Vying for the AI Payment Track: Traditional Card Networks Face Off Against Coinbase

Foresight NewsОпубликовано 2026-06-08Обновлено 2026-06-08

Введение

As AI agents increasingly conduct commercial transactions, a battle for control over the underlying payment infrastructure is unfolding. The competition centers on two divergent and incompatible technical approaches for autonomous AI payments. One camp, led by traditional card networks Visa and Mastercard, relies on tokenized card credentials within the established banking rails. Visa's "Intelligent Commerce" and Mastercard's "Agent Pay" services extend their existing tokenization technology to authorized AI agents for consumer retail transactions, leveraging decades of fraud protection and dispute resolution systems. Their partners include major AI firms like Anthropic, OpenAI, and Microsoft. The opposing camp, spearheaded by Coinbase, advocates for an open internet protocol using stablecoins. Coinbase's x402 protocol utilizes the HTTP 402 status code to enable direct, machine-to-machine micropayments with USDC on-chain. This model eliminates card fees and is designed for high-frequency, low-value transactions between AI agents, such as paying for API calls or data streams, where traditional card costs are prohibitive. Currently, application scenarios are clearly divided. Mainstream consumer-facing AI shopping services (e.g., ChatGPT's "one-click checkout," Amazon's AI-assisted shopping) predominantly use card channels due to their mature consumer protections and merchant networks. Conversely, the stablecoin channel dominates machine-to-machine payments, as seen in Amazon...


Author: Zennon Kapron

Compiler: Chopper, Foresight News


As AI agents increasingly undertake various commercial transactions, a battle for the underlying payment infrastructure is already underway.


Currently, the technical approach for enabling AI agents to autonomously consume diverges into two largely incompatible solutions: the channel through which the settlement and clearing of a transaction is ultimately completed when a software program acts as the payer. One camp constructs payment links based on tokenized bank card credentials controlled by Visa and Mastercard; the other, led by Coinbase, uses stablecoins to complete settlement based on open internet protocols. While the surface-level focus of AI agent commerce is shopping assistant applications, the core struggle behind it is actually about who will dominate the next-generation payment system.


Two Major Payment Channels, Suited for Different Application Scenarios


The traditional card networks moved first and acted swiftly. Mastercard launched its Agent Pay service in April 2025, built on its proprietary agent tokenization system. This tokenization technology was originally designed for contactless payments and card-on-file fast payment scenarios but has now been expanded to allow verified AI agents to complete transactions on behalf of users within authorized limits.


At launch, the service assembled a group of industry partners, signaling a clear strategic intent: collaborators included Microsoft, IBM's watsonx orchestration platform, and payment service providers Braintree and Checkout.com. A day later, Visa introduced its Visa Intelligent Commerce service, opening its payment network to AI developers, with AI-enabled bank cards as the core vehicle. This solution replaces the original card number with a tokenized credential to prove user authorization for a specific AI agent and define transaction boundaries. Visa also enlisted several top AI companies, including Anthropic, OpenAI, Perplexity, Mistral, and Samsung.


The solutions from both card networks keep transactions within the decades-old bank card payment model. AI agents are new actors, but behind them runs the same traditional payment channel that has served global commerce for half a century.


The stablecoin camp adopted an architecturally distinct solution. In May 2025, Coinbase launched the x402 protocol, reviving the long-dormant HTTP 402 "Payment Required" status code to enable direct settlement of transactions over the internet using the USDC stablecoin. The specific process is: a client requests access to a resource, the server returns a payment instruction; the client attaches signed stablecoin payment information to the request header; once the on-chain transaction is confirmed, the corresponding resource can be accessed normally. The entire process requires no account registration, card linking, and does not incur bank card transaction fees.


This solution is designed for machine-to-machine transactions. AI agents may need to complete thousands of micro-payments for API calls, data stream acquisition, or connecting with other agents. Such transactions are entirely unfeasible on traditional bank card channels from a cost perspective.


The two technical routes each have their strengths. The bank card channel excels in personal retail consumption scenarios, which place high demands on chargeback mechanisms, fraud protection, and dispute arbitration. The stablecoin channel demonstrates significant advantages in high-frequency, small-value, cross-border machine transactions, where traditional bank card fee structures and settlement timescales break down completely. The core of the contest lies in which type of scenario will become the mainstream for AI agent commercial transactions.


