IDC Analyst Highlights OpenAI's 'Structural Limitations' in Agentic Commerce and Acquisition Strategy

IDC Publishes Structural Analysis of OpenAI's Commerce Strategy

Is there a play for OpenAI in agentic commerce?

OpenAI's push into agentic commerce reveals structural limits. Explore why data ownership—not interfaces—will define winners in AI-driven commerce.

On March 18, 2026, IDC Research Director Gerry Murray published an analysis report on OpenAI's agentic commerce strategy. The report digs into why OpenAI failed to scale "pay in chat" and highlights a structural issue: "data ownership, not AI interfaces, determines the winner."

The analysis follows The Information's early March report on OpenAI's Instant Checkout retreat, adding a significant perspective to the industry-wide battle for agentic commerce leadership.

Background and Industry Trends

OpenAI launched "Instant Checkout" on ChatGPT in September 2025, but scaled it back just five months later in March 2026. Forbes' Jason Goldberg reported that only about a dozen Shopify merchants actually went live during that period. Even tax processing systems were reportedly not in place, revealing a fundamental lack of commerce platform infrastructure.

Meanwhile, platform-side AI implementations are advancing rapidly. Amazon's Rufus is used by over 300 million customers and contributes approximately $12 billion in incremental annual revenue. Walmart reports that [50% of mobile app users engage with AI assistant "Sparky," with those sessions showing 35% larger cart sizes](https://www.forbes.com/sites/jasongoldberg/2026/03/10/why-openais-checkout-retreat-spells-trouble-for-its-commerce-strategy/). In other words, "vertical agents" that own product inventory, pricing, and fulfillment are delivering results.

The "Three Strategic Assets" OpenAI Needs and IDC's Acquisition Candidates

IDC's Murray argues that for OpenAI to become a true agentic commerce platform, it needs to own or control at least one of three strategic assets.

These are the Commerce Graph (inventory, seller, and transaction data), Customer Graph (identity, purchase behavior, and lifecycle data), and Product Data Graph (high-frequency usage data and intent signals).

Based on this analysis, IDC identifies five specific acquisition candidates.

The highest-rated target is Klaviyo, aimed at acquiring the Customer Graph. It holds behavioral data from over 193,000 merchants and processes billions of events including email, SMS, and purchase conversions. With revenue growth of approximately 32% and net revenue retention exceeding 109%, its financial foundation is solid.

Next is Instacart. As one of North America's largest grocery delivery platforms, it processes approximately 33.8 million orders and over $37 billion in GTV. Groceries represent weekly or biweekly "high-frequency commerce," creating an ideal data loop for AI agent habit formation.

BigCommerce (including Feedonomics) provides structured product data feeds to major marketplaces including Amazon, Google, and TikTok, controlling a key element of the "Product Data Graph." Etsy complements with discovery commerce preference data, while Cart.com adds fulfillment operations data.

Murray analyzes that if OpenAI were to acquire the top three — Klaviyo, BigCommerce, and Instacart — it would gain multiple revenue streams including AI SaaS, transaction fees, retail media, and merchant services, effectively redefining itself as "the AI OS of commerce."

The Battle for the "Personal Shopper"

Another critical argument in the IDC report concerns who will own the consumer's personal shopper agent. Murray points out that consumers may ultimately default to a single AI agent, and identifies two structural vulnerabilities for OpenAI.

First, without exclusive distribution agreements with major device manufacturers, OpenAI remains a "third-party app." Second, if Gemini-powered Siri deploys across billions of Apple devices, overcoming that would be extremely difficult.

Amazon's $50 billion investment in OpenAI is primarily a cloud infrastructure (2GW of Trainium chips on AWS) and enterprise distribution partnership — it does not include an agreement to sell Amazon products within ChatGPT. Forrester has called "OpenAI + Amazon" a potential winner in consumer agentic commerce, but the premise is data integration, not interface provision.

Impact on E-commerce Businesses and How to Leverage

The biggest insight from IDC's analysis is that "AI interface providers" and "data owners" are different players, and the latter determines the winner. The specific actions for e-commerce operators are clear.

Structuring data is urgent. The Agentic Commerce Protocol (ACP) being jointly developed by OpenAI and Stripe is advancing standardization, and preparing product catalogs, inventory, and pricing data that AI agents can read becomes the entry ticket.

Strengthen first-party data ownership. The very fact that Klaviyo is listed as an acquisition candidate speaks to the strategic value of customer behavioral data. Solidifying your first-party data foundation with CDPs and marketing automation tools becomes the defensive line in the agentic era.

Prioritize vertical strategies. As Forbes' Goldberg points out, "full-stack vertical agents" like Amazon Rufus and Walmart Sparky are delivering results. Building an AI experience that integrates your own product inventory, pricing, and fulfillment is more reliable than depending on horizontal agents.

Summary

IDC's analysis makes clear that the main battleground for agentic commerce is shifting from "whether you can buy within AI chat" to "who controls the three graphs of commerce, customer, and product data." OpenAI's scaling back of Instant Checkout was not due to lack of consumer demand, but rather resource allocation priorities and insufficient data assets.

For e-commerce operators, the key question is whether OpenAI will pursue an acquisition strategy going forward. The trajectories of Klaviyo, Instacart, and BigCommerce have the potential to significantly reshape the agentic commerce landscape. In any scenario, preparing your customer data and structured product data to be "readable by AI agents" is the most reliable preparation.