In February 2026, OpenAI, Anthropic and Perplexity took opposite stances on advertising in conversational AI: ChatGPT now shows ads, while Claude and Perplexity refuse them. For users, the stakes go beyond the business model, because a single answer phrased like advice does not survive commercial suspicion well. At Fast Growth Advisors, we follow this choice closely, because it affects the credibility of the content B2B marketing teams produce with these tools.
What happened between OpenAI, Anthropic, Perplexity and Google in February 2026?
OpenAI chose advertising, Anthropic and Perplexity ruled it out, Google hedges.
On February 9, 2026, OpenAI switched on the first ads inside ChatGPT. Sponsored placements, displayed below responses, targeted on conversation topic, past chats and previous ad interactions. Covered by the test: free users and Go subscribers ($8 a month) in the United States. Plus, Pro, Business, Enterprise and Education plans are excluded.
Five days earlier, Anthropic had gone the other way.
In a post titled Claude is a space to think, the company committed to never displaying ads inside Claude: no sponsored links next to conversations, no advertiser influence on responses, product placements included. Its argument fits in one sentence. Users should never have to wonder whether the AI is genuinely helping them or subtly steering the conversation toward something monetisable.
Then, on February 18, Perplexity joined the no-ads camp. The search startup had tested sponsored answers in 2024 before reversing course. One of its executives told the Financial Times, with unusual candour, that a user must believe they are getting the best possible answer to keep using the product and pay for it.
Which leaves Google, playing both sides.
Google already places ads inside AI Overviews, the AI-generated summaries shown above traditional search results. In February 2026, it announced it was testing a new ad format in AI Mode, its conversational search experience. Gemini, its chatbot, remains ad-free. For now.
And “for now” is doing a lot of work in that sentence. Adweek reported on December 8, 2025 that Google had told advertisers about a Gemini ad rollout planned for 2026, and Google’s VP of Global Ads publicly denied it the same day. At Fast Growth Advisors, we read that split (advertiser briefings on one side, official denials on the other) as a sign of internal tensions.
Why do ads inside a conversational AI raise a different problem?
Because the tool delivers a single answer, in the tone of an adviser, where a search engine offers a list to sort through.
We have lived with online advertising for 25 years. We know how to spot a sponsored link on Google, ignore a banner on a news site, scroll past a promoted story on Instagram. An implicit but clear contract: the service is free, ads fund access.
With a conversational AI, the mechanics change.
This kind of assistant answers in natural language, in a dialogue, instead of displaying results. The user does not browse a list and pick the relevant items: they receive one answer, phrased in the register of a trusted adviser. And it is precisely this form, the natural conversation, that blurs the line between a sincere recommendation and a commercial suggestion.
OpenAI states that ads do not influence ChatGPT’s responses, which is probably true at the testing stage. Doubt, though, is a matter of perception.
If a user asks which project management tool to adopt and a Monday.com ad appears below the answer, that answer is tainted by doubt, even if it is objectively neutral. That is the point Fast Growth Advisors watches: perception, more than mechanics.
Will user trust in AI depend on the business model?
Nobody knows yet, and the two camps are betting on opposite answers.
Anthropic bets that trust is a durable competitive advantage. Its reasoning: in a market where several AIs are technically comparable, the absence of a conflict of interest becomes a selection criterion. A marketing director using Claude to evaluate vendors or structure a launch plan does not want to wonder whether the answer serves an advertiser.
OpenAI, for its part, bets on adaptation: users got used to advertising in Google Search, on YouTube and in podcasts.
The U-turn is striking. In October 2024, Sam Altman called the idea of mixing ads and AI “uniquely unsettling”, according to remarks reported by Tubefilter. Some eighteen months later, with about $1.4 trillion in capital expenditure commitments over eight years, mostly in data centres, OpenAI has to diversify its revenue.
At Fast Growth Advisors, we do not pick a side: both arguments are defensible. Yet they rest on opposite assumptions about how users will behave with a product that did not exist three years ago.
What scenarios are emerging for ads in AI agents?
Three scenarios coexist, and none of them is implausible.
The industry is at an inflection point comparable to the early web of the 2000s. Back then, people wondered whether users would accept search engines blending organic and paid results. They did, at the cost of a generalised cynicism towards search results that, twenty years later, opened the door to alternatives like Perplexity.
First scenario, habit: users get used to it, as they did with sponsored search, and ads in AI become the norm.
Another possible outcome is lasting segmentation. On one side, ad-funded AI, with broad access and lower trust. On the other, subscription AI, with restricted access and higher trust.
