At the Stripe conference, I saw the future of the AI economy

Author: Gao fei

Compiled by: Jiahua, ChainCatcher
In 1987, economist Robert Solow famously said, "You can see the computer age everywhere but in the productivity statistics."

This statement troubled economists for nearly a decade. It wasn't until the mid-1990s that the contribution of computers to productivity finally began to appear in the data.

Today, in 2026, the same confusion is unfolding around AI. The talk of bubbles is rampant. The academic community is in constant debate. Businesses are hesitant. The signals conveyed by macroeconomic data remain unclear.

But in one place, the impact of AI on the economy is beyond dispute.

That place is Stripe.

In the past few days, I attended the Stripe Sessions conference in San Francisco. Stripe processes transactions equivalent to nearly 2% of the global GDP, with an annual payment volume of $1.9 trillion and over 5 million businesses on its platform.

In the Forbes AI 50 list, 86% of the companies use Stripe. If the AI economy is a newborn baby, Stripe is the ECG monitor in the delivery room. It records the baby's heartbeat earlier and more accurately than almost anyone else.

A study released by the St. Louis Federal Reserve at the beginning of 2026 shows that AI-related investments have contributed nearly 40% to the marginal GDP growth in the U.S., surpassing the peak contribution of the tech sector during the internet bubble. And when these investments convert into revenue, a large portion of the settlements occur on Stripe.

More importantly, Stripe is not just recording the heartbeat of the AI economy.

At this year's conference, it announced the promotion of a brand new economic form: agentic commerce, where agents become the main entities in transactions.

In a media group interview, Stripe co-founder and president John Collison stated that he expects agents acting as buyers in commercial transactions to become mainstream within 12 to 18 months.

In just two days, 288 products and features were released. Over 10,000 attendees. One defining term: agentic commerce. Here are my observations and thoughts from the 2026 Stripe conference.

How Fast is the Development of the AI Economy?

Before discussing agentic commerce, it is necessary to look at the overall outline of the AI economy. Solow said in 1987 that there was no trace of computers in the statistics. Nearly forty years later, AI is clearly visible in Stripe's data.

On the first morning of the conference, CEO Patrick Collison presented a set of data. Since the pandemic, the number of new businesses created monthly on Stripe has remained high, but the curve has been relatively flat. Starting from early 2026, this curve has almost shown a vertical upward trend.

The most direct reason is that AI programming tools have lowered the barriers to entrepreneurship. Many developers can now build a billable product in just a few days through vibe coding.

Patrick defined this as a grander concept: the entire economy is undergoing a platform reconstruction around AI.

Maia Josebachvili, Stripe's Chief Revenue Officer for AI business, added an external comparison: before 2024, the number of apps released in the iOS App Store was declining. However, after the emergence of AI programming tools, the app release volume surged by 24% month-over-month.

This change is not only quantitative but also qualitative. Stripe Atlas is one of the easiest ways for entrepreneurs to register a company in the U.S.

Last week, it celebrated the birth of its 100,000th company. At the conference, I heard some astonishing numbers: companies registered through Atlas in 2025 generated twice the revenue of those registered in 2024 at the same lifecycle stage. The companies registered in 2026 have only been established for a few months, yet their revenue is already five times that of the same period last year.

In the AI economy report on the first afternoon, Maia Josebachvili listed several companies driving the rise of the AI economy.

Lovable achieved $100 million in revenue within eight months, followed by $400 million in the next eight months. Cursor reached an annualized revenue of $1 billion in less than two years, and three months later doubled to $2 billion.

Leading AI-native companies on Stripe grew by 120% in 2025. By 2026, the current growth rate has reached 575%.

Consumer growth is equally rapid. The highest spending group spends $371 per month on AI products, which is more than the total monthly spending of the average American on the internet, streaming services, and mobile phone bills combined. I roughly calculated my monthly token expenses, and they have long exceeded my mobile phone bill.

Patrick also made a comparison: the growth rate of businesses on Stripe is 17 times that of the global economic growth rate.

On the second day, John Collison directly mentioned the Solow Paradox and used a historical analogy.

In 1882, Edison lit the first electric lights for customers in Manhattan. However, for thirty years after electrification, productivity hardly improved. The reason was not that electricity was useless, but that factories were designed around steam engines. It wasn't until entire factories were rebuilt that productivity improvements became apparent.

John's judgment is that AI is in a similar stage. The transformation has occurred, but the old models have not had time to absorb it. "However," he said, "I don't think AI needs thirty years."

