This article is authorized to be reprinted by Automatic Insight, author: Rhythm Editorial Department, copyright belongs to the original author.
If you haven't been following the payment industry in the past month, you may have missed some important news.
On September 29, Stripe and OpenAI jointly announced that ChatGPT users can shop directly in the chat window without needing to navigate to the merchant's website. The next day, Visa launched a stablecoin preloading pilot, allowing financial institutions to use USDC and EURC for cross-border settlements. A day later, Stripe made another move, releasing a platform called "Open Issuance," enabling any business to issue its own stablecoin.
On October 9, market news emerged that Mastercard and Coinbase are competing to acquire the stablecoin infrastructure company BVNK, with bids ranging from $1.5 billion to $2.5 billion. Just last December, the company's valuation was only $750 million.
This is just the tip of the iceberg. If we extend the timeline to the entire month of September, you'll find that Mastercard, Google, Visa, and Stripe all made significant moves in the fields of AI payments and stablecoins within a similar timeframe.
Let's first review the key events of this month in full.
In one month, there were nine major news stories, a density that is rare in the payment industry. More importantly, these news items are not isolated product releases; they resonate with each other and build upon one another.
When AI agents begin to replace humans in initiating payments, a truly tricky problem arises—who authorizes, who is responsible, and how can we prevent AI from completing a mistaken transaction due to hallucinations?
Traditional payment systems are built on a simple premise: humans will personally click the purchase button. But when this premise is broken, the entire authorization and accountability mechanism must be redesigned.
Stripe and OpenAI's answer is "Shared Payment Tokens," abbreviated as SPT. This is a new payment primitive that allows AI agents to initiate payments on behalf of users without accessing the user's real account or card information. Each SPT is limited to a specific merchant and total cart amount, granting AI sufficient payment authority while protecting user privacy and security.
ChatGPT's instant checkout feature is based on this technology, allowing users to purchase items directly from Etsy in the chat. Soon, this feature will expand to Shopify merchants, including brands like Glossier, Vuori, Spanx, and SKIMS.
Google chose a different path. It proposed the AP2 protocol, using three types of verifiable digital credentials: Intent Mandate, Cart Mandate, and Payment Mandate. The Intent Mandate defines the conditions under which the user authorizes the agent to make a purchase; the Cart Mandate is the user's encrypted signature authorizing a specific shopping cart; the Payment Mandate signals to the payment network and issuer that this is a transaction involving an AI agent.
This mechanism provides fine-grained control and traceable audit trails. Google emphasizes that AP2 is an open protocol, an extension of A2A and Model Context Protocol, and does not belong to any single company.
Mastercard's strategy is more pragmatic. "Agent Pay" does not emphasize technological innovation; its core value lies in compatibility. Mastercard is collaborating with multiple platforms, including Stripe, Google, and Ant International's Antom, to ensure its payment network can seamlessly integrate into the mainstream AI agent ecosystem.
The three protocols were launched almost simultaneously. They attempt to solve the same problem but take completely different paths. Stripe chooses to first occupy the scene and then promote standards; Google establishes standards first and then attracts applications; Mastercard seeks not to dominate but to ensure it is not absent.
History has repeatedly proven that whoever controls the standards controls the future. This battle over protocols is quietly determining the power landscape of the AI commercial era.
The trading volume of stablecoins has already surpassed the combined total of the two payment giants, Visa and Mastercard. This figure has alerted the entire industry; stablecoins are no longer experimental products in the crypto world but are becoming the underlying infrastructure of the global financial system. With the rise of AI agent payments, this trend is further amplified.
AI agents require a payment method that is available around the clock, instantaneously settled, low-cost, and programmable. Traditional bank wire transfers can take days, and cross-border payments often go through multiple intermediaries. Stablecoins almost naturally fit this demand, settling transactions in seconds at very low costs, and can be combined with smart contracts to execute complex payment logic.
Google's AP2 protocol has clearly identified stablecoins as the primary payment method. In their design, stablecoins serve as a common language between AI agents, possessing digital throughput while maintaining currency stability.
Traditional payment giants have chosen different response strategies.
Visa launched a stablecoin preloading pilot, allowing financial institutions to use USDC and EURC to top up Visa Direct accounts. In other words, stablecoins are no longer competitors outside the Visa system but are being absorbed into the network. Mark Nelsen, Visa's product head, stated in an interview with Reuters that it is extremely difficult to rebuild the underlying software of the global payment system, and integrating stablecoin technology into existing processes is a more realistic path.
Stripe's Open Issuance is more aggressive. This platform not only supports stablecoin payments but also allows any business to issue its own stablecoin. More importantly, businesses can share the profits generated from reserves.
In the past, issuers like Circle and Tether would invest the dollars deposited by users in low-risk assets like government bonds, keeping all the profits for themselves. Stripe has broken this pattern, allowing issuers to share profits with businesses.
Stripe President William Gaybrick believes that the gradual clarification of the regulatory framework has significantly lowered the barriers for businesses to enter the stablecoin space. He expects to see dozens or even hundreds of corporate stablecoins in the future. Open Issuance supports multiple chains, including Ethereum, Solana, and Stripe's self-developed Tempo blockchain.
The bidding war for BVNK reveals the true value of stablecoin infrastructure.
