"For twenty years, retail tech debates were about platforms: Shopify vs. Magento, Amazon vs. direct-to-consumer. In 2026, the debate has shifted to protocols. The companies that control how AI agents discover, negotiate, and purchase products will control the next era of commerce. The protocol wars have begun."
Fred Skoler, AI Shopper News, analyzing 1,500+ articles from 310 sources
🏛️ The Protocol Wars: Open vs. Closed
On January 11, 2026, at the National Retail Federation conference, Google announced the Universal Commerce Protocol (UCP), an open-source standard for agentic commerce. It was co-developed with Shopify, Etsy, Wayfair, and Target, endorsed by over 20 global partners including Visa, Mastercard, American Express, Stripe, Best Buy, Macy's, Home Depot, and Walmart.
The bet: create a common language so any AI agent can discover products, check inventory, negotiate discounts, and complete checkout with any retailer. One protocol. No custom integrations. Since launch, Klarna has joined, Salesforce has built Agentforce Commerce on top of it, and the protocol has generated 14 articles in our database in just one month.
Four days before Google's announcement, Amazon closed its own stores. Not metaphorically. Amazon announced it was shutting down Amazon Go and Amazon Fresh locations, shifting investment away from physical retail toward AI-driven commerce. Amazon's response to the protocol question has been to build its own closed ecosystem rather than adopt an open standard.
On February 5, Amazon announced Q4 2025 results: $213.4 billion in revenue and a staggering plan to invest $200 billion in capital expenditure in 2026. CEO Andy Jassy revealed that Rufus, Amazon's agentic AI shopping assistant, was used by over 300 million customers in 2025, generating nearly $12 billion in incremental annualized sales. Rufus can now shop tens of millions of items on other retailers' websites and make purchases autonomously through its "Buy for Me" feature.
The Insight: This is the defining split of 2026. Google says: "Let's build an open standard so all agents can shop everywhere." Amazon says: "Our agent already shops everywhere, and 300 million people are using it." The two approaches represent fundamentally different visions for how AI commerce will be structured.
| Dimension | Google (UCP) | Amazon (Rufus) |
|---|---|---|
| Philosophy | Open protocol, any agent | Closed ecosystem, one agent |
| Partners | 20+ including Shopify, Walmart, Visa | Self-contained |
| User Base | 1.5B Google AI monthly active users | 300M Rufus users |
| Revenue Model | Ads + Commerce enablement | Transaction ownership |
| Merchant Control | Merchant remains Record owner | Amazon controls experience |
| 2026 Investment | $185B (Alphabet total capex) | $200B (Amazon total capex) |
📈 From Endcap to Every Aisle: 24,200% Growth in 12 Months
Our data reveals a dramatic coverage shift. In January 2025, the term "agentic" appeared in 1 article in our database. By February 2026, it appears in 243 articles with 386 total mentions.
The catalyst? Three events in rapid succession. In September-October 2025, Perplexity launched its Buy with Pro feature, OpenAI added shopping to ChatGPT, and Walmart partnered with OpenAI to enable shopping through ChatGPT. By November, the financial infrastructure started responding: Visa flagged escalating fraud risks from agentic commerce, then partnered with AWS to build agentic commerce infrastructure. Morgan Stanley projected AI agentic shoppers could drive $385 billion in online sales.
The term didn't just grow. It became the category. Our Agentic Commerce category now contains 278 articles. For comparison: Cashierless Stores has 49. Virtual Try-On has 70. A concept that barely existed 12 months ago now generates more coverage than technologies that have been developing for a decade.
