OpenAI is positioning 2026 as a turning point focused on what it calls “practical adoption,” according to the company’s chief financial officer. In a blog post published on Sunday, CFO Sarah Friar said the company’s priority is to narrow the gap between what artificial intelligence is capable of today and how it is actually used in everyday life by individuals, businesses, and governments.

Friar wrote that the opportunity for real-world impact is both large and immediate. She highlighted areas such as healthcare, scientific research, and enterprise applications, where improvements in AI-driven intelligence can directly lead to better results and measurable outcomes. In her view, the next phase of growth is less about showcasing what AI can do in theory and more about embedding it into daily workflows and decision-making processes.

In the blog, Friar also outlined how OpenAI thinks about monetizing its products, including ChatGPT, while at the same time ensuring access to the massive computing power required to run them. She explained that the company’s revenue growth closely follows the expansion of its technical infrastructure. According to her, OpenAI’s available compute capacity increased from around 0.2 gigawatts in 2023 to roughly 1.9 gigawatts in 2025. Over the same period, the company’s annual revenue run rate rose from about $2 billion to more than $20 billion.

She described this expansion as unprecedented, noting that growth at this scale has rarely been seen before in the technology sector. Friar added that OpenAI believes customer adoption and revenue generation could have accelerated even faster if more computing resources had been available earlier. In her words, access to compute has become one of the main factors determining how quickly AI products can scale.

Friar’s comments come at a time when OpenAI and the broader technology industry are facing increasing scrutiny over the enormous investments required to build data centers and secure the energy and hardware needed to support advanced AI systems. Many investors are questioning whether these capital-intensive efforts will translate into sustainable returns, especially as the industry is still in the early stages of widespread commercialization.

One of the most closely watched partnerships is OpenAI’s agreement with Nvidia, announced in September. Under that arrangement, Nvidia said it would commit up to $100 billion to support OpenAI’s efforts to build and deploy at least 10 gigawatts of Nvidia-based systems. To put that figure into perspective, 10 gigawatts of power is roughly comparable to the annual electricity consumption of about eight million U.S. households, based on energy data analysis.

However, uncertainty remains. In November, Nvidia cautioned investors that there was no guarantee the agreement with OpenAI would move beyond the announcement phase and become a finalized contract. This uncertainty underscores the complexity and long-term planning required to secure large-scale computing resources.

Friar acknowledged these challenges in her blog, noting that world-class compute capacity must often be secured years in advance and that growth rarely follows a perfectly smooth trajectory. She emphasized that discipline and long-term planning are essential in managing such an infrastructure-heavy strategy.

She also pointed out that OpenAI’s approach to computing has evolved significantly. Three years ago, the company depended on a single compute provider. Today, it operates within a more diversified ecosystem of partners. This shift, according to Friar, allows OpenAI to plan, finance, and deploy capacity with greater confidence in a market where access to computing power largely determines who can grow and compete.

Looking ahead, Friar argued that OpenAI’s business model is designed to scale alongside its services. As AI becomes more deeply integrated into fields like scientific research, drug development, energy management, and financial modeling, she expects entirely new economic models to emerge. These changes, she suggested, will reshape how value is created and captured across multiple industries.

The blog post follows OpenAI’s recent announcement that it plans to begin testing advertising for some ChatGPT users in the United States. The move signals a new phase in the company’s monetization strategy as it prepares for a potential public listing later this year.

Friar stressed that any monetization efforts must align closely with the user experience. She wrote that revenue-generating features should feel like a natural part of the product rather than an intrusion. If a feature does not clearly add value for users, she said, it does not belong in the platform.