^{Photo: Mint}

From rapid scaling to real-world impact

OpenAI is preparing to make 2026 a turning point, shifting its focus from rapid technological expansion to what it calls “practical adoption,” according to a detailed strategy update from Chief Financial Officer Sarah Friar. In a blog post published Sunday, Friar said the next phase for the artificial intelligence company will center on closing the gap between what advanced AI systems can do and how widely they are actually used by businesses, governments, and individuals.

The emphasis marks a natural evolution for OpenAI after several years defined by extraordinary growth in infrastructure, users, and revenue. Friar described the coming year as a moment where AI moves decisively from experimentation into daily operations across industries.

Explosive growth in compute and revenue

Friar outlined how closely OpenAI’s financial performance has tracked the expansion of its technical infrastructure. The company’s available compute capacity increased nearly tenfold in just two years, rising from approximately 0.2 gigawatts in 2023 to about 1.9 gigawatts in 2025. Over the same period, OpenAI’s annual revenue run rate climbed from around $2 billion to more than $20 billion.

She characterized the parallel growth curves as unprecedented, noting that few technology companies have scaled both infrastructure and revenue at this pace. Friar added that customer adoption and monetization could have accelerated even faster if additional compute had been available earlier.

The message was clear: in the AI economy, access to large-scale, reliable compute is not just a technical requirement but a core business driver.

Why 2026 is about adoption, not just innovation

According to Friar, OpenAI now sees its biggest opportunity in helping customers turn advanced models into tangible outcomes. She pointed to sectors such as health care, scientific research, and large enterprises, where better intelligence can directly improve productivity, decision-making, and results.

Rather than focusing solely on more powerful models, the company plans to invest heavily in usability, deployment, and integration. The goal is to ensure that AI tools like ChatGPT are not just impressive demonstrations, but essential parts of everyday workflows.

The high cost of building the AI backbone

The blog post comes amid growing scrutiny of the enormous capital required to support the AI boom. Data centers, specialized chips, and energy supply have become major constraints for the entire industry. OpenAI has been at the center of this debate due to the scale of its ambitions.

One of the most notable efforts is its announced partnership with Nvidia, under which the chipmaker said it could commit up to $100 billion to support OpenAI’s deployment of at least 10 gigawatts of AI systems over time. To put that figure in perspective, 10 gigawatts roughly matches the annual electricity consumption of millions of U.S. households.

However, uncertainty remains. Nvidia later cautioned investors that there was no guarantee the agreement would move beyond an initial announcement into a finalized contract. Friar acknowledged this reality, emphasizing that securing world-class compute requires long-term commitments and careful financial discipline.

A more diversified infrastructure strategy

Friar also highlighted a major shift in how OpenAI sources its compute. Just three years ago, the company relied on a single provider. Today, it operates within a diversified ecosystem of partners, giving it greater flexibility and resilience as demand fluctuates.

This diversification, she said, allows OpenAI to plan and deploy capacity with greater confidence in an environment where access to compute increasingly determines which companies can scale and compete.

Monetization that fits the product

As OpenAI looks toward broader adoption, it is also refining how it generates revenue. Friar reiterated that monetization must feel natural and add value for users. That philosophy is shaping decisions such as the company’s recent announcement that it will begin testing advertising for some ChatGPT users in the United States.

The move is widely seen as part of OpenAI’s preparation for a potential public offering, possibly as soon as this year. Friar stressed that any new revenue streams must align with the user experience rather than disrupt it.

A broader economic vision for AI

Looking ahead, Friar argued that OpenAI’s business model will scale alongside the expanding role of AI in the global economy. As artificial intelligence becomes embedded in drug discovery, energy management, financial modeling, and scientific research, entirely new economic frameworks are likely to emerge.

For OpenAI, 2026 is shaping up not as a year of flashy breakthroughs, but as a decisive step toward making advanced AI a practical, revenue-generating utility used at scale across the world.

‍