities that giants do not possess.

But compared to the other three categories, the competitive dynamics here are more favorable for mature companies, as layering AI on top of existing products and customer bases is much easier than the reverse.

The Real Competitive Point

So why are people still building these things? There are two reasons.

The first is motivation. Industry giants have ample cash flow to bet on a future that may take years to materialize. For them, the cost of entering five years early is merely a rounding error, while the cost of entering one year late could be devastating. So they must build.

The second is a cognitive blind spot. When your primary business is payments, every problem appears to be a payment problem. The Agent economy requires a payment layer, so let’s build that payment layer.

But payments are just one part of a much grander issue. The real challenge is not how to transfer funds between Agents, but how to coordinate work between Agents and humans, validate work outcomes, and settle results. Payments are just one part of settlement. Settlement is just one part of coordination. And coordination is the real big piece of cake.

Large-scale coordination will naturally give rise to a settlement mechanism as a necessity. Payments are merely an instrument in this symphony, not the entire piece. Companies that solve coordination problems will swallow payment businesses, not the other way around.

Most established companies are engaging in defensive construction to prepare for future scenarios of large-scale machine transactions. Because their funding runway is infinite, the timeline does not matter to them.

But startups do not have this luxury. We must seek out the true state of the market; we cannot simply wait for the wave to break on the shore.

A year of construction experience has led us in an unexpected direction. The market activities there are real, growing rapidly, and have yet to be adequately served. It exists outside the four categories we described.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。