Last hypothesis, rupture: a first scandal over real or perceived ad bias in an AI response, followed by a massive backlash that rewrites the market’s rules.
Which one will win? At Fast Growth Advisors, we do not have the answer. Neither do OpenAI, Anthropic or the analysts.
What does this mean for B2B marketing teams?
For teams that use these tools every day (writing, research, competitive analysis, content strategy), the business model of their chosen AI is not a technical footnote. It is a reliability parameter.
Picture a marketing director who asks ChatGPT to compare 3 marketing automation platforms and sees a Salesforce ad below the answer. That analysis cannot go as-is to a board meeting. Not that the answer is biased: the appearance of bias is enough to disqualify it.
That is the whole nuance.
Whether ads actually influence responses matters less than their effect on the perceived reliability of those responses. And nobody can measure that effect with certainty in 2026, neither the vendors nor the analysts.
Only users will decide, over time, through their usage, their subscriptions and their tolerance for doubt.
At Fast Growth Advisors, we are watching this closely, without taking sides. What interests us is how these infrastructure choices reshape the tools marketing teams rely on, and what they mean for the credibility of content produced with their help.
Sources
Methodology: Fast Growth Advisors opened and reread every link below, checking that it predates this article and that it says what the text attributes to it. Thirteen sources in all.
- OpenAI: Testing ads in ChatGPT (Feb. 9, 2026)
- OpenAI: Our approach to advertising and expanding access to ChatGPT (Jan. 16, 2026)
- Anthropic: Claude is a space to think (Feb. 4, 2026)
- Mac4Ever: Perplexity renonce à la publicité pour préserver la confiance des utilisateurs (Feb. 18, 2026)
- TechCrunch: ChatGPT rolls out ads (Feb. 9, 2026)
- The Register: Anthropic keeps Claude ad-free (Feb. 4, 2026)
- Axios: Anthropic pledges ad-free Claude, with a Super Bowl shot at ChatGPT (Feb. 4, 2026)
- Sherwood News: Anthropic pledges no ads for Claude (Feb. 4, 2026)
- eWeek: Anthropic Says Claude Will Stay Ad-Free (Feb. 5, 2026)
- Tubefilter: OpenAI’s CEO said ads were a “last resort” business model (Jan. 26, 2026)
- Adweek: Google Tells Advertisers It’ll Bring Ads to Gemini in 2026 (Dec. 8, 2025)
- 9to5Google: Google says there are ‘no plans’ to put ads in the Gemini app (Dec. 8, 2025)
- Google Blog: What to expect in digital advertising and commerce in 2026 (Feb. 11, 2026)
FAQ
Do ads in ChatGPT influence the responses?
No, according to OpenAI’s announcement of February 9, 2026.
ChatGPT’s responses are generated independently from the ads, which appear below them and are labelled “sponsored”. Free and Go accounts in the United States are the only ones to see them. The real risk lies elsewhere, in the perception of bias.
Which AI chatbots are ad-free in 2026?
Claude and Perplexity, which both committed to it in February 2026.
Anthropic pledged that Claude would remain ad-free, and Perplexity dropped the sponsored answers it tested in 2024. Gemini shows no ads at this stage, but Google already places them in AI Overviews and AI Mode.
Why did Perplexity abandon advertising?
To protect trust, the condition for paid subscriptions.
After testing sponsored answers in 2024, Perplexity concluded that any doubt about the independence of its answers would undermine user trust, and with it users’ willingness to pay for the service. For a subscription business, that trade was not worth making.
Are ads in a conversational AI comparable to search engine ads?
Only on the surface: the mechanics differ.
A search engine presents a list of results that users filter themselves. A conversational AI delivers a single answer in natural language, which makes the line between genuine advice and a commercial suggestion harder to perceive.
What are the risks for B2B marketing teams?
A risk of perception more than of actual bias.
At Fast Growth Advisors, we see it this way: an analysis produced by an ad-funded AI may look less reliable to a board, even when it is objectively neutral. The choice of tool becomes a credibility parameter.
How much does an ad-free ChatGPT subscription cost?
From $20 a month, the price of ChatGPT Plus.
In the test launched in the United States on February 9, 2026, the Plus, Pro, Business, Enterprise and Education plans are exempt from ads. Free accounts and the Go plan, billed at $8 a month, are the ones that display them.
Will advertising in conversational AI become the norm?
Nobody can say in 2026.
Infrastructure costs push toward ad monetisation, but the conversational nature of these tools creates an unprecedented trust risk. User behaviour, over the long run, will settle which model dominates, and B2B teams would do well to watch it before choosing their tools.