A major challenge facing both routes is identity verification. When a software program initiates a payment, merchants need to confirm that the operator is a legitimate agent authorized by a real user, not a malicious bot using stolen credentials. Simultaneously, users need a mechanism to request the reversal of a transaction mistakenly initiated by an AI agent.


Visa stated that AI traffic on US retail websites surged 47-fold, prompting it to collaborate with cloud service provider Cloudflare to launch a Trusted Agent Protocol for distinguishing legitimate AI programs from malicious crawlers. This highlights a structural advantage of traditional card networks: fifty years of accumulated risk scoring systems, chargeback rules, and dispute resolution mechanisms are well-suited to handle issues like an AI agent buying the wrong product. Stablecoin transactions, once on-chain, are permanent and irreversible, a problem for which no native solution currently exists within that system.


In the future, the key to winning the consumer-facing market may not be which payment channel has lower fees, but rather who can solve the challenges of agent identity verification and transaction dispute resolution.


Card Networks Hedge Their Bets, Covering Both Tracks


A telling signal is that Visa and Mastercard are not putting all their eggs in their own channel's basket; they are simultaneously investing in the stablecoin track.


As of April 2026, Visa's stablecoin settlement business reached an annualized transaction volume of $7 billion, a 50% increase quarter-over-quarter; the company added support for 5 new public blockchains, bringing the total number of partnered chains to 9, while also implementing over 130 "stablecoin + bank card" linkage projects in more than 50 countries. In October 2025, Visa doubled down by jointly launching the Trusted Agent Protocol with Cloudflare to help merchants distinguish legitimate agents from malicious programs, and publicly announced collaboration with Coinbase to promote interoperability between its network and the x402 protocol. Seemingly competing systems—the bank card system and stablecoin protocols—are now building bridges.


Mastercard has adopted a similar dual-track strategy. In March 2026, Mastercard announced its intent to acquire stablecoin platform BVNK for up to $1.8 billion. Prior to this, its Agent Pay service had already expanded to Latin America and the Caribbean, with adaptation completed for local card issuers in early 2026.


The core strategy of the two traditional card networks is evident: no longer simply defending the bank card channel, but striving to become the toll gate for all payment flows, whether through their own channels or stablecoin channels. This strategic move strongly indicates their judgment: if the industry ultimately settles on bank cards as the mainstream for AI payments, they would not need to invest heavily in acquiring stablecoin-related infrastructure.


Diverging Implementation Scenarios


Judging from currently launched products, the application boundaries of the two technical routes are quite clear.


Most mainstream products targeting ordinary consumers opt for the bank card channel. The "Checkout with ChatGPT" feature launched in September 2025, co-developed by OpenAI and payment service provider Stripe, relies on shared payment tokens to complete bank card clearing. These tokens are limited to specific merchants and shopping orders. It initially connected with Etsy sellers and later expanded to cover over a million Shopify stores. Amazon's "Buy for Me" feature, which calls upon AI agents to make purchases on third-party websites, automatically populates the user's linked bank card for settlement.


Personal consumption-oriented AI shopping services generally choose bank cards due to the system's mature anti-fraud tools, extensive merchant network, and long-established user trust.


Meanwhile, the stablecoin channel firmly occupies the machine transaction market. Amazon integrated the x402 protocol into its Bedrock agent core payment service, using Coinbase's Base public blockchain for settlement, with a single transaction taking about 200 milliseconds and fees under one cent; Stripe also joined the service as a payment integrator. According to Coinbase data, in its first year, the x402 protocol processed over 169 million payment orders, involving 590,000 buyers and 100,000 sellers.


These transactions are not typical user purchases like clothing; they are payments by AI agents for services like computing power, data, and API calls, where transaction frequency and individual amounts are incompatible with the logic of bank card design. In September 2025, Coinbase, together with Cloudflare, spearheaded the establishment of the x402 Foundation, aiming to promote industry-wide development of a universal standard rather than building a closed, proprietary product.


Summarizing five landmark AI commercial payment projects implemented by early 2026: 3 use bank card settlement, 2 use stablecoin settlement, with application scenarios largely divided along the lines of personal consumption and machine transactions.


Industry Outlook


In the short term, the industry landscape in 2026 is likely to maintain the status quo: bank cards dominate personal retail payments, stablecoins specialize in machine-to-machine transactions, with both coexisting and developing. However, by 2030, this situation may change, as both camps are vigorously competing for the converging zone between the two types of scenarios.


