Source: a16z

Translation: Felix, PANews

The era of Agentic Commerce has arrived.

OpenAI's AI Agent Commerce Protocol (ACP) and Google's Universal Commerce Protocol (UCP) promise to implement checkout functionality in ChatGPT and Gemini. Soon, hundreds of millions of consumers around the world will find better products, merchants will see improved conversion rates, and platforms will take a 5-10% cut.

However, the checkout feature in ChatGPT is just an incremental improvement. It will not reshape society like the Internet did in the early 21st century, but open-agent commerce can do that.

We must go back to the 1990s to understand why.

At that time, there were two competing versions of the "Internet."

• AOL version: unified pricing, email, weather forecasts, and eventually the entire Time Warner content library.

• Open protocols: HTTP, DNS, HTML, and a browser named Mosaic.

Compared to AOL's version, Mosaic seemed laughably absurd. There were so few websites on Mosaic that a search function was unnecessary; an alphabetical index was sufficient. Eight years later, AOL merged with Time Warner in an equal deal for $350 billion. The market had made its choice: curated content appeared to be the future direction.

But soon, Mosaic and the open protocols won, and human civilization officially entered the digital age. Why? Imagine if a closed ecosystem had ultimately prevailed.

In 2004, Zuckerberg wanted to create Facebook, and he needed to strike a distribution deal with AOL. Two students at Stanford wanted to build a web index, and they needed permission from CompuServe. A person wanted to sell books online from their garage, and they needed to pitch proposals to MSN's content team.

They would have said, “Go back to school, kids.” None of this would have happened. The entire digital economy that we take for granted would not exist at all.

Open protocols mean there are no "gatekeepers." Anyone with a server and domain name can access the entire Internet. Innovations emerge from the edges while the centers lag behind, ultimately resulting in one of the largest wealth creation events in human history. This is a fundamental principle of capitalism: innovation comes from the margins.

Back in 1997: Tim Berners-Lee, Marc Andreessen, and others were researching protocols and browsers. At that time, running a server cost hundreds of thousands of dollars. It was unclear why content servers should respond to requests from unfamiliar users; this was expensive to do and there was no known economic incentive.

They created a message status code called "402" that servers could send to users: “Pay a small fee to access this content.” But no reasonable means existed for digital payments at that time. PayPal had not yet emerged, and the fixed fees for credit cards could be tens of cents—too high for a one-cent microtransaction.

Nonetheless, the Internet took off.

Google found a unique business model for the Internet: advertising. In traditional media, the primary economic relationship existed between content producers and consumers. Google borrowed the economic model from broadcasting, introducing a third party: advertisers. They would pay to bridge the relationship between content producers and consumers.

This was quite clever. Now, producers can monetize the attention of viewers. They do not need any prior relationship with consumers. Google sits comfortably in the flow of funds, positioned between advertisers and content producers, taking any proportion it wants as a commission.

Thus, the demand for micro-payments was bypassed. Open-source software began to emerge, and the cloud computing revolution exploded, causing hosting costs to drop by an order of magnitude. Google became the biggest advocate for a free and open Internet. The more consumers searched, the more money Google made. Thus, they invested hundreds of billions of dollars to make the Internet fast, affordable, and ubiquitous.

Then came the 2010s, where everything seemed unchanged.

Interest rates were low, technological advancement slowed, and closed ecosystems continued to grow.

In 2022, ChatGPT launched, and the world was about to undergo change again. Large language models (LLMs) can do more than just provide results. They can also generate and compile many such results into a clear summary, often without directly touching the content itself.

By the time GPT-4 emerged, the direction was clear: Agents are the next step in development; LLMs excel at using computers like humans, but at lower costs and higher efficiency.

Thus, the economic landscape of the Internet changed.

From 1997 to 2024, the core of the business model is “distraction marketing.” Humans are easily distracted by advertisements while reading web pages, and advertisers profit from their limited attention. Meanwhile, low-grade agents / agents will not be distracted.

There is a wonderful irony here: advertising created a free and open Internet, and the Internet spawned massive datasets containing trillions of tokens, which led to LLMs and ultimately the decline of advertising.

Since the launch of GPT-4, traffic to Stack Overflow has dropped by 75%, and traffic to tech news has fallen by 60%. Tech users are early adopters, but this trend will eventually sweep through all information online.

