ChatGPT Ads Passed $1B Run Rate — But It Is Not in China. What Overseas Brands Should Do

On September 1, 2026, OpenAI said ChatGPT Ads passed a $1 billion annualized run rate less than 200 days after launch, and opened self-serve ad buying to India, Europe, the Middle East, and North Africa. The announcement did not include mainland China — because ChatGPT has no official operation there. That gap is the most useful piece of China marketing news an overseas brand will see this quarter, because it defines the one strategy that works today: you cannot pay to appear inside China's AI answers, so you need to earn the citation.

💰 The $1 billion run rate in under 200 days

OpenAI's numbers show conversational AI advertising is a working business model, not an experiment:

• ChatGPT Ads passed a $1 billion annualized run rate in under 200 days.
• Advertisers number in the tens of thousands across more than 40 countries.
• Ad buying now supports self-serve CPC and performance bidding, product catalog ads, and audience targeting.
• One e-commerce advertiser reported a 3x return on ad spend over 28 days; one partner said over 80% of ad-driven traffic came from new customers.
• OpenAI stresses that ads and answers are strictly separated and ads never influence answer content.

The message to the market is simple: ads inside an AI chat interface can work at scale. The question for brands is where this model exists — and where it does not.

🚧 What China looks like today: three gaps

For a brand targeting Chinese consumers, none of the pieces that make ChatGPT Ads work are available in mainland China:

1. No foreign AI ad platform. ChatGPT Ads does not cover mainland China, and ChatGPT has no official operation there. The same applies to Google's AI surfaces — blocked in China.
2. No local AI ad product yet. As of September 2026, the major Chinese assistants — Doubao, ERNIE Bot (Wen Xiaoyan), Kimi, Yuanbao, Qwen — do not run platform-wide ad products inside their chat answers. Chinese AI model makers still monetize mostly through API usage and subscriptions; Zhipu, for example, reported that API revenue makes up 86.5% of its total.
3. The biggest platform is deliberately not monetizing. Baidu has said it is "deliberately holding back" on monetizing its AI search to improve the product first.

So there is no ad slot, no auction, no budget line for "China's AI answers" in 2026. The commercial layer simply does not exist yet.

👥 The traffic is already there — but the inventory is not

The absence of ads does not mean an absence of users. Chinese consumers are already asking AI assistants for recommendations at scale. Monthly active users reported for the leading Chinese AI apps exceed 500 million across the category; Doubao alone reports hundreds of millions of MAU. When a user asks "which overseas brand is reliable for X" in Doubao, ERNIE, or Kimi, the answer is generated — and nothing in that answer is influenced by advertising spend, because no ad system exists.

That is the unusual situation this moment creates: a massive, growing recommendation surface with real purchase influence, and no way to buy placement on it. The only way to appear is to be referenced by the model's sources.

⏳ Why waiting for "China's version of ChatGPT Ads" is the wrong strategy

It is tempting to plan around a future ad product: wait until Baidu or Doubao launches chat ads, then buy visibility like you do on Google or Baidu search today. Three problems with that plan:

1. You will wait at least a year, likely more. Baidu is explicitly delaying AI-search monetization. Doubao and the others are still focused on usage and API revenue.
2. The brands that appear in AI answers first will keep the citations. Models are improving their source-selection logic, and cited sources tend to stay cited. First-mover citations compound.
3. When ads do arrive, the organic/AI-answer layer will already matter for quality score. On Baidu today, authority signals affect both organic ranking and ad quality. The AI answer layer will be the same — brands without citation history will pay more later.

Waiting converts a problem you can act on now into a more expensive problem later.

🎯 What to do now: claim your place before the ads exist

The practical play is GEO — getting your brand structured, sourced, and cited so Chinese AI assistants recommend it when the answer is generated:

1. Get your entity right. Make sure Baidu Baike, business registries, and Chinese industry directories describe your company consistently — name, category, what you sell.
2. Build a citation base in Chinese. Product details, specs, and comparisons published across authoritative Chinese sources give models something to cite instead of your own marketing claims.
3. Structure your own web content. Schema.org and JSON-LD markup on your pages help models extract facts about your brand reliably.
4. Watch the answers. Test how Doubao, ERNIE, Kimi, and Qwen describe your brand and your competitors, and close the gaps they reveal.

This is not SEO on a search page. It is preparing your brand to be the answer in a conversation where you cannot buy a word.

🧗 The obstacles — and why most overseas brands need help

There are three real barriers between an overseas brand and this work. First, you need a China presence and Chinese-language assets, which most non-Chinese companies do not have in-house. Second, Chinese AI assistants draw from Chinese-language sources — Baidu Baike, WeChat ecosystem content, local industry media — that are hard for an overseas team to navigate. Third, the work needs to be maintained as models and sources change.

This is exactly the gap BPP works in: we help international brands open Baidu accounts, run campaigns, and — increasingly — build the AI-answer presence that no paid slot can replace. The window is open now, while the answer layer is still young and before ad auctions raise the cost of visibility. When China's AI ads eventually launch, the brands that prepared their citations will pay less and rank higher. The others will be catching up.