ChatGPT Turned Ads On, Perplexity Turned Them Off: Buy AI Visibility or Earn It?

Key takeaway (as of 2026-05-28): In 2026 the two answer engines moved in opposite directions. OpenAI brought ads to ChatGPT; Perplexity removed ads entirely and kept only organic citations. As a result, the paths for a brand to appear in AI answers split clearly into two — paid ad slots you buy and citations you earn with content. The point is this: ads opening does not end the citation race. The citation inside the answer body is a separate layer, carries more trust, and is therefore worth more, not less.

What happened

In 2026, ChatGPT and Perplexity took opposite directions on monetization — one switched ads on, the other switched them off.

OpenAI rolled ChatGPT ads out in stages. Per OpenAI's official note, ads appear in a clearly labeled box at the bottom of a response and do not influence the actual answer content. They are matched to the conversation's topic and context. According to Axios reporting on 2026-05-05, a self-serve ads manager beta opened the same day so anyone can buy directly, CPC bidding was added on top of the existing CPM model, and the minimum-spend requirement was dropped. Winbuzzer reported on 2026-05-27 that OpenAI is widening this channel even to smaller advertisers.

Perplexity went the other way. According to Search Engine Land, Perplexity stopped testing ads in February 2026 and shifted its weight toward subscriptions. Campaign US reported the reason as trust concerns, quoting one executive: "A user needs to believe this is the best possible answer." The logic is that the moment ads appear, users begin to second-guess the integrity of the answer.

Two paths into AI answers, now clearly split

A brand's path into AI answers now divides cleanly into two: paid ad slots you buy, and citations you earn with content.

Item	ChatGPT	Perplexity
Ad slots	Yes (labeled box at bottom)	No (removed Feb 2026)
Citations (organic sources)	Yes (sources in the answer body)	Yes (the only path in)
How a brand appears	Buy ads + earn citations	Earn citations only
Trust framing of exposure	Ads separated as "labeled"	Answer = promise of the best answer

The core difference is simple. On ChatGPT you can buy part of your exposure, but on Perplexity there is no exposure to buy. And even on ChatGPT, what ads buy is the bottom box — the citation inside the answer body still has to be earned with content. In other words, the asset both engines require in common is not an ad budget but a citation.

Ads and citations are entirely different exposure mechanisms. Compared item by item:

Criterion	Ads (Paid)	Citation (Earned)
Cost structure	Charged continuously while running (CPC/CPM)	No extra charge after content is made
Persistence	Disappears the moment budget stops	Persists once it accrues
Placement	Labeled box at the bottom of a response	Evidence source inside the answer body
User trust	Perceived separately as "Sponsored"	Perceived as the basis of the answer
Reach	Only on engines that have ads	Common across ChatGPT, Gemini, Perplexity
Control	Direct, via bidding and targeting	Indirect, via content structure and trust

Why citations matter more even after ads arrive

Ads opening does not end the citation race; the in-answer citation is a different layer that carries more trust, so its value rises.

First, ads and citations are different regions of the same screen. OpenAI itself stated that ads do not influence the answer content. Put another way, a brand being cited in the answer body is exactly the region ads cannot buy. Even if you run ads, the body citation has to be earned separately.

Second, the weight of trust differs. Users trust a source cited as evidence in the answer body more than a box labeled "Sponsored." The very rationale Perplexity gave for abandoning ads was "trust wobbles once ads are visible." Citations bring the trust that ads cannot buy.

Third, citations travel everywhere. On an engine with no ads at all, like Perplexity, citation is the only entrance. On ChatGPT, which has ads, citation is still the path to body exposure. A single citation asset can work both engines at once, whereas an ad works only inside that engine, and only while you pay.

This pattern was already visible in search. According to Brandlight analysis cited by HubSpot, the overlap between AI-cited sources and traditional top search links fell from about 70% to below 20% (single-source estimate). Just as a high search rank no longer guarantees an AI citation, running ads does not guarantee a citation in the answer body.

The numbers behind the ad-versus-citation shift

The weight of money is tilting toward ads, while the weight of conversion and trust is tilting toward citations — that is where 2026's public figures point.

Ad revenue is growing explosively. Per Axios reporting on 2026-04-09, OpenAI's internal projection is to grow ad revenue from about $2.5 billion in 2026 to roughly $100 billion by 2030. Ads have become an enormous market. Yet the quality of conversion shows up more sharply on the citation side. The figures below come from different sources and are industry estimates, so treat them as directional rather than absolute.

Metric	Figure	Source
OpenAI ad-revenue target (2026)	~$2.5 billion	Axios (2026-04-09)
OpenAI ad-revenue target (2030)	~$100 billion	Axios (2026-04-09)
ChatGPT chatbot market share (Jan 2026)	~68% (down from ~87% a year earlier)	Similarweb-based reporting (single-source)
AI-referral conversion vs. organic search	~14.2% vs. 2.8%	Industry analysis (single-source)
Overlap of top search links and AI-cited sources	~70% → below 20%	Brandlight analysis (single-source)

The table's message is clear. Ads are a fast-growing revenue source, but the power to move users to action and the trusted seat inside the answer body belong to citations. This dovetails with analyses like Similarweb's 2026 generative-AI statistics, which note that AI-referred users arrive with high intent, having already finished comparing and evaluating.

What brands should do now

The strategy that works on both engines is not raising your ad budget, but first measuring which answers cite your brand right now.

Ads are a supplement you can switch on and off. Citations, by contrast, are a base asset that accrues only as content and trust build up — and once they accrue, they work on both engines at once, with no spend. That is why order matters.

1. Measure first — Ask the same question and check whether ChatGPT, Gemini, and Perplexity cite your brand in their answers, and how often you appear versus competitors. RanketAI focuses on measuring this brand citation inside AI answers and mention share versus competitors.
2. Build a structure worth citing — Clear paragraphs that answer the question directly, freshness, author information, external evidence, crawler access. Shape content into a form AI finds easy to pick as a source.
3. Ads come next — Running ads while your citation base is weak produces lopsided exposure: visible in the bottom box but absent from the answer body. Lay the base first, then add ads as a supplement.

The ad era opening does not change the job to "buy ads or not." The real question is: "Before I buy an ad, am I being cited in answers right now?"

Frequently asked questions (FAQ)

Related reading

• From the Age of Links to the Age of Answers: ChatGPT Search vs. AI Mode vs. Perplexity
• Why SEO Alone Is Not Enough in the Age of AI Search
• The New Standard for AI Shopping: How Google UCP and OpenAI ACP Reshape the Buying Journey

Update basis

• Content reference date: 2026-05-28 (KST)
• Update cadence: Monthly
• Next scheduled review: 2026-06-28