The checkout feature in ChatGPT is not important. The Internet has become a civic square, and traditional economic contracts are long outdated.

There are small areas on the Internet that have successfully fended off Google's "erosion," resembling "walled gardens" with genuinely unique content: Facebook, TikTok, and LinkedIn. These areas have been preserved thanks to thousands of high-paid engineers working day and night, keeping automated bots at bay.

However, the defenses of the walled gardens have also been breached. Agents with computer usage capabilities can perfectly simulate the traffic of real human users. Scammers will peddle various "magic pills" over the next decade, and capital from Sand Hill Road (Note: Silicon Valley venture capital center) will follow suit. But in reality, there is no magic pill. The fortress walls have been destroyed by warplanes.

What’s next?

Open-agent commerce.

The checkout feature in ChatGPT is like AOL in the era of agent commerce. It is a curated directory, a walled garden with a better user experience. To sell through it, merchants need months of business development, strict legal documents, concrete five-year plans, substantial revenue, a strong user base, and a story that makes shareholders happy when featured on the front page of The New York Times.

Open-agent commerce is like today’s HTTP. It is a simple protocol that allows agents to pay for anything they need. Data, cloud hosting, communication, and many things we have yet to imagine.

Coinbase's x402 and Tempo's MPP launched in collaboration with Stripe are two leading players. After 28 years since the invention of the "402" status code, we finally have a viable implementation. The cost of stablecoin transactions on modern blockchains is less than a cent, solving the fixed fee issue that doomed micro-payments in 1997.

An agent that can only buy from pre-approved merchants is like an employee with a company credit card that can only spend with three vendors. An agent with an open protocol is like an entrepreneur with a bank account.

There are no business development teams, no whitelists, only simple, permissionless standards.

These protocols focus on two things:

• Agent: “How do I make a payment?”

• Merchant: “How do I ensure the agent has paid?”

LLMs excel at using tools they have never seen before. Starting with Claude 4.5+ and Codex 5.2+, agents can discover APIs, read their patterns, and use them correctly without prior training.

Current discussions mainly focus on "skills." You can think of them as natural language programs that can be combined like building blocks. A non-technical founder can write a Slack message and have it execute like software:

• Order pizza from a nearby highly rated pizzeria, checking the delivery status every 10 minutes.

• Turn on the porch light when the delivery person is 5 minutes away.

• If delivered within 30 minutes, tip the delivery person $5.

No coding is necessary, nor is a computer science degree. An agent reads the intent, writes a computer-native program in real time, executes it, and then discards it. Programming no longer needs to be a discipline; mastering the human native language is enough.

Skills are indeed effective. However, they are merely a transitional product, the first way we thought to build once we discovered that agents can call upon unfamiliar tools. They need someone to write, publish, perform security checks, and update them. Moreover, agents need to pre-load them. This is cumbersome.

The discussion around skills masks a deeper breakthrough: agents can combine various capabilities in unprecedented ways.

Ordering pizza is a simple example. Here’s a more realistic case: an agent managing a small business supply chain notices that due to tariffs, the prices from a packaging supplier have increased by 15%. It finds three local alternatives, requests samples from each supplier, negotiates bulk pricing, and ultimately completes the switch. All of this is done before the business owner finishes their morning exercise.

No API collaborations, procurement teams, or bidding processes are required. Just an agent with a balance in an account and using an open protocol.

Discovery

Agents can make payments and combine capabilities, but they still cannot find the content they need.

The remaining question is "discovery." For agents: “How can I find what I want to buy?” For merchants: “How can I introduce my services to agents?”

AgentCash emerges. It is a unified account that provides access to all APIs on the Internet. When agents are blocked, they can still access thousands of APIs and continue operating with minimal costs.

The key is that AgentCash integrates payment functionality with merchant discovery. Merchants can register their servers at x402scan.com or mppscan.com and immediately showcase their services to all over 2,000 AgentCash agents.

In 1997, there was no business model for the Internet, and no one knew why servers should communicate with strangers. Open protocols and a strategy called “advertising” cleverly addressed this issue, ushering civilization into the digital age. By 2026, this clever strategy is fading. Open protocols and a status code that is already 28 years old are set to take its place.

Welcome to the era of open-agent commerce.

Further reading: Understanding x402 and MPP: Two Paths for Agent Payments

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。