ChatGPT Ads: A Cheap Test for 2027, Not a Strategy for Years
ChatGPT Ads are cheap and almost uncontested for now, but the platform's audience has stalled and 95 % of it overlaps with Google. Search Engine Land advises testing without betting on it. What this means for business in Russia.
OpenAI has launched advertising in ChatGPT and in August 2026 expanded it to 31 European countries. As Search Engine Land writes, the inventory is cheap right now and competition is minimal, but the long-term prospects of the platform itself are murky. Google brings users to Gemini through Search, Android, YouTube and Workspace, while ChatGPT has to fight for every user separately. For a small business owner the conclusion is this: cheap contacts are worth buying now, but building a long-term strategy on this channel is risky.
The difference in approach is simple. Google does not ask permission to show its AI: it is already in the phone, the inbox, the browser. ChatGPT has to earn every user, one app download, one habit, one return visit at a time. That changes the economics of the channel and determines how much attention and budget it deserves.
The AI Platform War Looks Like the Browser War
The author, Rémi Kerhoas of the agency Eskimoz, draws a parallel with the history of browsers. Statcounter's market share data from 2008 to 2026 shows Internet Explorer collapsing, Firefox fading into a loyal but marginal niche, and Chrome climbing a slow, unstoppable staircase and never coming back down. Chrome launched in September 2008 and grew on distribution through Google's services and then through Android.
SensorTower's monthly active user data for major AI assistants paints a similar picture. ChatGPT rocketed up, then flattened hard from September 2025 onward. Gemini draws a clean, steady diagonal straight through the ChatGPT line. Claude shows solid traction but remains tiny compared to the two leaders.
Swap the logos and the graph is the same: ChatGPT sits in the Internet Explorer seat, the incumbent that defined its category and now has to fight for the next one; Claude lines up with Firefox, a well-loved product with a passionate but structurally limited base; Gemini takes the Chrome seat, not because it is the better product but because it is plugged into an ecosystem that never has to ask permission to show up.
ChatGPT's Audience Has Stalled and Google Already Has Almost All of It
Gemini inherits the same compounding advantage as Chrome, just further along the curve: it sits inside Search, Android, Workspace and YouTube and does not need to win users from zero. ChatGPT has no equivalent.
ChatGPT's audience, according to SensorTower, basically flatlined in mid-2025. Worse, according to SimilarWeb data cited by the publication, only 5 % of that audience does not overlap with Google's. In other words, ChatGPT has no exclusive user base, and it is not cannibalizing search behavior either.
For comparison the author takes a platform marketers have long treated as a secondary channel: TikTok has roughly two billion monthly users globally, twice ChatGPT's reach. Look at what share of your media budget currently goes to TikTok and you have a fairly honest benchmark for how much strategic weight ChatGPT Ads deserves.
OpenAI Is Betting on Enterprise, Not Advertising
In an October 2024 fireside chat at Harvard, Sam Altman called advertising a “last resort” business model and said that ads plus AI was “uniquely unsettling” to him. Fourteen months later OpenAI launched ads anyway. In the view of the Search Engine Land author, that reversal says the company needed the revenue, not that advertising became the center of its strategy.
Leadership moves reinforce that reading. When OpenAI lost COO Brad Lightcap, a finance person, he was replaced by CRO Denise Dresser, formerly CEO of Slack. She was then replaced by Dali Rajic, an enterprise security veteran. None of them came from ad sales or media; all are enterprise B2B people.
OpenAI's own 2030 revenue forecast, as reported by Bloomberg: $280 billion, of which “only” $100 billion from ChatGPT Ads. The remaining $180 billion comes from B2B products. ChatGPT Work, released last month, hooks agents into inboxes, Slack, Notion and Figma to handle multistep knowledge work, which, in the author's words, is a great way to sell more tokens. The Information reports that OpenAI is testing an outcome-based payment model for enterprise clients.
Advertising is the bridge; enterprise products and the API are the destination. When both a company's hiring and its products point away from advertising, do not expect the ad business to get its focus.
Why It Is Still Worth Testing
Losing the long-term platform war and being worth testing right now are not contradictory positions, and treating them as one is exactly the mistake to avoid, the publication writes.
Start with stability. According to a Bloomberg analysis, Microsoft booked $24 billion in revenue from OpenAI in its latest fiscal year, roughly 70 % of Microsoft's total AI business. In most contexts that level of customer concentration would be a red flag. Here it works the other way: Microsoft does not underwrite a partner at that scale without confidence in its staying power. For anyone hesitating because “the platform is too new”, that number settles whether ChatGPT Ads will exist in two years.
The more useful historical parallel is not Internet Explorer but early TikTok Ads. Plenty of advertisers stayed on the sidelines for years, often for reasons closer to politics than marketing, even while the signals, a real base, genuine incremental reach, clear momentum, were already visible. Those who tested early built a knowledge edge that took competitors years to close.
Timing matters too. The expansion to 31 European countries in August 2026, the largest in the platform's history, came just ahead of the fourth quarter, and that is not an accident. OpenAI gets performance data to defend 2027 budget conversations with advertisers, and early advertisers buy inventory and audience insight while CPMs and CPCs are still low, before competition and pricing catch up.
Competition really is low. In an SE Ranking study, one advertiser, BestMoney, accounted for 13.56 % of all ChatGPT ads. ChatGPT shows no more than one ad unit per user prompt, while Google stacks up to four ads in its search results. The author's advice: test while it is cheap and uncrowded.
What This Means for Business in Russia and the CIS
The publication deals with global data and the European expansion; it says nothing about the availability of ChatGPT Ads to Russian companies. In practice, a test makes sense for those who already have foreign ad accounts and customers outside Russia: exporters, services with an international audience, companies selling to the CIS and Europe. If you have no such access, it is not worth setting it up for a single experiment.
The article's logic, though, applies more widely than to one platform, and here is what can be taken from it.
- Measure a channel's weight by share, not by hope. The author's device is simple: whatever share of budget you give a secondary channel like TikTok is what a new platform with a comparable audience deserves. For new channels in Russia, AI assistants, new platforms, the rule is the same.
- Look at distribution, not the product. The winner is not whoever has the better AI but whoever is already in front of the user. In Russia that means the main money goes where your audience already lives, and new channels are a “tests” line in the report.
- Set exit conditions in advance. Before launching a test, write down which metrics and what timeframe will tell you the channel is not working, and compare it with control channels on cost per lead, not clicks. Link tagging and end-to-end analytics are mandatory, otherwise the result is indistinguishable from noise.
- Remember the audience overlap. If 95 % of ChatGPT users are already on Google, advertising there is a second and third contact with the same people, not new reach. Count it as an addition to search, not a replacement.
- Buy knowledge, not only impressions. The main value of an early test, in the author's view, is an understanding of the format and the audience that competitors then spend years catching up on. This applies to any new channel: the results of the first campaigns are an asset even if the campaigns themselves did not pay off.
Conclusion
ChatGPT Ads today are a tactical opportunity: cheap inventory, near-zero competition, one ad unit per prompt and a platform backed by Microsoft. But the audience is not growing, 95 % of it is already on Google, and OpenAI's hiring and products are steering the company toward enterprise.
Search Engine Land's advice: test while it is cheap, build up knowledge of the format and the audience, and do not confuse a tactical bet with a long-term platform strategy. Keep the main resources where distribution works for you for free: in search and on the platforms where users already live.