Founded in 2021, this company focuses on helping businesses achieve seamless conversion between stablecoins and fiat currencies, boasting extensive banking partnerships and multiple financial licenses, having processed over $20 billion in transactions.
Last December, BVNK's valuation was only $750 million. In less than a year, its valuation soared to between $1.5 billion and $2.5 billion. Both Mastercard and Coinbase are vying for this company, while Visa and Citigroup are participating through investments.
The significance of BVNK lies in its role as a bridge between the traditional fiat currency system and the rapidly expanding stablecoin network. In the context of AI payments, the value of this bridge is being redefined. Whoever controls it holds the key passage between the old and new financial systems.
For Mastercard, acquiring BVNK means quickly filling the gap in stablecoin infrastructure, avoiding marginalization in the new wave of technology. For Coinbase, it represents a strategic expansion opportunity, moving from an exchange to a broader payment field, aiming to build a Stripe for the crypto world.
The surge in BVNK's valuation reflects the market's re-evaluation of stablecoin infrastructure. In the era of AI payments, the roles these companies play are akin to clearinghouses in the traditional financial system; they handle not just transactions but also the underlying pipelines of value flow.
Protocols and infrastructure are the armaments, but the real battlefield is at the application layer. Whoever can get users accustomed to completing shopping on AI platforms will hold the throat of future commerce.
ChatGPT's instant checkout is a milestone event. This is the first shot in the transition of AI agent payments from concept to reality. Users can directly purchase items from Etsy in their conversation with ChatGPT, with the entire process requiring no navigation to the merchant's website. Stripe provides the payment infrastructure, and OpenAI provides the traffic entry, creating a brand new shopping experience.
This feature will soon expand to Shopify merchants, with brands like Glossier, Vuori, Spanx, and SKIMS already preparing to integrate. Sam Altman stated that this is the starting point of AI Commerce.
Google is also accelerating its actions. It announced that it will expand the shopping interface of AI Mode in the coming months, adding price tracking and direct purchase features. Users will be able to browse, compare, and place orders in AI Mode, with transactions ultimately completed through Google Pay.
Perplexity is also not falling behind. This AI search engine has launched the "Buy with Pro" feature in collaboration with PayPal, allowing users to check out directly in the chat interface. It has also integrated Firmly.ai, a platform backend that makes it easy for merchants to connect.
A report released by BCG on October 6 disclosed a set of key data. By July 2025, traffic to U.S. retail websites from GenAI browsers and chat services is expected to grow by 4700% year-on-year. The behavior of these users also differs from traditional visitors; they spend 32% more time on websites, view 10% more pages, and have a 27% lower bounce rate.
More importantly, when they arrive at websites, they are often already in the latter half of the purchasing decision process. Data from Adobe further confirms this, with over half of consumers expecting to use AI assistants for shopping by the end of 2025.
The entry points for traffic are being rewritten. In the past, people entered e-commerce websites through search engines or direct visits; now, AI platforms are becoming the new entry points. As consumers become accustomed to completing shopping in ChatGPT or Google AI Mode, the retail websites may gradually lose their significance.
The impact of this change is profound. The direct customer relationships that brands have spent decades building may be re-taken over by AI platforms. Consumer behavior data and transaction records will no longer belong to retailers but will flow into AI's databases.
In the past month, we have witnessed the comprehensive offensive of payment giants on three fronts.
At the protocol level, Stripe's ACP, Google's AP2, and Mastercard's Agent Pay are all competing for a core proposition: who will set the rules for AI agents. These protocols define how AI agents initiate payments, how they are authorized, and how they are held accountable. Whoever controls the protocols holds the discourse power in the era of AI Commerce.
At the infrastructure level, Visa's stablecoin pilot, Stripe's Open Issuance, and the bidding war surrounding BVNK are answering another question: who can control the pipeline of value flow. The trading volume of stablecoins has surpassed traditional payment networks, and it is becoming the preferred tool for AI agent payments. Whoever owns the stablecoin infrastructure possesses the clearing and minting rights of the new era.
At the application level, ChatGPT's instant checkout and Google's AI Mode are competing for the final gateway: who can become the new traffic entry point. As users begin to get accustomed to completing shopping on AI platforms, the retail websites and brand entry points are being quietly replaced. The shift in traffic signifies a transfer of commercial power.
These seemingly scattered actions are actually pointing towards the same goal: redefining the underlying rules of business operations at the moment when AI agents become the new type of consumer.
This is a restructuring of power, from humans to agents, from brands to algorithms, from payment networks to stablecoin infrastructure. Every technological revolution brings about a re-drawing of the power map, and AI payments are no exception.
In this war, what is perhaps most noteworthy is not who will win, but who will be excluded.
BVNK's valuation has tripled in less than a year, and this signal could not be clearer. The market is re-evaluating the entire payment ecosystem. Those companies still on the sidelines may find that they have missed the window to enter.
What has happened in the past month is not the starting point of change, but the beginning of acceleration. The contours of regulation have taken shape, technological capabilities have matured, and market demand has emerged. What remains is merely execution and competition.
A new commercial order is taking shape, and those companies that have not yet realized their positions have changed will pay the price in this reconstruction of order.
Related: Can SWIFT's blockchain challenge Ripple's dominance in the payment field?
Original: “What Have the Payment Giants Been Doing Over the Past Month”
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。