The ultimate deciding factor will depend on whether AI-driven commercial transactions ultimately lean more towards traditional retail forms or evolve into a massive network of micro machine transactions. If the former, traditional card networks will remain dominant; if the latter, the stablecoin channel will capture a large volume of entirely new transaction flows.


Visa and Mastercard have made the safest bet: hedging by investing in both tracks, ensuring they can collect fees regardless of where future transaction flows go. Those who need to be truly wary are companies betting solely on a single payment channel. The two major card networks have already mitigated this risk, a clear reflection of their assessment of the industry's future.

Связанные с этим вопросы

QWhat are the two main payment channels for AI agent autonomous transactions described in the article, and what are their primary use cases?

AThe two main channels are the traditional card scheme channel (led by Visa and Mastercard using tokenized cards) and the stablecoin channel (led by Coinbase using protocols like x402). The card channel excels in personal retail consumption scenarios requiring fraud protection and dispute resolution. The stablecoin channel is optimized for high-frequency, low-value, cross-border machine-to-machine transactions.

QWhat are the key strategic moves by Visa and Mastercard regarding stablecoins, and what does this indicate about their strategy?

AVisa and Mastercard are not solely defending their traditional card networks; they are actively investing in the stablecoin sector. Visa has expanded its stablecoin settlement volume, partnered with multiple blockchains, and even announced collaboration with Coinbase. Mastercard plans to acquire stablecoin platform BVNK. This dual-track strategy indicates their goal is to be the fee-collecting gateway for *all* payment flows, regardless of the underlying channel, hedging their bets on the future of AI commerce.

QAccording to the article, what is a major technical and operational challenge that both payment channels face in AI agent commerce?

AA major challenge for both channels is identity verification and transaction dispute resolution. Merchants need to verify that a payment is initiated by a legitimate AI agent authorized by a real user, not a malicious bot. Users also need mechanisms to dispute or reverse transactions made by AI agents in error. Traditional card schemes have decades of experience in risk scoring and dispute handling, while stablecoin transactions are typically immutable on-chain, lacking native solutions for chargebacks.

QHow do current AI payment implementations from major companies like Amazon, OpenAI/Stripe, and Coinbase reflect the split in payment channel use?

AMajor implementations clearly split based on the transaction type. For personal consumer-facing services, companies use the card channel: OpenAI/Stripe's ChatGPT 'one-click checkout' and Amazon's 'Buy for Me' feature both settle via tokenized cards. For machine-to-machine transactions, companies use the stablecoin channel: Amazon integrated Coinbase's x402 protocol into its Bedrock agent core payment service for fast, low-cost settlements for services like API calls and data.

QWhat does the article suggest is the likely determining factor for which payment channel becomes dominant in the long-term future of AI commerce?

AThe long-term dominance will be determined by whether AI-driven commercial transactions evolve to resemble traditional retail (favoring card schemes) or become a vast network of high-volume, micro-value machine transactions (favoring stablecoins). The ultimate 'decisive factor' is which of these two scenarios becomes the mainstream model for AI agent commerce.

Похожее

Jensen Huang 'Saves' South Korean Stock Market: Locks In SK Hynix Memory, Chip Shortage to Continue

On June 5th, South Korea's stock market experienced a sharp decline, with major chipmakers like Samsung and SK Hynix dropping nearly 10%. Amidst the turmoil, NVIDIA CEO Jensen Huang's visit to Seoul played a dramatic role in boosting market sentiment. Following a dinner meeting with SK Group Chairman Chey Tae-won and SK Hynix CEO Kwak Noh-Jung, Huang confirmed that NVIDIA's new Vera CPU will utilize SK Hynix DRAM. The companies announced a multi-year technical partnership to co-develop next-generation memory for NVIDIA's AI infrastructure, covering products from data centers to personal AI and robotics. This collaboration extends beyond memory supply. SK Hynix is integrating NVIDIA's AI and Omniverse platform into its own semiconductor design and manufacturing processes, including computational lithography and creating digital twins of its fabrication plants for autonomous operation. While strengthening ties with SK Hynix, NVIDIA is diversifying its supply chain for the upcoming HBM4 memory, with Samsung, SK Hynix, and Micron all certified as suppliers for its Vera Rubin platform. Despite this, Huang warned that the global chip shortage, driven by relentless demand from AI factory construction, is expected to persist for several years across the entire supply chain. His visit underscores NVIDIA's systematic effort to deepen integration with South Korea's broader tech industry.

marsbit3 мин. назад

Jensen Huang 'Saves' South Korean Stock Market: Locks In SK Hynix Memory, Chip Shortage to Continue

marsbit3 мин. назад

Nasdaq Plunges 4.2% in a Single Day: Does "Black Friday" Burst the U.S. Stock Market Bubble?

