Criteo's Agentic Commerce Play: Assessing Its Position on the AI Shopping S-Curve

The shift to agentic commerce is not a distant future-it is the next phase of the digital shopping S-curve. This isn't about replacing traditional e-commerce; it's about expanding the ecosystem. The macro trend is clear: a new channel is forming. Market projections point to a staggering $5 trillion in global agentic commerce volume by 2030. For context, that represents a multi-trillion dollar incremental channel on top of today's digital retail base.

Yet adoption remains in its infancy. The data shows a massive gap between consumer interest and current transaction flow. While 39% of consumers already use AI for product discovery, and traffic to retail sites from AI tools has exploded, the actual commerce happening is minimal. In the 2025 holiday season, LLMs and other AI tools remained less than 1% of total traffic for retail and travel merchants. More telling is that 23% of Americans bought something via AI in the past month, a figure that highlights real, recent demand but also the nascent state of the infrastructure to support it at scale.

This is where the core friction lies. AI shopping assistants are powerful, but they lack the commerce-grade intelligence needed to drive trusted, transaction-ready recommendations. They are built on generic language models, not the rich behavioral data of actual shoppers. This creates a visibility gap: retailers lose track of the customer journey the moment it moves into an AI assistant's interface. The result is a performance cliff. ChatGPT referrals convert 86% worse than affiliate links, and true agentic payments are at the very early stages of deployment.

Criteo's new service is a direct play on this infrastructure gap. It targets the fundamental need: AI assistants require commerce-grade data, not just product descriptions. By leveraging its unmatched scale of 720 million daily shoppers and $1 trillion in annual transactions, Criteo aims to provide the real-world shopping signals that generic models lack. Its Agentic Commerce Recommendation Service is built to deliver the trusted, personalized results consumers expect, potentially improving recommendation relevancy by up to 60%. In this paradigm shift, Criteo is positioning itself not as a retailer, but as the essential data layer that will power the next generation of shopping interfaces.

Criteo's Current Position: Market Context and Financial Reality

The stock market is pricing Criteo as a company in a holding pattern. Shares trade at $18.52, down roughly 10% year-to-date and a steep 22% over the last 120 days. This underperformance is reflected in its valuation, with a trailing P/E of just 5.6. The consensus from Wall Street is a cautious "Hold," with a wide range of price targets from $27 to $47, averaging $38.11. That spread signals significant uncertainty about the company's near-term path. For an investor betting on a paradigm shift, this setup presents a tension: the stock is cheap, but the market is clearly skeptical about the near-term catalysts.

Yet beneath this financial reality lies a formidable infrastructure asset. Criteo's core Commerce AI platform already operates at a scale that is foundational for any agentic commerce play. It leverages a 720 million daily active users and a catalog of 4.5 billion product SKUs. This isn't just data; it's the real-world shopping signal network that generic AI models lack. The company's new Agentic Commerce Recommendation Service, announced today, is a direct application of this existing scale, aiming to translate shopper queries into transaction-ready results. The internal testing showing up to a 60% improvement in relevancy is a proof point for this data advantage.

The financial metrics suggest the company has the runway to invest in this future. With a market cap of $947 million and an enterprise value of $669 million, the valuation is low. The P/E of 5.6 implies the market is discounting future growth heavily. This could be a rational assessment of near-term headwinds, but it also means the cost of capital for funding the next phase of development is relatively low. The key question is whether the current financial position provides enough flexibility to fund the aggressive scaling needed to capture a share of the nascent agentic commerce S-curve before competitors build their own data moats. The stock's steep decline may reflect that doubt.

Infrastructure Moat and Competitive Landscape

Criteo's new service is built on a strategic foundation: the Model Context Protocol (MCP). This isn't just a technical detail; it's a critical moat. MCP standardizes how AI agents connect to external tools, and since its introduction in late 2024, it has seen rapid adoption by Microsoft, OpenAI, and Google. By anchoring its Agentic Commerce Recommendation Service to MCP, Criteo positions itself as a plug-and-play solution for the dominant AI infrastructure. This gives it immediate access to a vast network of AI agents and developers, accelerating its path to integration. The protocol's rapid adoption is the first-order signal that the agentic commerce ecosystem is forming around standardized interfaces, and Criteo is betting it can become the essential commerce tool within that stack.