Stripe's data seems to confirm his optimistic view. On its platform, the AI economy has already exploded. In almost all traditional businesses I encountered, top leadership is pushing for AI deployment with a strong sense of urgency.

Born Global

In addition to speed, another characteristic of these AI companies that impressed me is that they have been global from day one. Stripe has a saying: go global by default.

Since becoming an AI blogger, I have often had an experience. AI content creation has no time zone. AI news from across the Pacific is as important as local news.

The operation of AI products is similar. Large language models blur the interface language and interaction habits that traditional software relied on. A unified chat box allows global users to use products through natural language. In this sense, large language models have made a unified global software market possible for the first time.

Data from the conference confirmed this observation. In the early waves, the fastest-growing SaaS companies covered about 25 countries in the first year and reached 50 countries by the third year.

The development speed of AI companies is completely different: they reached 42 countries in the first year and expanded to 120 countries by the third year.

Maia noted that Kazakhstan has now appeared on the market lists of many AI companies. In the second day's "Index Economy" session, Stripe provided a median data point: the top 100 AI startups had already sold products to 55 countries in their first year.

One company provided a specific example. Emergent Labs was founded in the U.S. in 2024, but nearly 70% of its revenue already comes from overseas. At least 16 countries each contribute at least 1% of its revenue.

Among leading AI companies, 48% of revenue comes from outside their home markets. Three years ago, this figure was only 33%. Global revenue is no longer a supplement but a fundamental base.

Speed and globalization are two core characteristics of the AI economy, both of which have a direct connection to Stripe. AI companies need to quickly establish payment capabilities. They need to accept payments in 40 countries and regions within the first week of their establishment. This is precisely what Stripe has been doing since its inception.

Here, let's briefly review the founding background of Stripe.

Stripe's founders, Patrick Collison and his brother John Collison, are both Irish. They are multinational entrepreneurs themselves.

At the conference, I met an Irish colleague who told me that in the eyes of AI entrepreneurs in Ireland, these two brothers are heroes. After arriving in the U.S., the brothers found it ridiculously difficult to collect payments online. Integrating payment systems required signing contracts with banks, undergoing PCI compliance reviews, and integrating with multiple intermediaries. The entire process could take weeks or even months.

Thus, in 2010, two twenty-year-olds dropped out of school, moved to San Francisco, and wrote a solution that allowed developers to accept payments with just seven lines of code.

The birth of these seven lines of code coincided with the rise of mobile internet and SaaS. Shopify needed to help millions of merchants collect payments. Uber needed to enable frictionless payments for passengers. Salesforce needed to handle global subscription services.

They all chose Stripe. As Stripe grew alongside these global clients, it established localized capabilities in 46 countries, covering 195 markets and supporting 125 local payment methods.

For consumers, Stripe is not a company that stands in the spotlight.

It hides behind the checkout pages of Shopify, the subscription confirmation emails of OpenAI, and the fare notifications of Uber. But this invisibility has not prevented it from becoming the financial pipeline underlying the internet economy.

In the AI era, this global financial infrastructure has given Stripe a head start in serving AI companies going global.

At this year's conference, I also met Abhi Tiwari, Stripe's global product head.

He took on this role three months ago and moved to Singapore. Stripe has engineering centers in San Francisco, Dublin, and Singapore, and has established a Latin America office in São Paulo. Abhi told me that many AI companies approach Stripe with the same opening line: "We are born global by default. It doesn't matter where our users are."

The old model of developing products at headquarters and then pushing them globally is being replaced by local teams directly developing in target markets.

Reaching global users is one thing. Getting paid by them is another. The latter is quite complex because each market has its own currency and payment habits.

Here, Stripe primarily helps AI companies and many other clients in two ways: pricing in local currencies and accessing local payment methods.

The former allows Brazilian users to see product prices in Brazilian reais instead of dollars, increasing cross-border revenue by 18%. The latter allows Indian users to pay using UPI and Brazilian users to use Pix, resulting in over a 7% increase in conversion rates.

After the AI demonstration tool Gamma added UPI payment in India, its revenue in India surged by 22% that month. At the booth, I also saw the presence of the Chinese company MiniMax. From what I understand, many Chinese companies going global use Stripe's financial services through their overseas entities.

These AI-native companies also share a common characteristic: they have very few employees. Many are solo founders. One or two people plus a group of agents can support a global company with actual revenue.