The Nasdaq plunged 4.18% on June 5, 2026, its worst single-day drop in over a year, as a much stronger-than-expected US jobs report triggered fears of economic overheating and delayed Federal Reserve interest rate cuts. The selloff, centered on high-valuation tech and AI stocks like Nvidia and Broadcom, spread across major indices. The article examines whether this signals a market top. The strong May non-farm payrolls data, nearly double expectations, pushed bond yields higher, directly hurting rate-sensitive tech stocks. This exposed vulnerabilities in the crowded AI trade, where valuations had soared on narratives of infinite growth, despite emerging signs of slowing order momentum and corporate AI monetization challenges. Prior to the drop, market indicators flashed warning signs: historically high valuations (e.g., Shiller CAPE ratio near 39.5), extreme bullish sentiment, and high levels of leverage. Technical charts showed key support levels being breached. Wall Street is divided on the outlook. Bears, citing risks of "stagflation" and AI bubble comparisons to the dot-com era, warn of a potential significant correction. Bulls view the drop as a healthy correction within a bull market, underpinned by a strong economy and expected corporate earnings growth of around 7% in 2026. The immediate future hinges on upcoming key events: the May CPI inflation data and the mid-June FOMC meeting. Their outcomes will critically shape market expectations for the Fed's rate path. The article concludes that conditions for a major market top are aligning, marking a fragile transition from narrative-driven gains to a phase demanding validation from macroeconomic data and corporate fundamentals. Caution is advised.

marsbit7 мин. назад

Nasdaq Plunges 4.2% in a Single Day: Does "Black Friday" Burst the U.S. Stock Market Bubble?

marsbit7 мин. назад

Nasdaq Plunges 4.2% in a Single Day, Did 'Black Friday' Pop the U.S. Stock Bubble?

The Nasdaq Composite plummeted 4.18% on June 5, its biggest single-day drop since April 2025, triggering widespread debate over whether the U.S. stock market has peaked. The sell-off was sparked by a stronger-than-expected U.S. non-farm payrolls report, which fueled fears of economic overheating and pushed back market expectations for Federal Reserve rate cuts, leading to a sharp rise in Treasury yields. The AI sector, the primary driver of the recent bull market, suffered severe losses, with the Philadelphia Semiconductor Index crashing over 10%. Stocks like Nvidia, Broadcom, and Micron led the decline. Concerns are mounting about the sustainability of AI capital expenditures and high valuations, with signs of order cuts for next-generation chips emerging. Analyses point to several warning signs: historically high market valuations (e.g., elevated Shiller CAPE ratio, Buffett Indicator), extreme bullish sentiment indicators, and significant insider selling. The sell-off also caused a key technical breakdown, with the S&P 500 breaking below its short-term moving average and testing its 200-day moving average. Wall Street is divided on the outlook. Bears warn this could be the start of a bubble deflation or a "stagflation" scenario, while bulls view it as a healthy, overdue correction within a bull market driven by solid corporate earnings growth. A more moderate view suggests the easy liquidity-driven rally is over, and markets are entering a phase of fundamental stock-picking with potential for consolidation. The immediate future hinges on key upcoming events: the May CPI report and the mid-June FOMC meeting. Their outcomes will be critical in determining whether this is a temporary pullback or the beginning of a more significant trend reversal. The consensus is that the era of one-directional market gains may be ending, requiring increased investor caution.

Odaily星球日报13 мин. назад

Nasdaq Plunges 4.2% in a Single Day, Did 'Black Friday' Pop the U.S. Stock Bubble?

Odaily星球日报13 мин. назад

The First Case on AI Agents: What Was Adjudicated?

