Just three years ago, OpenAI was a niche research lab known primarily in Silicon Valley. That all changed after the launch of ChatGPT in November 2022 – the company has become a symbol of the global AI race.
OpenAI is investing billions to build a new internet infrastructure
Today, OpenAI strives to transform itself from just a chatbot developer, but into a next-generation platform that will unite tools for communication, work, learning, and entertainment. At its Dev Day 2025 developer conference on Monday, the company unveiled its strategy for creating an “AI internet ecosystem” and announced massive investments in physical infrastructure.
Billions in Computing Power
To support the growing demands of its models, OpenAI is investing billions of dollars in data centers. New agreements include the installation of systems with a combined capacity of 6 gigawatts based on AMD processors – in addition to existing contracts with Nvidia and Oracle.
In January, the company launched the Stargate project with Oracle and SoftBank, an investment of up to $500 billion in data center construction in the United States. The first facility, covering over 90,000 square meters, is already under construction in Abilene, Texas, with additional sites planned in New Mexico and the Midwest.
In July, OpenAI signed another agreement with Oracle for $300 billion over five years to expand Stargate’s infrastructure.
A separate deal with Nvidia provides for the purchase of chips to provide 10 gigawatts of computing power, in exchange for a $100 billion investment from the manufacturer.
“We need as much computing power as possible,” OpenAI President Greg Brockman emphasized in an interview with CNBC.
Besides the United States, the company is building data centers abroad, in the United Kingdom and the United Arab Emirates.
From Chatbot to Digital Platform
While Google, Amazon, and Meta once defined the architecture of the modern internet through search, shopping, and social media, OpenAI wants to become the core of the AI era.
According to Sam Altman, ChatGPT has 800 million weekly active users. The company is expanding its product capabilities: now you can create Spotify playlists, search for housing on Zillow, or complete purchases using the Instant Checkout feature directly from the chat.
In education, a learning mode has been launched that adapts responses to students, and the new Sora 2 app offers a feed of short AI videos—a direct response to Meta and TikTok.
Furthermore, OpenAI is collaborating with former Apple designer Jony Ive to create an AI device capable of competing with smartphones.
Competition and Risks
OpenAI’s expansion is taking place against a backdrop of fierce competition: Meta, Microsoft, and Google are already integrating their AI systems into ecosystems like Gmail, Docs, and Teams.
According to The Information, despite a $500 billion valuation, OpenAI is still unprofitable, with losses of $7.8 billion in the first half of 2025 alone.
“Today, ChatGPT competes not with Copilot, but with the entire Microsoft ecosystem,” noted Columbia University professor Daniel Keum.
To stay in the race, OpenAI is forced to simultaneously create new products and expand the infrastructure that supports them. It’s a vicious cycle—but it’s precisely this strategy that once enabled Google and Amazon to become tech empires.
“AI revenue is growing faster than any product in history,” Greg Brockman said in an interview with Bloomberg. “The more powerful our computing resources, the more opportunities and profits they will generate.”