In a speech on the second day, Emily provided a statistic: on Atlas, the density of solo founders is approaching 5,000 per million Americans, and an increasing number of them have annual incomes exceeding $100,000.

Emily used the term solopreneur: a one-person company. John explained this phenomenon using Ronald Coase's "theory of the firm." Firms exist because the cost of internal coordination is lower than the cost of market transactions.

But AI may disrupt this logic. When agents can help you discover services, integrate software, and handle payments, the cost of external coordination will drop sharply. You no longer need a room full of employees to accomplish what once required an entire department.

From Human Economy to Agent Economy

The AI economy described above, no matter how fast it develops or how globalized it is, still has humans as the transaction entities. It is humans purchasing AI products. It is humans using AI tools to start businesses.

But the strongest signal I felt at this year's Sessions conference is that Stripe's next major focus is another transformation: an economic form where agents become market participants. This is agentic commerce.

This transformation has quietly begun to appear in Stripe's own data.

Stripe's product and business president Will Gaybrick presented a set of data. For years, Stripe's command line interface (CLI) has been used by a small group of technically skilled users, with usage remaining almost unchanged.

However, entering 2026, usage suddenly surged. The reason is that agents do not need a polished graphical interface. A simple CLI is often more practical.

Maia's data shows that in 2025, the traffic of agents reading Stripe documentation increased by about tenfold.

If this trend continues, by the end of this year, the number of agents reading Stripe documentation will surpass that of humans. The API documentation that Stripe has refined over more than a decade is finding a group of its most loyal new readers.

If the idea of agents spending money still sounds a bit strange, consider two real scenarios that are currently happening.

The first is that the shopping interface may have already shifted to the model chat window. Consumers now typically use ChatGPT, Gemini, or Instagram to search for products. The distance between searching and transacting is being compressed into a single interface. There are also related examples in China, including the now-familiar story of buying milk tea in AI applications.

In a media group interview, John Collison used his experience of purchasing a travel power adapter to explain why this compression is hard to reverse.

If an agent completes the entire process from search to order, and the product arrives at his home days later, he will not go to another website to fill in personal information from scratch, even if the product on that site might be slightly better. Once the shopping agent completes the search process, the logical next step is checkout.

The second example is even more interesting: OpenClaw. Those who have followed the "claw" wave know that it is currently one of the hottest open-source autonomous agent frameworks.

Users issue commands to agents through messaging applications like Feishu, Telegram, and WhatsApp, and the agents autonomously execute tasks.

The key point is that OpenClaw can consume hundreds or even hundreds of dollars in token costs within a day. It manages its token consumption and usage autonomously. Although human authorization is still required in many cases, ultimately, it is the agent that consumes the tokens. And tokens can be directly converted into money.

There is only one step between agents managing token consumption and agents spending money directly. At this year's conference, Stripe's demonstration crossed that line.

Agent Buying and Selling

On the main stage on the second day, a demonstration received rounds of applause.

John Collison gave an agent a simple instruction: research how AI demand is affecting the energy market. The agent began searching and found that Alpha Vantage had a dataset it needed on the energy market, priced at 4 cents.

The agent determined that the price was within budget, then autonomously purchased and downloaded the data using a stablecoin wallet in the Tempo CLI, as it was clearly unreasonable to use a credit card for a 4-cent transaction.

Next, it generated a complete analysis report. Just this step was already astonishing. But John then told the agent, "Publish and sell this report. Set a price you think is reasonable and make it available for other agents to find and purchase."

The agent checked the licensing terms of the Alpha Vantage dataset, confirmed that commercialization was allowed, built a website, published the report, and generated an instruction file allowing other agents to purchase the data through a single request.

Within minutes, an agent completed the entire chain of operations: research, procurement, production, compliance review, publication, pricing, and sales. It was both the buyer and the seller. After the demonstration, John said, "Agent commerce has arrived."

The other two demonstrations on the first day were equally impressive. Will Gaybrick built an API code review application that allowed agents to obtain review services on behalf of users. Throughout the process, he did not inform the agent of any payment information.

While executing the task, the agent automatically discovered that the application used a machine payment protocol (MPP) and independently completed a $2 payment. The only thing humans did was press a fingerprint for authorization. This zero-configuration payment discovery capability is at the core design of MPP. Developers do not need to write payment logic specifically for agents. The agents can find it themselves.

