
OpenAI Reports $38.5 Billion Loss in 2025, Citing Azure Costs
LLM, AI Agents & AI Infrastructure Specialist

LLM, AI Agents & AI Infrastructure Specialist
OpenAI reported a $38.5 billion loss in 2025, driven by $34 billion in operational costs, of which $17 billion (50%) were tied to Microsoft Azure. Despite $13 billion in revenue, its dependence on third-party infrastructure raises concerns about the sustainability of AI development.
In 2025, OpenAI reported a net loss of $38.5 billion, a dramatic increase from its $5 billion loss in 2024. The company achieved $13 billion in revenue, yet its operational costs surged to $34 billion, primarily driven by the immense compute power required for advanced AI models like GPT-4 and the anticipated GPT-5.
While OpenAI’s revenue growth reflects its increasing market presence, the escalating operational costs threaten its financial sustainability, particularly as the company prepares for a potential IPO in 2026.
OpenAI’s partnership with Microsoft has been instrumental in providing the computational resources necessary to power its AI models. However, this collaboration has also become a financial burden, with half of OpenAI’s operational costs in 2025 directly tied to Azure services.
The scale of OpenAI’s financial losses is raising critical questions about the viability of AI companies that depend heavily on third-party cloud services. The ripple effects are likely to impact various stakeholders in the industry:
OpenAI’s ability to navigate these financial challenges will likely determine its future in the AI landscape. Several strategies could help stabilize its position:
OpenAI's $38.5 billion loss was primarily due to $34 billion in operational costs, half of which ($17 billion) were spent on Microsoft Azure for cloud infrastructure.
OpenAI's dependency on Azure accounts for 50% of its operational costs, limiting cost control and exposing the company to risks like pricing changes from Microsoft.
OpenAI could diversify its cloud partnerships, invest in proprietary infrastructure, increase product prices, and optimize its operational efficiencies.
💡 Dica Pro: Vertical integration, as seen in companies like Google and Meta, significantly reduces dependency on costly third-party infrastructure. For companies like OpenAI, developing proprietary data centers could cut long-term costs despite high initial investment.