What about the dip? Sharp-eyed readers will notice January’s 70 articles and February’s pace trailing December’s peak of 95. That’s not decline. It’s normalization, and the pattern is worth examining. In retail terms: agentic commerce moved from the promotional endcap to permanent shelf space. December’s spike was amplified by year-end trend roundups, holiday shopping analyses, and “2026 prediction” pieces. By January, “agentic” stopped being the headline and started being the assumed context. Articles about payment infrastructure, search, and checkout now treat agentic capabilities as a given rather than a novelty. When a concept stops generating its own coverage because it’s embedded in everyone else’s coverage, it suggests the concept has moved from novelty to infrastructure.
| Category | Articles | vs. Jan 2026 | Trend |
|---|---|---|---|
| Retail AI | 1,180 | +126 | Established leader |
| Agentic Commerce | 278 | +70 | 🔥 Fastest growing |
| Payment Technology | 196 | +24 | Surging (agentic payments) |
| Supply Chain AI | 122 | +8 | Steady |
| Shopping Assistants | 105 | +5 | Consolidating |
| Computer Vision | 86 | +3 | Maturing |
The Insight: Agentic commerce increasingly overlaps with every other category in our database. Payments, search, checkout, and personalization stories now routinely reference agentic capabilities, suggesting the concept is becoming a cross-cutting layer rather than a standalone category.
⚔️ The Three Gatekeeper Models
Our data reveals three distinct approaches to controlling the AI shopping experience. Each model has a clear champion, growing ecosystem, and fundamentally different philosophy about who owns the customer.
| Model | Champion | Coverage | Key Partners |
|---|---|---|---|
| The Open Protocol | Google (UCP) | 200 mentions, 110 articles | Shopify, Walmart, Target, Visa, 20+ |
| The Walled Garden | Amazon (Rufus) | 396 mentions, 235 articles | Self-contained ecosystem |
| The Platform Layer | OpenAI (ChatGPT) | 166 mentions, 98 articles | Walmart, Shopify, Klarna, PayPal |
Google's open protocol play (51 Gemini mentions) positions UCP as the TCP/IP of commerce. As Shopify's engineering team wrote when announcing their co-development: the protocol models the complexity of real commerce, from payment handlers to fulfillment options, so any AI agent can transact with any merchant. Google keeps the search gateway. Retailers keep the customer relationship.
Amazon's closed ecosystem play needs no protocol because it is the protocol. With 300 million Rufus users, $12 billion in incremental sales, and the ability to shop on competitor websites through Buy for Me, Amazon is building a self-contained commerce loop. Customers who use Rufus are 60% more likely to complete a purchase, giving Amazon a significant retention advantage.
OpenAI's platform layer play takes a different approach. ChatGPT already has 800 million weekly active users. With Walmart integration, Shopify checkout (merchants pay 4% on ChatGPT-driven sales), and PayPal for instant checkout, OpenAI doesn't need its own commerce infrastructure. It routes intent to whoever can fulfill it. As one Retail TouchPoints headline put it: “When Bots Shop: Why Brands Should Stop Ignoring ChatGPT Commerce.”
💰 Follow the Money: The Infrastructure Race
The investment numbers tell the story of how seriously Big Tech is taking this moment:
| Company | 2026 Capex | Key AI Commerce Move |
|---|---|---|
| Amazon | $200B (up ~50%) | Rufus: 300M users, Buy for Me, $12B incremental sales |
| Alphabet/Google | $185B | UCP launch, Business Agent, AI Mode in Search |
| Microsoft | ~$80B+ | Copilot shopping + Shopify checkout integration |
Combined, these three companies are investing nearly half a trillion dollars in 2026 capex, with AI commerce as a primary use case. Meanwhile, analyst forecasts continue escalating: McKinsey projects agentic commerce could generate $3-5 trillion globally by 2030. Morgan Stanley estimates $385 billion in AI-driven online sales. Gartner projects AI agents will command $15 trillion in B2B purchases by 2028.
The Insight: When companies invest $500 billion and analysts project $15 trillion in outcomes, the question isn't whether agentic commerce will happen. It's whether your business is ready for the protocol wars that will determine how it happens.
💡 What Our Sources Reveal: The Ecosystem is Exploding
Perhaps the most telling metric in our entire dataset isn't about any single company. It's about the ecosystem itself.