Next, Gaybrick combined a real-time billing engine Metronome, a blockchain designed for payments Tempo, and stablecoins to demonstrate streaming payments (breaking down funds into countless tiny amounts that sync in real-time and continuously as services like AI computing power are consumed).

An application billed in real-time based on AI's token consumption, priced at $3 per million tokens. Multiple agents were running simultaneously. The dashboard on the left showed the rising token consumption, while the right side showed the micro-payments of stablecoins flowing in synchronously.

Opening the Tempo blockchain explorer, one could see a total payment of $3.30 composed of thousands of micro-payments, each equivalent to one-thirtieth of a cent.

Credit cards cannot do this. ACH clearing cannot do this. UPI and Pix also cannot do this. Gaybrick announced on stage that this is the world's first streaming payment business.

The Return of Micropayments and New Consumption Logic

Shopping in chat windows and OpenClaw are examples of agents representing human consumption. But in the group interview, Collison made a more grand judgment: agents may also create entirely new demands.

He believes that agents may revive a business model that has been discussed for years but has never truly succeeded: micropayments. Humans are not good at making extremely fine consumption decisions. The reason Spotify replaced single song payments with a monthly subscription fee of $9.99 is that no one wants to weigh whether a song is worth 15 cents every time they hit play.

Agents do not have this cognitive burden. If this judgment is correct, an entire class of business models that failed due to human cognitive resistance may suddenly become feasible in front of agents.

Maia expressed a similar view in my one-on-one conversation with her. She said she had just spoken with dozens of AI founders, and when they discussed agentic commerce, pricing was the most frequently mentioned topic.

Every transaction has two parties: the buyer and the seller. If the buyer becomes an agent, what should merchants do?

In an interview, I asked Stripe's product head Jeff Weinstein a question: Humans often say "the customer is king." Merchants need to please consumers. So how should they please agents?

Jeff's answer was to imagine agents as the best programmers you know. They want perfect information, structured formats, quick readability, and all the background information needed for decision-making.

Human consumers like beautiful images and smooth animations. Agents want raw structured data, precise logistics information, and to complete transactions with as few steps as possible.

In another conversation, Meta's VP of product Ginger Baker summarized this transformation more radically: payments will shift from an "instant" to a "strategic" approach.

Human consumer purchases are discrete.

You walk to the checkout, take out your wallet, swipe your card, and the transaction is complete.

Agent consumption is continuous.

You set a group of rules, such as "spending no more than $50 on daily necessities this week," "always prioritize using this card," or "never automatically authorize any transaction over $500." Then, the agent will continuously make autonomous purchases within the authorization framework you set.

Computing Power is the New Cash

If agents truly become a new type of consumer, they will also bring new risks. These risks are fundamentally different from traditional SaaS transaction risks and are entirely different from the risks faced by human consumers.

During the Sessions conference, I paid special attention to this topic and discussed related issues with several Stripe executives.

Stripe's data and AI director Emily Glassberg Sands described three rapidly growing fraud patterns. The first is multi-account abuse. The same person repeatedly registers different accounts, each of which can receive free credits.

According to Stripe's network data, one in six AI company registrations involves such abuse. The second is malicious consumption during free trials. This is particularly deadly for AI companies because every trial burns real inference costs.

She gave an example: for a partner company, the token cost to acquire each paying customer exceeds $500 because converting one customer requires 25 free trials, of which 19 are fraudulent.

The third pattern she referred to as "dining and dashing." Customers consume a large number of tokens and then refuse to pay at the end of the month. Emily also quoted a saying: "Computing power is the new cash." When traditional SaaS is abused, the marginal cost is almost zero. But every inference call for AI companies incurs real costs. Stealing tokens is stealing money.

However, there is a particularly tricky dilemma here. Many AI founders respond to abuse by simply shutting down free trials.

Emily said she had asked everyone who claimed to have "solved" this problem how they did it, only to find that their solution was merely to close the free tier. But Jeff believes this will raise another problem.

Agents are increasingly becoming the primary means of discovering new services. If agents cannot trial a service on their own, they will simply jump to another link.

Emily added that if the options presented to agents are "join the waiting list" or "contact sales," the agents will immediately leave. Closing self-service registration to prevent fraud may mean handing over the most important growth channel to competitors.

Stripe's solution to this dilemma is its fraud prevention system, Radar. The logic of Radar is easy to describe: every time a transaction is completed on Stripe, Radar learns.

