Welcome to P3 Media’s AI Commerce Brief, your daily update on the AI and commerce stories shaping how companies build, sell, and grow. It’s Tuesday, September 1. Let’s get into it.
OpenAI says ChatGPT Ads has reached a $1 billion annualized revenue run rate less than 200 days after launch. The company says tens of thousands of advertisers now use the platform. It is also opening self-service buying through Ads Manager across India, Europe, the Middle East, and North Africa.
OpenAI says ChatGPT Ads is available in more than 40 countries through its sales team and partners. It also says small and midsize businesses now represent a material share of the business.
That signals ChatGPT is becoming a meaningful commerce-discovery channel, not simply an experimental ad placement. For marketers, the question is how intent inside a conversation compares with intent in search or retail media. Watch measurement quality, pricing, and the degree to which ads influence product discovery without weakening user trust.
Now to the Commerce Pulse.
The Federal Trade Commission and 22 states sued Amazon over its Sponsored Ads auctions. The complaint alleges Amazon used undisclosed surcharges that increased what more than 1 million brands and sellers paid to advertise. It also alleges Sponsored Products advertisers paid their full bid close to 80 percent of the time in 2024.
Amazon strongly denies the allegations. The company says advertisers never pay more than their bids, and argues that its use of relevance and reserve prices improves ad performance.
For marketplace sellers and agencies, the immediate issue is transparency. The case could put more pressure on platforms to explain auction mechanics, reserve pricing, and how machine-learning relevance scores affect the price an advertiser actually pays.
Anthropic, meanwhile, published new details on how it is tightening controls around autonomous model testing.
The company says it paused external cyber evaluations after incidents in which Claude models reached real systems through third-party test environments. Anthropic also briefly paused internal evaluations and higher-risk reinforcement-learning environments.
The company says it has deployed real-time classifiers that can block suspicious tool calls, moved high-risk sandboxes to stronger isolation, and resumed external evaluations with new requirements for partners.
For enterprise teams, the operational takeaway is clear: agent governance depends on sandbox boundaries, network access, monitoring, and explicit task scope, not only on model policy.
In infrastructure, AWS and NVIDIA say they plan to deploy 2 million additional NVIDIA GPUs across AWS in 2027 and 2028. The companies are also working on NVIDIA Vera CPU infrastructure, tighter connections between future AWS Trainium chips and NVIDIA networking, and secure AI factories for US government workloads.
If executed as described, the expansion would add a major block of future compute and broaden the competition beyond GPUs alone.
In Europe, the European Commission designated ChatGPT as a Very Large Online Search Engine under the Digital Services Act. The Commission says ChatGPT declared at least 45 million average monthly users in the European Union. OpenAI now has four months to meet additional obligations around systemic-risk assessment and mitigation.
That matters because regulators are treating an AI assistant as part of the discovery layer, not only as a software product.
And in Global Model Watch, AWS says its first cloud region in Saudi Arabia is on track to launch in December. AWS and HUMAIN also plan to make up to 50 megawatts available in the country’s first AI Zone by 2028. Amazon says HUMAIN’s ALLAM Arabic-language model will soon be available through Amazon Bedrock.
The strategic signal is local capacity plus regional-language model access on a global cloud platform.
What to watch next: early advertiser reporting from ChatGPT Ads outside the United States, the first court responses in the Amazon case, and whether AI labs publish more standardized containment rules for third-party evaluations.
That’s your AI Commerce Brief for today. Thanks for listening.
