Amazon vs. Perplexity: Who Owns Your Shopping Data?
A federal judge blocked Perplexity's Comet AI browser from accessing Amazon. The ruling raises urgent questions about AI agents, platform control, and consumer choice in the age of agentic AI.
You ask your AI assistant to find the cheapest headphones on Amazon. It logs into your account, browses the listings, and buys them — without Amazon's permission. Who's in the wrong?
What Happened
A federal judge handed Amazon a preliminary win on Monday, temporarily blocking Perplexity AI's Comet browser from accessing Amazon's website. U.S. District Judge Maxine Chesney ruled that Amazon had provided "strong evidence" that Comet accessed its site "without authorization," and found Amazon likely to succeed when the full case goes to trial.
Amazon originally sued Perplexity in November, alleging the startup deliberately concealed its AI agents to keep scraping Amazon's site after being told to stop. Comet — Perplexity's AI-powered browser — lets users ask a chatbot to find products on Amazon and complete purchases on their behalf. That means Comet needs to enter password-protected customer account areas, not just browse public pages.
The court found that Amazon spent more than $5,000 and "numerous employee hours" developing tools to detect and block Comet's access — costs that Judge Chesney treated as concrete, essentially undisputed harm. Perplexity has one week to appeal before the injunction takes full effect. The company told CNBC it "will continue to fight for the right of internet users to choose whatever AI they want."
Amazon's Real Concerns: Ads and Data
Amazon's objections go well beyond a territorial "keep off my lawn" instinct. Two concrete business interests are at stake.
First, advertising revenue. Amazon's ad business generated roughly $56 billion in 2024 — its fastest-growing segment. Advertisers pay only for impressions from real human users. When AI agents browse and buy, they generate traffic that has to be identified, filtered, and excluded before advertisers are billed. Amazon's original complaint noted this requires "developing new detection mechanisms" and "system adaptations" — costs it didn't sign up for and can't pass on easily.
Second, customer data security. Comet doesn't just browse Amazon's public storefront. To complete purchases, it accesses private account areas protected by passwords. Amazon argued this creates genuine security risks — an AI system acting inside a customer's account, handling payment data, purchase history, and saved addresses, without Amazon having any visibility or control over what happens next.
Meanwhile, Amazon has been quietly building its own walled garden. It's already blocked dozens of external AI agents — including OpenAI's ChatGPT — from accessing its shopping site, while investing heavily in its homegrown assistant Rufus, which is embedded directly in the Amazon app and website.
The Two Sides, Side by Side
| Amazon | Perplexity | |
|---|---|---|
| Core claim | Unauthorized access = security risk + contract violation | Blocking AI = restricting consumer choice |
| Legal basis | Computer Fraud and Abuse Act (CFAA) | Open internet principles |
| Business interest | Protect ad revenue and own AI ecosystem | Capture AI browser market |
| Consumer framing | "Trusted shopping experience" | "Users' right to choose their AI" |
| Current status | Preliminary injunction granted | Preparing appeal |
Why This Ruling Matters Beyond Amazon
This isn't just a dispute between two companies. It's a preview of a much larger battle.
Every major e-commerce platform — Walmart, Shopify, eBay — faces the same question: what happens when AI agents start acting on behalf of users, entering platforms without explicit platform consent? The legal framework governing this is thin. The Computer Fraud and Abuse Act was written in 1986, long before AI agents existed. Courts are now being asked to apply 40-year-old law to behavior its authors couldn't have imagined.
For investors, the implications are significant. Perplexity was valued at $9 billion as of its last funding round. Comet is central to its strategy of becoming the AI layer through which users interact with the web. A sustained injunction doesn't just block one feature — it challenges the entire agentic browsing model. Conversely, Amazon's stock barely moved on the news, reflecting how minor this skirmish is relative to its scale. But the precedent it sets could determine whether Amazon — or any platform — can legally lock out third-party AI agents entirely.
For consumers, the stakes feel more personal. If platforms successfully block third-party AI agents, your choice of AI assistant may be dictated by which shopping site you use, not the other way around. Amazon gets Rufus. Walmart gets its own assistant. The open, AI-powered web starts to look a lot like the early days of closed app stores.
This content is AI-generated based on source articles. While we strive for accuracy, errors may occur. We recommend verifying with the original source.
Related Articles
Amazon's site crashed for six hours last week. Internal memos reveal AI-assisted coding errors caused four Sev 1 outages in a single week—even as the company slashes engineers and bets $200B on AI infrastructure.
America is preparing strict new AI guidelines amid tensions with Anthropic, signaling a major regulatory shift that could reshape the global AI landscape and investment strategies.
Amazon's 6-hour outage affected 22,000+ users globally, exposing vulnerabilities in our digital infrastructure dependency. What happens when convenience meets fragility?
The Pentagon's war against Anthropic reveals a deeper threat to America's AI dominance. Can the US win against China while destroying its own companies?
Thoughts
Share your thoughts on this article
Sign in to join the conversation