"The First 'Agent' Ruling: What Was Decided?" On April 30, the Guangzhou Internet Court issued a ruling—China's first behavior preservation order in the intelligent agent (AI agent) field. The defendant, an open-source AI agent software, was ordered to stop downloads, cease actions that bypassed a platform's technical protection measures, and delete related tutorials and data. The core issue: the software used the operating system's "accessibility service" permissions to automate user interactions within other apps without those platforms' authorization. This mirrors a recent US case where Amazon sued Perplexity for similar reasons—bypassing Amazon's API to directly scrape and interact with its pages—and won a preliminary injunction. Both rulings establish a crucial legal boundary for the AI agent era: agents cannot operate unchecked. The article argues the fundamental legal principle emerging is one of **dual authorization**. An AI agent requires both **user consent** AND **platform consent** to operate legitimately within that platform's ecosystem. Bypassing platform rules through system-level permissions, even with user permission, undermines platform responsibilities for content moderation, data security, and user privacy, creating liability issues. The piece uses the evolution of "Doubao Phone" (an AI-integrated smartphone) as a case study. Its initial, aggressive version that bypassed platform controls faced roadblocks. Its upcoming 2.0 version is reportedly pivoting to negotiate API access and authorization deals with major platforms (like Alibaba's ecosystem), seen as a strategic adaptation to the new regulatory reality. A global trend is identified: the era of unregulated, "wild west" growth for AI agents is ending, replaced by a **compliance race**. This raises barriers to entry, as securing platform authorizations becomes a new cost. Open-source status is also not a legal shield if the code facilitates bypassing technical protections. In conclusion, these first rulings target not the largest, but the most **aggressive and representative** cases. By setting precedent with them, regulators are efficiently steering the entire industry towards a new, more regulated operating paradigm defined by dual authorization and platform cooperation.

marsbit18 мин. назад

The First Case on AI Agents: What Was Adjudicated?

marsbit18 мин. назад

Fired by Google Over a 14-Page Paper, Over 4,000 Rallied for Her. 6 Years Later: She Almost Predicted the Entire AI Era Back Then.

In late 2020, Google AI researcher Timnit Gebru was effectively dismissed following a conflict over a 14-page, unpublished research paper she co-authored titled "On the Dangers of Stochastic Parrots." The paper, which has since been cited over 14,000 times, raised critical early warnings about the risks of large language models (LLMs). It argued that these models, trained on vast, biased internet data, are essentially "stochastic parrots" that mimic language without true understanding, potentially amplifying societal biases, generating plausible but false information (later termed "AI hallucination"), consuming massive energy, and obscuring their training data contents. Gebru's stance led to a clash with Google management, who requested the paper's withdrawal. Her subsequent internal criticism of the company's diversity efforts and handling of the matter culminated in her termination, which sparked protests from over 4,000 Google employees and researchers. Six years later, the paper's predictions have proven remarkably prescient. Issues like AI hallucination, embedded bias (evident in resume screening and healthcare algorithms), soaring energy consumption from AI data centers, unvetted training data containing harmful content, and the risk of "model collapse" from AI-generated internet content have become central industry challenges. The incident also highlighted concerns about AI development being driven primarily by commercial competition within a handful of powerful tech companies, often at the expense of ethical considerations. After leaving Google, Gebru founded the Distributed AI Research Institute (DAIR) to explore these issues independently. The controversy underscores how her early, critical insights into the fundamental limitations and societal impacts of LLMs anticipated many of the most pressing dilemmas in today's AI era.

marsbit19 мин. назад

Fired by Google Over a 14-Page Paper, Over 4,000 Rallied for Her. 6 Years Later: She Almost Predicted the Entire AI Era Back Then.

marsbit19 мин. назад

Торговля

Спот
Фьючерсы

Популярные статьи

Неделя обучения по популярным токенам (2): 2026 может стать годом приложений реального времени, сектор AI продолжает оставаться в тренде

2025 год — год институциональных инвесторов, в будущем он будет доминировать в приложениях реального времени.

1.8k просмотров всегоОпубликовано 2025.12.16Обновлено 2025.12.16

Неделя обучения по популярным токенам (2): 2026 может стать годом приложений реального времени, сектор AI продолжает оставаться в тренде

Обсуждения

Добро пожаловать в Сообщество HTX. Здесь вы сможете быть в курсе последних новостей о развитии платформы и получить доступ к профессиональной аналитической информации о рынке. Мнения пользователей о цене на AI (AI) представлены ниже.

活动图片