OpenAI Targets “Practical Adoption” in 2026 as Revenue Surpasses $20 Billion

SAN FRANCISCO – In a comprehensive strategy update that signals a major shift from experimental research to industrial-scale utility, OpenAI’s Chief Financial Officer Sarah Friar announced that 2026 will be the company’s year of “practical adoption.”

The directive, outlined in a company blog post on January 18, 2026, marks a pivotal moment for the artificial intelligence leader. As the initial “wow factor” of generative models matures into a demand for measurable ROI, OpenAI is repositioning itself not just as a creator of frontier models, but as the foundational infrastructure for global productivity.

Closing the “Capability Gap”

According to Friar, the primary mission for the coming year is narrowing the divide between what AI is technically capable of and how it is actually integrated into the daily operations of businesses and governments.

“The priority is closing the gap between what AI now makes possible and how people, companies, and countries are using it day to day,” Friar wrote. She specifically highlighted healthcare, scientific research, and enterprise as the sectors where “better intelligence translates directly into better outcomes.”

The strategy moves beyond simple chat interfaces. Friar described an evolution toward AI agents—systems that maintain context over long periods, operate autonomously, and execute multi-step tasks across various software tools. For enterprises, this means transitioning from using AI for “drafting emails” to using it as a “persistent operational layer for knowledge work.”

Unprecedented Financial Growth

The shift in strategy is backed by staggering financial performance. Friar revealed that OpenAI’s annualized revenue run rate surpassed $20 billion in 2025, a massive jump from $6 billion in 2024 and just $2 billion in 2023.

Friar linked this financial trajectory directly to the availability of physical infrastructure. She noted that OpenAI’s computing capacity grew from 0.2 gigawatts (GW) in 2023 to 1.9 GW in 2025.

“This is never-before-seen growth at such scale,” Friar stated, adding that the company firmly believes more compute in earlier periods would have led to even faster monetization. This “compute-to-revenue” correlation serves as the justification for the company’s massive ongoing capital expenditures.

New Economic Models and Partnerships

As OpenAI enters this new phase, the way it makes money is also evolving. While consumer and team subscriptions remain a core pillar, Friar hinted at the emergence of “new economic models” as AI becomes embedded in specialized fields like drug discovery and financial modeling.

Potential future revenue streams mentioned include:

• Outcome-based pricing: Charging based on the value or success of an AI-generated result.

• IP-based agreements: Licensing specific intelligence for specialized industrial use.

• Advertising: OpenAI recently confirmed it is testing ads within ChatGPT to support a free tier that drives broader adoption.

The company is also leaning heavily into partnerships to scale its reach. A recent landmark agreement with ServiceNow aims to bring OpenAI’s frontier models, including the newly discussed GPT-5.2, directly into enterprise workflows. This allows businesses to use AI for end-to-end actions—such as resolving IT incidents or managing HR requests—within their own secure environments.

Navigating the Infrastructure Challenge

Despite the optimism, Friar acknowledged that “compute is the scarcest resource in AI.” Securing the energy and hardware required for 2026 and beyond requires “discipline and foresight.”

OpenAI has moved away from relying on a single compute provider, instead building a diversified ecosystem of partners. This includes a massive, though still complex, $100 billion commitment from Nvidia to support the deployment of up to 10 GW of systems.

The focus on “practical adoption” also serves as a response to growing investor scrutiny. As the “AI hype cycle” cools, the market is demanding proof that the billions of dollars poured into GPUs and data centers can generate sustainable profit. By focusing on sectors where AI can provide immediate, tangible value—like accelerating drug discovery or automating complex financial audits—OpenAI aims to prove its longevity.

Looking Ahead

Beyond software, OpenAI is reportedly “on track” to unveil its first hardware device in the second half of 2026. This, combined with the push for agentic workflows, suggests a future where OpenAI is as ubiquitous in the physical and professional world as the internet itself.

As Friar concluded in her missive, the goal is for AI to become a “foundational utility,” comparable to electricity or clean water—essential, reliable, and fundamentally practical.