Yet the durability of this moat faces a fundamental test: the performance gap. The market is showing real demand, but the current infrastructure is broken. Data reveals that ChatGPT referrals convert 86% worse than affiliate links. This is the core problem Criteo's service aims to solve. The gap stems from a visibility cliff: when a shopper uses an AI assistant, the rich behavioral data that powers effective recommendations and attribution vanishes. Criteo's solution is to plug that gap with its own data. Its internal testing suggests the service can leverage its scale to deliver up to a 60% improvement in recommendation relevancy. If this translates to real-world conversion lifts, it directly addresses the friction holding back the entire channel. The service's value is in closing the loop between discovery and transaction, a function that generic models cannot provide.

The competitive landscape, however, is not static. The biggest threat is the potential for LLM platforms themselves to build in-house recommendation engines. If Microsoft or OpenAI decides to internalize this commerce layer, they could bypass Criteo's intermediary role. This is the classic "platform as competitor" risk. Other data providers with retail or consumer behavior data could also emerge as rivals. Furthermore, the security vulnerabilities identified in MCP implementations-like command injection and server-side request forgery-pose a tangible risk to Criteo's own service if not rigorously managed. These are not theoretical concerns; they are the early signs of a crowded, high-stakes race to own the infrastructure layer of the next shopping paradigm. Criteo's advantage is its unmatched real-world shopping signal network, but the race is on to see if that data moat can be built faster than the platforms can replicate it.

Adoption Scenarios and 2026 Catalysts

The financial impact of Criteo's agentic commerce play hinges on a single, critical transition: the shift from AI-driven traffic to AI-driven transactions. The company's new service is designed to be the essential data layer that closes the visibility gap, but its revenue potential is directly tied to the acceleration of true agentic payments. This is the primary catalyst for 2026.

Right now, the channel is stuck in a traffic trap. While AI tools are driving traffic to retail and travel merchants that has increased more than 10x over the last 15 months, that traffic is not converting. The performance cliff is stark: ChatGPT referrals convert 86% worse than affiliate links. The new service aims to change that by providing the commerce-grade intelligence that generic models lack. The primary financial impact will be a new revenue stream from serving both the LLM platforms building the agents and the retailers building their own AI assistants. It supplements Criteo's existing Commerce AI offerings by monetizing its unmatched shopper graph data for this nascent channel.

The key catalyst for validating this thesis is the acceleration of true agentic payments. This is the feature at the very early stages of deployment, where an AI agent can initiate and complete checkout on a customer's behalf. Flagship research notes that true agentic payments are at the very early stages of deployment. When this feature moves from pilot to production, it will create an immediate, high-value demand for the kind of trusted, transaction-ready recommendations Criteo's service promises. The company's internal testing suggests a potential 60% improvement in recommendation relevancy, which could directly translate into higher conversion rates from AI traffic. This is the moment the service moves from a theoretical data advantage to a revenue-critical infrastructure component.

For investors, the watchlist for 2026 is specific. The first milestone is adoption velocity. Watch for the number of LLM partners integrating the service via the Model Context Protocol. Rapid adoption by major players like Microsoft, OpenAI, and Google would signal strong network effects and a faster path to scale. The second, more telling metric is the improvement in conversion rates from AI traffic. Early signs of a lift-say, from the current less than 0.2% of e-commerce sessions coming from ChatGPT referrals-would be the clearest signal that the visibility gap is closing and the service is delivering tangible value. Success here could re-rate the stock from a holding pattern to a growth play on the agentic commerce S-curve. Failure to gain traction would confirm the market's skepticism and likely cement the stock's low valuation.