Future of ads in AI search platforms

Advertisers may view AI search platforms as a major new opportunity, but the space is still so unformed that few understand how ads will actually function within it. Some analysts even suggest placements on tools like ChatGPT could command prices closer to connected TV than to traditional search ads.

For now, it is still uncertain on how advertising will look like on these platforms, but brands are actively exploring where AI might fit into their marketing strategies.

To shed light on these uncertainties around AI search advertising, Ad Age consulted several industry experts, whose insights offer a glimpse into how this emerging ad model might take shape for brands.

How much will AI search ads cost?

So far, most platforms haven’t formally introduced ad products, and the few that have remain vague about pricing. Google, for instance, doesn’t let advertisers buy ads specifically for AI Mode or Overviews. placements run across both standard search and AI search, a spokesperson told Ad Age – making it hard to gauge the unique value of AI search inventory. Perplexity, meanwhile, has only run limited ad experiments, and it’s unclear if brands paid to be involved; the company declined to comment.

As more platforms begin rolling out ad options, costs will likely vary widely by provider. According to Hans Fischmann, CPMs could mirror those in connected TV advertising, which typically range from $8 to $60 – and AI search could even surpass that spread. The variance will hinge on perceived ad quality. With its massive scale and mainstream appeal, ChatGPT may command the highest prices, Fischmann noted. Perplexity, by contrast, has a smaller, more specialized audience, which could mean lower ad rates.

Will AI’s energy consumption lead to a higher cost of AI search ads?

AI firms are spending enormous sums to keep their models running, but advertisers probably won’t be asked to shoulder those energy costs, according to Lillie Ratliff, head of strategic initiatives at ad tech company Scope3. The assumption is understandable: large language models require vast amounts of power to train on data, forcing tech giants to build or adapt massive data centers worldwide. 

Google recently estimated that a single Gemini text prompt consumes about as much electricity as less than nine seconds of TV viewing. Scaled across billions of prompts, the energy use becomes staggering, as highlighted by MIT Technology Review. While the rising costs might encourage tech companies to look for ways to share the burden, potentially passing some of it onto advertisers through AI search ad pricing, Ratliff noted this is unlikely to meaningfully increase the cost of advertising.

AI companies have strong incentives to keep ad prices low in order to attract brands to this entirely new channel. Passing along higher costs tied to energy use would clash with advertisers’ budgets, sustainability goals, and overall willingness to participate. Instead, providers could build trust by being transparent about electricity consumption and reduction efforts while maintaining stable pricing.

Will ad waste be a big problem in the AI search market?

Ad specialists are hopeful that AI search will lead to significantly less ad waste compared to the current digital ad ecosystem, where 23% of total spending is lost, according to the ANA.

A key factor behind this expectation is the likely absence of a programmatic marketplace. As Hans Fischmann from AdRoll explained, when brands purchase ads directly from AI platforms, it reduces the risk of fraud and eliminates opportunities for low-quality, made-for-advertising websites to emerge, since “you have to come in through one pipe.”

AI search could also pave the way for new ad formats that avoid waste. According to Scope3’s Lillie Ratliff, these formats might take advantage of AI’s ability to recall previous user interactions, enabling it to serve ads that are more aligned with users’ interests