Agentic Commerce and Crypto

Contents

• 1. Current state of the Internet, advertising, and the original sin
• 2. How is the ad landscape changing?
• 3. New business models and how crypto comes into the mix
• Conclusion

Introduction

Traditional search is evolving before our eyes. As if Google’s $75 billion planned expenditure in AI in 2025 wasn’t clear enough, the highlight of I/O 2025 was how AI would become embedded across all of their products.

At a faster and faster pace, it’s becoming clear that the interface of the internet is about to change. Agents can offer much deeper recommendations with fewer steps, simplifying shopping experiences while enabling users to make more informed purchases. However, this change calls into question much of the financial plumbing that the internet has been built off of.

How does the business model of eCommerce work in the age of agents? How do advertisements work when traditional page rank search shifts? And how does payment infrastructure support agentic checkout processes?

Importantly, this transition represents an opportunity to fix "the original sin of the web" -- and crypto might be the key to doing it right this time.

Agenda:

1. Current state of the Internet, advertising, and the original sin
2. How is the ad landscape changing?
3. New business models and how crypto comes into the mix

1. Current state of the Internet, advertising, and the original sin

The web's financial foundation rests on what Ethan Zuckerman calls "the original sin of the web" - advertising as the default model to support online content and services.

Interestingly, through the evolution of the early Internet, many of today's large tech companies initially resisted ads.

Google: "We expect that advertising-funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers." - Page & Brin, The Anatomy of a Large-Scale Hypertextual Web Search Engine (1998)

Yet, over time, the incentives of the advertising business model made it the most powerful economic force of the past two decades. More than anything, advertising allowed platforms to monetize the attention flow of free users. At the same time, advertising was born from necessity. As Marc Andreesen has put it, the early web could not build payments into browsers because credit card companies and banks wouldn't cooperate.

However, this dominant advertising model set a new social standard for privacy by explaining every violation of privacy as the price that consumers needed to pay for “free” internet services. Shoshanna Zuboff writes in The Age of Surveillance Capitalism that these justifications hid the emergence of a new market—that of surveillance capitalism, a market that claims human experience as raw material that can be translated to behavioral data.

Much of last cycle's talk on Web3 and the vision of a free and permissionless Internet stemmed from a rebellion against this surveillance capitalism world. This model worked when humans were the primary interface to the web. But AI agents are about to break it entirely.

2. How is the ad landscape changing?

The stage is now set for agents to drive the new search interface. AI agents fundamentally disrupt the advertising model because they drive recommendations based on data, not ad placements. A large part of their appeal is their ability to be unbiased and circumventing sponsored listings.

Beyond that, many of the incentives underlying the ad model are also changing, opening up the scene to new players and interfaces.

1/ Shift to eCommerce advertising: Advertisers are racing to drive ROI, and in this paradigm, platforms that have the closest proximity to buying decision are winning. Consumers have become increasingly immune to traditional advertising methods that are focused around attention, a step above conversion in the user funnel. While Google can monetize a user intent to purchase, Amazon comes one step closer by owning the actual platform of purchase. As a result, their proximity to buying has been supercharging their ad business.

2/ Shift from social media to media (it's TikTok's world, and we're just living in it): TikTok ad revenue is projected to hit $32B in 2025, a 24.5% YoY growth. This underscores a shift from consuming content by friends to consuming content from other creators as a replacement for other media channels. As AI content proliferates, trust in purchasing comes from finding "genuine individuals" who can give ostensibly honest opinions.

3/ Death of attention based monetization -- "gray bots":

AI bot requests, which can come in the form of data crawls or agent-led browser automation, are obscuring traditional ad metrics and throwing off publishers. For example, Wikipedia reported that crawlers drove a 50% rise in infrastructure costs

since Jan 2024. For eCommerce businesses, this could mean inflating engagement metrics, triggering ad calls without delivering value, and completely distorting campaign performance. For other websites, this means that gated or ad-driven content would not be monetized, even though the same content could be leveraged by LLMs to deliver value to users.

Why does this matter?

In response to these trends, new and existing companies are racing to build the apps and infrastructure for a world of agent led content and commerce. These apps:

1. Integrate buying into the search process
2. Integrate media and entertainment instead of just search
3. Consider different methods of monetization, like affiliate revenue
4. Automates the end-to-end commerce journey within one platform, hoping to decrease drop off at every point in the journey. Agents are embedded all the way through, from discovery to payment

The proof is in the pudding. Here are examples of how players are shaking up the user journey:

Agentic commerce integration

• OpenAI poised for a partnership with Shopify
• Microsoft has launched their , an in-chat storefront to eligible sellers Copllot Merchant Program
• Perplexity has a shopping feature for pro users
• @doji_com recently raised $14M for a virtual try-on app powered by their in-house diffusion models. One of the next steps is to integrate the buying process in the app
• Daydream raised $50M last year to build a new e-commerce search engine leveraging AI
• Phoebe Gates launched Phia, a mobile app/browser extension that compares pricing from 40k retailers online + surfaces lower cost alternatives

All of these new interfaces are also demanding agent-led payment processes to reduce friction and own the buying step of the user journey. This means payment processors and infra need to adapt to a new paradigm where agents are able to transact.