Transaction data from 5 million businesses feeds into a shared risk identification network. If one company encounters a certain fraud pattern, all companies can benefit from it. Last month, Radar intercepted over 3.3 million high-risk free trial registrations among eight high-growth AI companies.

Jeff also proposed a counterintuitive viewpoint: agent shopping may ultimately be safer than human shopping on web pages. Human web shopping trust verification relies on inference: how long users stay on the site, whether the click path looks normal, etc.

However, agent transactions can be verified programmatically. Stripe's shared payment tokens tokenize payment credentials, so agents never touch the original credit card numbers. Users authorize through biometrics and can set transaction limits, time windows, and merchant whitelists.

When the trust mechanism shifts from inference to confirmation, the security baseline may actually improve.

Ecosystem, Protocols, and a Piece of History

At this point, you should be very clear that the realization of agent commerce relies on a well-functioning ecosystem. At the 2026 Stripe Sessions, I met someone from the food industry. He said he attended the conference to understand whether agent commerce could become a new opportunity for his company. This is the seller's perspective.

So this cannot be accomplished solely by Stripe. It requires an ecosystem.

After wandering around the Sessions exhibition hall for two days, I saw numerous booths from companies spanning various segments of the financial industry chain.

Stripe has also launched or joined a series of protocols with upstream and downstream partners to connect various parts of the ecosystem: buyers and sellers, humans and machines, and machines and machines. The machine payment protocol (MPP) allows agents to discover and complete payments via HTTP.

The agent commerce suite allows consumers to complete purchases directly within the AI applications of Google, Meta, OpenAI, and Microsoft. The Universal Commercial Protocol (UCP) is a cross-platform commercial protocol initiated by Shopify and joined by Meta, Amazon, Salesforce, and Microsoft. Stripe joined the UCP's general council.

A group of companies that are both partners and competitors agreed to collaborate on a shared protocol because fragmentation makes it difficult for agents to consume smoothly across platforms. This benefits no one.

Speaking of protocols, I saw a special Stripe partner at the exhibition: Visa. In my view, Visa is essentially a protocol platform.

Seeing Visa immediately reminded me of a book I have long loved: "One from Many: VISA and the Rise of the Chaordic Organization" by Visa founder Dee Hock.

One theme of this book is how banks, money, and credit cards can be redefined in the electronic age. Money no longer has to be coins and paper bills. It can also be data that is institutionally guaranteed, recorded on networks, and flows globally.

In the late 1960s, the American Bank issued the U.S. Bank Card, expanding nationwide. A large number of interstate consumers flooded in, causing the old system to collapse. Hock realized that the problem lay in the organizational structure. Dozens of competing banks needed to share infrastructure, but no existing organizational form allowed them to cooperate while competing.

He utilized decentralized design principles to make all banks equal members of the new organization, and the American Bank relinquished its exclusive control over the system. That organization was later renamed Visa.

So two different companies from different eras are doing similar things. Is there some kind of inheritance between them?

With the help of any agent, the answer is easily found. Patrick Collison has publicly paid tribute to Hock. After Hock passed away in 2022, Patrick referred to him as "a severely underrated innovator" and stated that Hock inspired him and his brother.

A more obvious sign of his influence is a hiring decision: David Stearns, the author of the authoritative academic history of Visa, later joined Stripe.

Another detail that would make anyone familiar with payment history smile is that on stage, Georgios Konstantopoulos, the CTO of Tempo blockchain, showcased the validator lineup. One of the names was Visa.

The institution founded by Hock, Visa, has now become a participating node in a blockchain network incubated by Stripe. The students have built a new network, while the teacher has become one of the nodes.

When Patrick traced the intellectual origins of Stripe at the opening ceremony of the conference, he said he was initially a programmer writing code in Lisp. One of the core ideas of Lisp is "code is data."

He translated this idea into Stripe's own terminology: "The fundamental idea of Stripe is that money is data. When we launched Stripe in 2011, this was not the mainstream view in the industry."

Hock explored the essence of money from the perspective of organizational theory and concluded that money is merely "a guarantee of value exchange." The medium that carries it can be anything. Collison, on the other hand, approached it from the perspective of programming languages, directly equating money with data: data that can be programmed, called via APIs, and operated by agents.

The two expressed the same idea in different languages. On the same stage, Ginger Baker put it more bluntly: "Isn't money just another form of digital content?"

If money is data, then consumers of data will naturally become consumers of money.

