WASHINGTON — OpenAI announced on Friday that it has surpassed one billion active users, marking a symbolic milestone reached less than four years after the launch of its widely used chatbot, ChatGPT.

"Our models now reach more than one billion active users and more than two million businesses. As people gain confidence in the technology, they use it more deeply," the company said in a blog post.

According to multiple US media outlets, the San Francisco-based startup had originally anticipated reaching the milestone earlier this year. OpenAI reported 900 million weekly active users in late February but subsequently encountered intensified competition from rivals.

Founded in 2015, OpenAI did not release its first consumer product, ChatGPT, until November 2022. For context, Facebook took approximately six years after launching its social network to reach the same one-billion-user benchmark — a comparison that underscores how rapidly generative AI has been adopted relative to previous consumer technology cycles.

The figure announced on Friday encompasses all of OpenAI's artificial intelligence interfaces. These include ChatGPT as the primary platform, along with Codex, a model designed for AI-assisted programming, and ChatGPT Work, a platform enabling users to access external applications. OpenAI's reach is further amplified through its partnership with Microsoft, which has invested billions of dollars in the company and integrates OpenAI's models across its Copilot products and Azure cloud services, giving the startup distribution channels that few competitors can match.

OpenAI's business customer segment, which produces higher revenue and margins per user, became a strategic priority for the company in late 2025. Since that shift, OpenAI has narrowed the gap with its primary competitor Anthropic, which has traditionally focused on serving enterprise clients. Other major players, including Google with its Gemini model family and Meta with its open-source Llama series, are simultaneously racing to capture overlapping enterprise and consumer demand.

The rise of AI agents — systems capable of performing tasks on demand, evaluating outcomes, and making autonomous improvements — has driven a sharp surge in demand for computing power, pushing costs higher for businesses and developers alike. This compute-intensive shift has placed pressure on data center capacity and semiconductor supply chains, with NVIDIA's GPUs remaining a critical bottleneck for companies scaling AI services.

In an effort to manage these expenses, many users are increasingly adopting less sophisticated models than the cutting-edge systems offered by OpenAI and Anthropic. This includes a growing turn toward open-source models, particularly those developed in China.

In early July, OpenAI released three versions of its latest model, GPT-5.6. Among them was Luna, a "low-cost" variant priced at one-fifth the cost of Sol, the highest-performing version in the lineup.

The competitive landscape is shifting so rapidly that on Thursday, OpenAI further reduced Luna's price by 80 percent, positioning it among the most affordable advanced AI models on the market, including those developed by Chinese firms. The move signals that pricing power in the AI model market is eroding even as aggregate user numbers grow, a dynamic that could reshape how developers and businesses select AI providers going forward.