OpenAI’s ambitions continue to expand beyond anything the tech industry has seen before. According to reports, the company plans to spend over US$1 trillion in the next decade to fuel the infrastructure needed for its AI systems, even as its annual revenue stands at about US$13 billion.
Roughly 70% of that revenue comes from ChatGPT subscriptions, a service that has gained over 800 million regular users worldwide. However, only about 5% of those users are paying subscribers, highlighting a massive gap between OpenAI’s user scale and its monetization potential.
The company’s spending spree underscores its desire to dominate AI infrastructure. Recent partnerships with Oracle, Nvidia, AMD, and Broadcom represent commitments of more than 26 gigawatts of computing capacity, an amount that dwarfs even the largest hyperscaler deployments to date.
Partnerships Powering the Expansion
A closer look at the contracts reveals just how vast OpenAI’s infrastructure ambitions are. The company has reportedly signed a $300 billion deal with Oracle over five years, beginning in 2027, to secure the cloud compute power necessary for training and running advanced models.
Additionally, CoreWeave, a specialized AI cloud provider, is said to have a $22.4 billion agreement, while Broadcom is expected to supply custom AI chips worth around $10 billion. These partnerships point to OpenAI’s efforts to diversify its supply chain and secure the specialized hardware needed to scale the next generation of AI systems, including its massive Stargate data center project.
Yet, this scale of spending raises immediate questions. OpenAI’s filings do not specify whether these contracts include take-or-pay terms, revenue triggers, or financing contingencies, details that could make or break such an enormous investment plan.
The Financing Gap and New Business Frontiers
To bridge the yawning gap between its $13 billion in revenue and its $1 trillion in spending plans, OpenAI is venturing into new business areas. These include government contracts, AI-powered shopping tools, video services, and even consumer hardware.
The company’s management reportedly projects $44 billion in cumulative losses before achieving profitability by 2029. Much of this will be driven by roughly $60 billion per year in Oracle cloud costs, underscoring just how front-loaded its infrastructure expansion will be.
Analysts suggest that OpenAI’s spending is less about short-term profitability and more about cornering the AI infrastructure market before competitors can catch up. If successful, the company could emerge as both a model provider and a backbone infrastructure supplier for AI itself.
Market Risks and Industry Implications
However, OpenAI’s massive commitments also carry systemic risks. Several major U.S. corporations depend heavily on OpenAI’s APIs and enterprise tools. Should OpenAI face financial strain, market stability could be affected, given the dependency of Fortune 500 firms on its systems for automation and productivity.
This concentration risk has already spurred interest in multi-model resilience tools, failover systems, and AI routing platforms that distribute workloads across different providers such as Anthropic (Claude) and Google (Gemini). Investors are watching closely, as startups offering such solutions may benefit from organizations seeking vendor diversification and cost optimization.
In this sense, OpenAI’s trillion-dollar plan is not just reshaping its own future, it’s altering the entire AI ecosystem. If the company can sustain this level of spending, it could cement its dominance for a decade. But if it falters, the ripple effects could redefine the balance of power in AI infrastructure worldwide.