Stripe's Content Gene

At this point, the story of the AI economy is nearing its end. But let's take a slight detour. Stripe can almost be considered a peer of content workers.

This company is not only good at financial services. It is also skilled at creating content products. Its publishing department, Stripe Press, has excellent taste. Many people know it thanks to its publication of "Poor Charlie's Almanack."

Its podcast "A Cheeky Pint" is also unique and has a wide audience. Google CEO Sundar Pichai, Anthropic CEO Dario Amodei, and a16z co-founder Marc Andreessen have all been guests.

During Sessions, I met Tammy Winter, a senior editor at Stripe Press, and designer Pablo Delcan. Tammy joked, "Stripe is a publisher with a billion-dollar company attached."

Pablo Delcan talked about his understanding of taste. He said taste is something that develops over time and needs to be refined. In terms of design trends, he believes that without sacrificing simple concepts and clear expression, the new challenge is how to add a certain complexity through details and precision.

When the topic turned to books, Tammy told me that within Stripe Press, the series of books published for founders and builders is called the "Turpentine" series.

These books focus on operational knowledge, tools, techniques, maintenance, and practical content that keeps work running smoothly. They are not abstract theories. They aim to help readers solve specific operational problems.

The name comes from a story about Picasso: when art critics gather, they discuss form, structure, and meaning; when artists gather, they talk about where to buy cheap turpentine.

What this series aims to do is become the cheap turpentine in the hands of founders. Think about it: for AI companies going global, Stripe's financial services are another form of turpentine. You don't have to worry about payments, compliance, or foreign exchange. You can focus on building products.

This interlude may seem unrelated to the main story, but it actually has a potential connection.

Stripe also has a magazine called "Works in Progress," which focuses on how the economy grows. Its podcast interviews leaders in the AI economy. Sessions itself, to some extent, resembles an economics lecture.

On the second morning, John Collison spent an entire session discussing economic data, Coase's theory of the firm, and the Solow Paradox. I suspect that a financial services company cares so much about economics because understanding the changes in economic structures is the way it discovers the next product opportunity.

As a podcast enthusiast, when I met John Collison on the first day of the conference, the question I most wanted to ask was not about finance. It was about podcasts. I asked him if, after interviewing so many different people, there was a fundamental question that ran through all the conversations.

He thought for a moment and said that he was genuinely interested in how these people's companies actually operate, what kind of competitive equilibrium they are in, and how they understand their businesses.

Coincidentally, a small incident occurred at the end of the first day. The last fireside chat was originally scheduled for Patrick to interview OpenAI co-founder Greg Brockman. But shortly before taking the stage, the guest was changed to Sam Altman. Patrick explained that after all, "AI is a rapidly evolving field."

So the surprise turned into a celebration. The audience erupted in cheers.

The two have known each other for nearly 19 years. Altman was one of Stripe's earliest angel investors; when he invested, the Collison brothers were not yet 20 years old. Because of this, Altman appeared very relaxed throughout the conversation.

Towards the end, Patrick asked a personal question: why did Altman invest in two teenagers back then? Altman said he remembered that the product they wanted to create addressed problems they personally encountered, and he also saw that this opportunity could scale because many others needed the same thing.

I believe his answer about podcasts and his answer about investment point to the same thing: finding real demand and solving real problems.

In the conversation, Altman divided OpenAI's transformation into three stages: from a research lab to a product company, and then to a "token factory" providing intelligence to the world. Each stage corresponds to a different mission.

Stripe is similar.

In 2010, the problem that two Irish teenagers solved was that online payments were too difficult. Along the way, they solved the same problem for 5 million users. And in 2026, they discovered a new problem: the customers of these businesses may soon no longer be human.

With one hand doing podcasts and the other doing publishing, discussing Coase's theory and the Solow Paradox on stage, while the exhibition hall is filled with protocols and APIs, Stripe is not only creating the AI economy. It is also recording the AI economy.

At the conference, a thought flashed through my mind that might sound a bit crazy: Stripe holds transaction data equivalent to nearly 2% of the global GDP. It can see where every dollar of AI revenue comes from, where it goes, and how fast it grows.

If Solow had such an ECG monitor back then, perhaps he wouldn't have had to wait a decade to find the trace of computers in the statistics.

Maybe one day, Stripe can provide a model for the AI economy. Not a large language model, but a Nobel Prize-level economic model. Who says that is impossible? Just a few years before DeepMind founder Demis Hassabis won the Nobel Prize, who could have predicted that scene?