Agent-enabled payment infrastructure

• Stripe has been consistently releasing new agentic workflows. They've already released issuing that creates single-use virtual cards that allows agents to spend funds. Last month at Stripe Sessions, they demo'd Order Intents API that allows you to create commerce agents that checkout on your behalf via browser automation.

• Visa recently introduced a framework to allow AI to shop and buy Visa Intelligent Commerce,
• Mastercard introduced Agent Pay around the same time as Visa last month. Agent Pay is an agentic payments program that integrates seamless payments into conversational AI platforms. They're working with Microsoft alongside acquirers and checkout players like Braintree and checkout.com to deliver agentic payments.
• PayPal has introduced an that incorporates PayPal API capabilities into agent workflows agent toolkit
• Obviously agents managing payments introduces a new world of problems and worry around managing credentials, authorizations, and setting up security guardrails. New players are also fundraising to try and solve agent authorization, like Nekuda ($5M round in May led by Madrona + Amex Ventures and Visa Ventures).

Here's another article from @drmelseidy that overviews the changing commerce tech stack:

3. New business models and how crypto comes into the mix

As the nature of transacting on the Internet is changing, so comes a new chance to right the original sin. As agents disrupt the current economic model of content creation, there may be new ways to monetize.

1/ Stablecoins for payments

Probably the highest probability and largest TAM way that crypto can help address a world of agentic commerce is via stablecoins, where crypto rails can help drive payments. Stablecoins solve the core payment infrastructure problem that Marc Andreesen explained was missing in the 1990s:

• No transaction fees for small payments
• Infinite divisibility allowing payments worth fractions of a cent
• Programmable logic that agents can execute automatically
• Global and permissionless accessibility without traditional banking infrastructure

Despite the long list of teams driving AI shopping agents, the reality is that current infrastructure does not support agentic payments. Perplexity's AI shopping agent can take up to a 3-8 hour checkout process, with frequent failures.It's no surprise that

Coinbase recently introduced x402, a payment protocol that enables instant stablecoin payments over HTTP, allowing agents to transact seamlessly.

Additionally, @tryskyfire is a financial stack for agentic commerce that empowers agents to make and receive payments. Importantly, they enable agents to transact with all kinds of digital resources, including LLMs, API services, websites, etc. Skyfire leverages stablecoins to enable agents to pay programmatically, meaning no manual intervention or pre-established accounts needed.

While stablecoins can initially help solve problems with payment infrastructure and agents, the broader vision is that they can enable new revenue streams all together. For example, Skyfire writes of a world where where agents can pay websites to access content, thereby letting websites monetize without ads.

2/ A shift to affiliate revenue

When responding to how OpenAI could drive a sustainable business model in a

Stratechery interview, Sam Altman responded

SA: The kind of thing I’d be much more excited to try than traditional ads is a lot of people use Deep Research for e-commerce, for example, and is there a way that we could come up with some sort of new model, which is we’re never going to take money to change placement or whatever, but if you buy something through Deep Research that you found, we’re going to charge like a 2% affiliate fee or something. That would be cool, I’d have no problem with that. And maybe there’s a tasteful way we can do ads, but I don’t know. I kind of just don’t like ads that much.

Agent led commerce disrupts the traditional ad model as people go to LLMs for unbiased information. Ad placements are antithetical to what makes LLMs so effective right now. While OpenAI is currently struggling to monetize, given that 96% of their users are free, affiliate revenue may be a way to help agents/AI companies monetize while preventing ad placements.

However, affiliate marketing is much harder in an agent led world, as traditional affiliate links drop cookies into user browsers to track user information and confirm revenue rates. When agents drive checkout experiences, many of the solutions rely on browser automation, meaning much of the attribution information does not get captured.

@henrysocialxyz is working on confirming attribution information in agent sessions via zkTLS, enabling agents to monetize day 1.

3/ New ad models -- Cost Per Result

With the increasing gray bot traffic disrupting traditional ad metrics, cost-per-click is increasingly ineffective. As brands try to get closer to downstream purchasing, the only ROI that truly matters is actual user growth / purchase. Crypto, via user wallets and stablecoins, enables brands to directly pay users for their attention, as opposed to paying publishers in the middle.

@earnos_io uses zkTLS to verify direct user interaction with brands, and leverages stablecoins to payout to users.

Conclusion

The shift from human-driven search to agent-mediated interactions is an opportunity to fundamentally realign the internet's economic incentives.

Where the advertising model created a three-way tension between users, content creators, and platforms, agentic payments via stablecoiuns could create direct value exchange. Users (or their agents) can pay for premium content access or get paid themselves for proof of real work. Creators could monetize their work directly, and brands could directly drive ROI by engaging on the purchasing level. And underlying all of this, agents could facilitate frictionless transactions at scales impossible for humans.

Additional Reference

• The Siren Song of Advertising, Eric Flaningham
• The Evolution of Digital Advertising, Eric Flaningham
• The Current State of Digital Advertising, Eric Flaningham
• The Agentic Web and Original Sin, Stratechery