When we published our first analysis in December 2025, we tracked 97 sources. One month later, 100+. Today: 310 unique sources are generating AI shopping coverage that meets our curation criteria. That's a 220% increase in less than three months.
| Source | Articles | Focus |
|---|---|---|
| amazon.science | 196 | Amazon AI research papers |
| PYMNTS | 183 | Payments & commerce |
| Bloomberg | 60 | Business & financial impact |
| Retail TouchPoints | 54 | Retail technology |
| Digital Commerce 360 | 51 | E-commerce data |
| + 305 more sources | 972 | Across every vertical |
The source explosion tells us something important: AI shopping is no longer a "tech story." It's a payments story (Finextra: 33 articles). A fashion story (Business of Fashion: 11). A grocery story (Supermarket News: 10). A logistics story (FreightWaves, Supply Chain Brain). When 310 publications from every retail vertical are covering AI shopping, the category has gone mainstream.
⚠️ The Casualties: Retail Darwinism Accelerates
While Big Tech invests hundreds of billions, the retail landscape is splitting into AI-haves and AI-have-nots. Our database captures both sides:
The innovators: Amazon investing $200B, closing physical stores to double down on AI commerce. Google launching UCP with 20+ partners. Shopify co-developing the protocol and integrating with both ChatGPT and Microsoft Copilot. Albertsons creating an end-to-end digital party planning hub. Swan Beauty launching AI smart mirrors.
The casualties: Eddie Bauer filing bankruptcy (February 9). UK fashion retailer Quiz entering administration (February 5). Amazon simultaneously shutting Amazon Go and Amazon Fresh stores while pivoting entirely to AI-driven commerce.
The pattern in our data is unmistakable: retailers that are building with AI are investing at historic levels. Retailers that aren't are facing increasing pressure. The gap between the two groups is widening.
But AI alone isn’t armor. Consider the paradox of Shein and Temu. Both companies use machine learning extensively for trend prediction, supply chain optimization, dynamic pricing, and hyper-personalized recommendations. By any operational measure, they are AI-native. Yet as we documented in our 2026 Insights analysis, Shein and Temu generated zero AI innovation coverage across our database despite a combined $50 billion in revenue. They use AI, but they are absent from the AI conversation. That distinction matters now. After the U.S. ended the de minimis tariff exemption on Chinese imports, Temu’s U.S. daily active users dropped 52%, Shein’s fell 25%, and both slashed U.S. ad spending by 70-95%. Forever 21 blamed Shein and Temu for its own liquidation, then those same disruptors got disrupted by trade policy. The lesson: using AI operationally is not the same as participating in the AI commerce ecosystem. In 2026, retailers need both AI capability and a commerce model positioned to survive protocol shifts and regulatory changes.
🔮 Three Predictions from the Data
1. The protocol war will have a winner by year-end. Google's UCP has momentum with 20+ partners, but Amazon's 300 million Rufus users represent the largest agentic shopping install base on earth. Watch for three signals: Does Amazon join UCP? Do major retailers adopt UCP checkout on Google AI Mode? Does OpenAI create or adopt a competing protocol? The answers will determine whether the next decade of commerce runs on an open or closed standard.
2. Agentic commerce will reach $100 billion in influenced transactions in 2026. Amazon's Rufus alone generated $12 billion in incremental sales in 2025. With 300 million users growing, Google UCP enabling checkout across AI Mode and Gemini, and ChatGPT's 800 million weekly users now able to complete purchases, the combined influenced transaction volume will be massive. Forrester notes that consumer trust remains the key frontier, but the infrastructure is now in place.
3. The "agent fee" becomes the new platform tax. Shopify merchants already pay 4% on sales made through ChatGPT checkout. As Google, Amazon, and OpenAI all position AI agents as the front door to commerce, merchants will face a new cost: the agent commission. This will reshape retail economics as fundamentally as marketplace fees did in the 2010s. Marketplaces that block AI shopping agents, as some already are, will face a difficult choice between control and relevance.
About This Analysis
AI Shopper News curates and filters AI shopping and retail technology stories from 310 unique sources, updating every 4 hours. This analysis covers 1,516 articles through February 11, 2026. All statistics reference our searchable database and link to verifiable source queries. We don't editorialize the news we aggregate, but we do pay attention to what the data reveals. Search our database →