OpenAI Addresses Monetization, Hires Ad Engineers

What it means to be a software engineer will change as more people learn how to get computers to do what they want, but the nature of the job — typing or debugging code — will evolve, according to Sam Altman, CEO at OpenAI, who spoke at a Town Hall meeting streamed Monday on YouTube.

“It has happened many times in engineering" throughout the years, Altman said to a group of developers. "Demand for software isn’t slowing, but the way it’s customized will."

Altman is building a self-sustaining, full-stack advertising business -- from processing data its own centers that will run the ecosystem to reduce its reliance on ad serving, to designing its own custom chips, working with Broadcom to build them. 

Recent job listings for OpenAI shows the company is looking for experienced software engineers to help build an ad-delivery platforms and systems that determine when, where, and how monetized experiences appear across OpenAI products.

advertisement

advertisement

The engineers will architect and implement large-scale, high-performance delivery and optimization systems to remove latency, and improve on accurate and safe content, according to the job description. Ad content must be safe to protect brands and users.

Earlier this month, OpenAI announced plans to begin testing ads in the U.S. for free and low-cost subscription ChatGPT Go tiers. The ads will be labeled.

Ad-supported tiers will allow people to use its AI tools with fewer limits or without having to pay for the technology. Plus, Pro, Business, and Enterprise subscriptions will not include ads.

As the newcomer enters the ad space backed by a growing AI-driven chatbot and agentic agents, some search industry veterans question whether OpenAI's model to charge advertisers about $60 per 1,000 ad views, or impressions, is the correct choice.

Recent reports indicate OpenAI plans to prioritize impressions or views rather than clicks in its initial advertising push.

The Information had reported that OpenAI will use something similar to a CPM model to monetize the ads. The idea is to use an impression-based model (CPM/CPV) rather than a performance-based model like CPC.

Some industry executives do not agree with this initial strategy.

“It’s important for AI reasoning models to adopt CPC, rather than CPM, monetization for several reasons,” said Adam Epstein, AdMarketplace co CEO. “It’s a sustainable business model. It’s a business model that can subsidize the cost of compute for users and can keep access to the AI models open, and is used to do revenue shares with content producers.”

Search has historically relied on a pay-per-click model that focuses on payment only when a user interacts with the ad by clicking it, rather than when it is shown.

AI models have largely been trained on content funded by advertising, and that model needs to continue, Epstein said.

“We are seeing the importance of advertising and traffic models to the open web,” he said. “Finding a way to monetize AI reasoning models is important. OpenAI took a step in that direction, but in the end, it will be the CPC model that will bring the yields to fund this ecosystem.”

MediaPost asked Epstein the existential question of what happens if AI consumes the entire web to provide the answer, would there be any reason for the web to exist.

"Yes," Epstein said. "The web needs to continue to exist or AI consumes itself. Content must be generated and monetized. The fact that AI is consuming web content shows the importance of it." 

He said "it's not a sustainable model if it sucks up all the information in the world and not pay for it, because it will lead to no more content being produced."

Next story loading loading..