November 13th-November 19th // Estimated Reading Time: 9 minutes

In This Edition!

  • Claude may not do ads but it definitely does marketing

  • Amazon is planning world (product search) domination

Top Headlines🔥

Summary: Anthropic raised $81.3B on Wednesday, selling 980M shares at $83 each, making it the largest IPO ever. Then the stock opened on Thursday at $111.50, up 34% from its IPO price, which put its market value at about $2T. By the close, shares had slipped to $97.17, still up 17% from the IPO price, for a market value of about $1.8T. Only Apple, Microsoft, Nvidia, Alphabet, Amazon, and SpaceX are worth more among US companies. Through the first nine months of 2026, Anthropic's revenue reached $33.24B, up about 1,200% from a year earlier. It also reported a $104B net loss during the period, but most of that is an accounting charge tied to early funding deals, not cash spent.

Opinion: OpenAI, Google, and Meta are betting on ads as an AI revenue stream. Anthropic has sworn off ads (for now). But perhaps its $2T valuation will come from “doing the work” of marketing and ads. Claude keeps creeping into the marketers’ workflows and tools, from Salesforce’s Claudeforce announcement to last week’s launch of Meetings, Sheets, Data, and Whiteboards. If marketers can do everything in Claude, the rest of the martech stack will become a wasteland of UIs, and Claude could eat up a lot of that market. OpenAI, Google, and Meta want your media budget. Anthropic wants your martech budget and your headcount. 

One thing to monitor: Anthropic’s “no-ads” stance is easy to maintain while revenue grows 1,200% a year. Public investors will push back as soon as they see signs of revenue growth slowing down. That’s when sh*t will get real.

Summary: Amazon wants shoppers to do all their product research on its site instead of on Google, brand sites, or review sites. An internal plan called World Product Graph would gather data on almost every product for sale, including products Amazon doesn’t carry. Brands like Coca-Cola and Bayer have agreed to send data directly to Amazon. Better data would help Alexa for Shopping, Amazon’s AI chatbot, provide more accurate and thorough answers. Amazon called the plan a “brainstorming document” and said it “doesn't necessarily reflect our current plans.”

Opinion: “World Product Graph” would be Amazon’s response to external AI agents: a better shopping agent, built on better data. It helps Amazon justify keeping external AI agents out. Unfortunately for Amazon, personal AI agents are starting to manage your calendar, inbox, travel, and shopping, and having context across all of those areas will make them better at each individual job. Like shopping! Amazon will likely always be swimming upstream against these personal AI agents, and despite how awesome “World Product Graph” might be specifically for shopping, Amazon may still lose the agentic commerce battle. If Amazon eventually relents and lets AI agents in, brands that sell on Amazon won’t be as beholden to Amazon for visibility anymore.

New Products & Features 🚀

What It Does: ChatGPT Wallet lets dots (ChatGPT’s agents) buy things for users. Users add a card once. Before each purchase, they check the items and total, then approve it with Face ID or a fingerprint. The business gets a stand-in card number, so neither the agent nor the store sees the real card details. Card rewards and dispute protections still apply. OpenAI built ChatGPT Wallet with American Express, Mastercard, and Visa.

Quick Take: Every AI company wants a piece of agentic commerce, because commerce is where the money is. As stated above, Amazon is in their way. The more tools and features these companies launch to make agentic commerce easier, and the more adoption those tools get, the bigger the threat to Amazon. And if Amazon eventually relents because of overwhelming consumer pressure, and lets AI agents in, these platforms win in a big way, as do brands that get more visibility for their products across multiple agentic platforms and become less reliant on Amazon.


Google releases Gemini 4 Neon and Genie 4

What It Does: Gemini 4 Neon is Google’s new mid-tier AI model for complex work like coding and data analysis. It sits between Carbon, Google’s most powerful model, and Hydrogen, its fast everyday model. Google says it costs about a third less per task than Carbon. Genie 4, the AI model Google introduced in October for creating interactive 3D worlds, is also now available to Google AI Ultra subscribers. Users can change a world by typing prompts as they explore it, and the world comes with sound.

