OpenAI is considering deals where it would take a percentage of sales from products discovered or developed using its AI, according to CFO Sarah Friar. She raised the idea on The OpenAI Podcast this week, using drug discovery as her example.
"Let's say in drug discovery, if we licensed our technology, you have a breakthrough," Friar said on the Monday episode. "The drug takes off, and we get a licensed portion of all its sales. It's great alignment for us with our customer."
A business model for AI's next phase
The comments expand on language Friar used in a company blog post published January 19. There she wrote that "licensing, IP-based agreements, and outcome-based pricing will share in the value created" as AI moves into scientific research, drug discovery, energy systems, and financial modeling.
That blog post was mainly a victory lap: OpenAI's annualized revenue hit $20 billion in 2025, up from $6 billion a year earlier. Compute capacity tripled alongside it, from 0.6 gigawatts to 1.9 gigawatts. Friar called the parallel growth "never-before-seen at such scale."
But the numbers also explain why OpenAI is exploring new revenue streams. The company has committed roughly $1.4 trillion in infrastructure spending over the coming years, per previous reporting. Subscriptions and API fees alone won't cover that. Ads are coming, Friar confirmed, though OpenAI will maintain an ad-free tier.
What "outcome-based pricing" might look like
OpenAI hasn't announced specific licensing terms. Friar's podcast comments were speculative, a glimpse at how the company is thinking about monetization rather than a product announcement.
Still, the drug discovery example is telling. OpenAI already works with Moderna on personalized cancer vaccines and has invested in Chai Discovery, a biotech startup building AI for antibody design that raised $130 million in December at a $1.3 billion valuation. If outcome-based pricing becomes real, deals like these could involve revenue share on downstream sales.
The model would represent a significant shift from current enterprise terms. OpenAI's existing contracts give customers full ownership of inputs and outputs. Enterprise users get copyright indemnification. A royalty arrangement would layer new conditions on top of that framework, though presumably customers would negotiate those terms explicitly rather than having them buried in standard service agreements.
The timing
Friar's comments came during a media blitz at Davos. In a CNBC interview Wednesday, she said OpenAI's enterprise business now accounts for 40% of revenue, up from 30% a year ago, and should approach 50% by year end. One million businesses use the platform, she said.
The company completed its restructuring in October, converting to a for-profit public benefit corporation while keeping the nonprofit foundation in control. That structure is supposed to make fundraising easier as OpenAI pursues a potential IPO, with internal discussions targeting a filing in the second half of 2026.
Friar's "practical adoption" framing for 2026 suggests the focus is less on model improvements and more on extracting value from capabilities that already exist. The licensing trial balloon fits that strategy. So do the ads.
Whether pharma companies or other enterprise customers will accept royalty arrangements is another question. Drug development involves years of R&D, clinical trials, and regulatory approvals that have nothing to do with OpenAI's models. Attributing a percentage of final sales to AI assistance in the discovery phase would require some creative accounting.
OpenAI prepares to test ads in ChatGPT in the coming weeks for US users on the free tier.
