Inside OpenAI’s Financial Landscape: A Closer Look at Revenue and Costs
Following a whirlwind year marked by high-profile deals, speculative IPO chatter, and an unprecedented surge in AI enthusiasm, OpenAI’s financial operations have come under intense scrutiny once again.
Revealing Revenue Streams Through a Recent Leak
A recent disclosure by tech commentator and Silicon Valley insider Ed Zitron has shed new light on the fiscal health of OpenAI, arguably the most scrutinized AI enterprise globally. The leak reveals that Microsoft received approximately $493.8 million in revenue-sharing payments from OpenAI during 2024.
By the first three quarters of 2025, this figure surged dramatically to $865.8 million. This substantial increase is reportedly tied to Microsoft’s 20% revenue share from OpenAI, stemming from their high-profile partnership often dubbed the “$13 billion AI dominance pact.” While neither company has officially confirmed the exact percentage, this figure is widely referenced in industry circles.
The Complex Revenue Sharing Between Microsoft and OpenAI
However, the financial relationship is more intricate than it appears. Microsoft also disburses payments back to OpenAI, including roughly 20% of revenue generated from Bing and Azure’s OpenAI Service, which provides access to OpenAI’s AI models. According to insiders, the leaked figures represent Microsoft’s net revenue share after these payments are deducted.
The gross revenue figures and the precise contribution from Bing remain undisclosed, as Microsoft maintains strict confidentiality over these details, akin to the secrecy surrounding the training of future AI models like GPT-6.
Estimating OpenAI’s Revenue and Growth Projections
Applying the commonly cited 20% revenue share model, OpenAI’s gross revenue likely exceeded $2.5 billion in 2024 and reached at least $4.33 billion in the first nine months of 2025, with actual numbers potentially higher. CEO Sam Altman has publicly projected an annual revenue run rate surpassing $20 billion, with ambitious forecasts aiming for $100 billion by 2027-claims that have sparked considerable debate and skepticism within the tech community.
Balancing Revenue Against Soaring Operational Expenses
Revenue, however, tells only part of the story. Zitron’s analysis highlights that OpenAI’s expenditure on inference compute-the costly process of running AI models-was approximately $3.8 billion in 2024 and escalated to an astonishing $8.65 billion in just the first three quarters of 2025.
While Microsoft subsidizes much of the training costs through credits, the inference expenses are paid in actual cash, representing a significant financial burden. If these estimates are accurate, OpenAI may currently be spending more on operating its AI models than it earns in revenue.
Implications for the AI Industry and Investment Climate
This financial dynamic intensifies ongoing discussions about the sustainability of the AI boom. If OpenAI-the sector’s leading entity-is operating at a loss, it raises critical questions about the viability of the broader AI investment landscape, where billions of dollars are being funneled into startups and emerging technologies.
As the AI gold rush continues, stakeholders must weigh the balance between rapid innovation and financial prudence, especially as operational costs soar alongside ambitious growth targets.