Quick Take: Could Genie 4 help marketers build ads and experiences people can walk around in? Could Genie 4 help build worlds with ads in them

What It Does: Knotch’s Ace for Agents rewrites a brand’s existing web content into a format AI search tools and agents can easily read and cite. It breaks pages into small chunks, tags each one by topic and audience, and updates them when the original page changes. Human visitors still see the same website. Knotch launched it alongside a Scott Brinker report predicting that traditional websites will disappear by 2030.

Quick Take: Agents don’t browse like humans do. They grab a few nuggets of information and move on. They need to be able to do that efficiently. Most brand websites are built for people who scroll and click, and those sites are getting fewer visits every month. Time to create the agent-friendly version of your website if you want to show up favorably in AI!

Check out the latest AI Edge podcast episode with Anda Gansca, CEO and Co-Founder of Knotch. We get into the value of website personalization, where product knowledge should live, the future of websites, and much more…

The Setup: Philips Sonicare runs Sponsored Ads on Amazon. After learning that 86% of shoppers in the oral care category weren’t even considering Philips, it wanted to reach those people on streaming TV to increase consideration without having to buy from each streaming service separately.

The AI Solution: Philips used Amazon Ads Agent to simplify its CTV buying. Philips combined its own customer data with Amazon’s shopping data to build audiences of people shopping for oral care who had never bought from Philips. Then it used Amazon Ads Agent to buy ads on Paramount+, Roku, Tubi, Peacock, TelevisaUnivision, and Warner Bros. Discovery. Running everything through one platform let Philips target the same audience on every service, control how often people saw its ads, and track purchases.

The Results:

  • 24% incremental reach among likely buyers

  • 75% lower cost per acquisition

  • 15x higher purchase rate among people who saw the streaming ads, compared with those who only saw Philips search ads on Amazon

  • More people searched for Philips by name after seeing the ads

Why This Matters: When you buy streaming services one by one, you can’t see the audience overlap, so you end up paying to reach the same people again and again. Consolidated TV offerings from companies like Amazon, Walmart, and Samsung will become more pervasive as CTV buyers demand deduplicated reach and frequency controls. 

Your Action: Paste your current CTV plan into AI: the streaming services, the budget, impressions for each, and who you’re targeting. Then prompt: “Act as a buying agent that plans streaming TV across all of these streaming services against one audience. Estimate how much these services’ audiences overlap, how many unique people this plan actually reaches, and how many times the average person sees my ad. Then reallocate the same budget to reach the most new people, with a cap of [X] views per person per week.” If the agent finds a lot of the same viewers getting hit on Roku, Tubi, and Peacock, you might be paying for the same users, and buying through one platform (Amazon, Walmart, Samsung, or your DSP) could help.

Other Notable Headlines📌


‘Do not search Amazon’: Meta’s Muse now shops elsewhere by default - Amazon blocked Muse in September. Meta now appears to be going along with it. Muse’s shopping instructions skip Amazon unless a user asks for it and point the agent to Walmart, Target, Best Buy, Home Depot, and Lowe’s instead.

Dentsu looks to AI to restore its “competitive edge” as it underperforms in Q3🔒 - After a 3.7% drop in organic revenue, Dentsu plans to grow its business that helps clients adopt AI, which has handled about 500 projects so far.

Claude’s Duet agent is a little neurotic and surprisingly helpful - A Verge journalist connected Anthropic’s new personal agent to her email, calendar, Google Docs, and Slack. It caught a deadline crunch she’d missed and drafted a fix, but it also nagged her about deadlines that weren’t urgent.

That’s It For This Week 👋

The U of Digital AI Edge Newsletter is intended for AI Edge Pro subscribers, but occasional forwarding is okay!

To subscribe visit Uof.Digital/Newsletters/ai-edge or AI Edge Pro, or contact us directly for group subscriptions.

And remember, U of Digital helps teams drive better outcomes through structured education on critical topics in marketing / advertising like AI, mar tech, ad tech, CTV, commerce media, and more. Interested in learning more about how we can supercharge your team? Let’s talk!

Thanks for reading!