As OpenAI Imposes Ads On Free Users, Initial Buys "Can't Prove Measurable Results"

To absolutely no one's surprise, OpenAI is about to impose advertisements on free users of ChatGPT. The move is seen as a means of defraying the company's mind-boggling expenses while also paving the way for a more sophisticated ad strategy to come. 

There is massive interest in the potential for advertising on AI. The problem, so far, according to advertisers and agencies that have participated in tests, is that OpenAI is not providing any measurable data to prove the ads' impact, let alone that it will work at all (though that seems assured). The question is whether this is because OpenAI is trying to fine tune its service before rolling it out or, as has become endemic for the AI industry, they just don't want to share any data or information as their business and legal posture has been that everyone's data belongs to AI for free. That is unlikely to stimulate advertiser participation, so some sort of alternative is to be expected. JL

Steven Desaulniers and Julia Boorstin report in CNBC, Barry Schwartz reports in Search Engine Roundtable and Reuters reports:

OpenAI will show ads to all U.S. users of ChatGPT’s free ‌and Go versions. (But) the (financial) commitments required to participate in initial tests are unusually high for this type of experiment. (And) so far, the test is moving too slowly to meet the hype. OpenAI’s is frustrating many partners because of the conservative rollout. Executives at agencies working with early ChatGPT advertisers said they haven’t been able to prove the ads have driven any measurable business outcomes for their clients. "OpenAI hasn’t yet offered marketers any automated way to buy ad space. Buyers have had to rely on phone calls and sending spreadsheets and emails to OpenAI representatives. More importantly, advertisers found it hard to tell whether the ads were paying off." The opportunity remains massive: the number of ads halfway through March increased 600% from the first of the month.

OpenAI will show ads to all U.S. users of ChatGPT’s free ‌and Go versions, Reuters reported Saturday (March 21).

And as the company prepares for the launch, it has hired Dave Dugan, a former top advertising executive for Meta, to oversee its as sales, The Wall Street Journal reported Monday (March 23) 

When OpenAI first announced it was rolling out ads on ChatGPT, brands and agencies across Madison Avenue were eager to test the new format to figure out their artificial intelligence advertising strategies.

The high-profile announcement, which was far more public than a typical “alpha” test of a new format, presented a massive opportunity.

Three of the world’s largest ad agencies are part of the testing program, including WPP, Omnicom and Dentsu. 

So far, the test is moving too slowly to meet the hype, according to multiple ad industry sources who spoke on condition of anonymity in order to discuss the details.

The sources told CNBC that OpenAI’s test program is frustrating many of its partners because of the conservative rollout.

The ad commitments required to participate in the test were unusually high for this type of experiment.

Some brands dedicated between $200,000 and $250,000 to the test, which is double a typical experimental ad commitment. For some brands, this money came from funds dedicated to innovative new formats, while others drew from search or social ad budgets.

With the pilot program running through the end of March, some of the sources told CNBC they are concerned with the slow pace of rollout, which means their full budget commitments are unlikely to be spent by the end of the month. "Two executives at agencies working with early ChatGPT advertisers said they haven’t yet been able to prove the ads have driven any measurable business outcomes for their clients."

"OpenAI hasn’t yet offered marketers any automated way to buy ad space. Buyers have had to rely on making phone calls and sending spreadsheets and emails to OpenAI representatives, one ad executive said. More importantly, advertisers found it hard to tell whether the ads were paying off."

While any excess will be returned, the budget was already committed to the trial and so can’t be deployed elsewhere during the quarter. Advertisers also won’t get the volume of insights they were hoping for.

Omnicom did not return requests for comment. WPP declined to comment. 

OpenAI told CNBC that the slow rollout of the ads program was intentional.

“We’re in the early testing phase of ads in ChatGPT, and the goal right now is to learn and refine the experience for consumers before expanding it more broadly,” the company said. “We’re encouraged by early signals from users and participating brands, and continue to see strong interest from advertisers.”

Japanese ad giant Dentsu told CNBC it set realistic expectations for its clients going into the test, pulling from a pool of funds dedicated to testing and innovation.

Dentsu EVP and Head of Paid Search Meredith Spitz said it is early on, but the firm is “eager to partner with OpenAI to further test, learn and evolve the offering.”

“So far, ad delivery is quickly building momentum, with volume increasing week-over-week as the environment scales,” she said. 

Despite some early frustrations, sources said they have been encouraged by OpenAI’s response to feedback and how quickly the company has been able to make changes, and more recently, ramp up.

The sources told CNBC that the caution is a good sign of OpenAI’s commitment to building a sustainable and successful ad business. But the frustration stems from the enthusiasm for this new category, and an eagerness to put budgets into ChatGPT ads and get more insights into how they’re working.

According to recent data from research firm Sensor Tower, the number of ads served halfway through March increased about 600% from the first of the month.

Sensor Tower estimated that ads have now rolled out to about 5% of ChatGPT mobile users, up from 1% at the beginning of March.

The opportunity for OpenAI and the AI ads landscape remains massive. 

A recent Truist analyst note called 2026 an “inflection year” for large language model-powered ads.

“Within the next several years, we would expect LLM-powered ad channels to become one of the most important pillars of the digital ad industry alongside Search, Social, and Retail Media,” the analysts wrote.

Truist estimates OpenAI will generate under $1 billion in ad revenue this year, with that figure growing to over $30 billion by 2030.

Dentsu noted that the most value of these new ads can come from brands looking to reach ChatGPT users with very specific queries.

“Overall, we’re seeing the continued importance of aligning ad relevance with user intent, reinforcing a broader pattern in conversational discovery; that when user intent is precise, brands with focused offerings and tailored messaging are best positioned to deliver relevance and value in the moment,” Spitz said.

While ads embedded in AI search are seen as having massive potential by ad industry insiders and analysts alike, Anthropic is wary.

The AI giant last month took shots at OpenAI in a Super Bowl commercial, criticizing its move into ads and proclaiming that Anthropic’s own platform will remain ad-free. Perplexity recently removed ads from its platform after beginning testing in 2024.

Google, meanwhile, has not yet announced official plans for ads within Gemini, but the company has signaled in recent reports that it is not ruling them out.

The company already has plenty of ad inventory around the AI overview results that appear alongside Google search results. 

The question is whether OpenAI’s slow rollout will prove an advantage for industry leader Google, which will sell an estimated $252 billion in search ads this year, according to Truist.