OpenAI told prospective investors that its dependence on Microsoft for “a substantial portion of our financing and compute” represents a material business risk, according to a financial document viewed by CNBC. The disclosure, shared alongside OpenAI’s recent $110 billion funding round, reads like a preview of the risk factors section in an eventual IPO prospectus.

“If Microsoft modifies or terminates its commercial partnership with us, or if we are unable to successfully diversify our business partners, our business, prospects, operating results and financial condition could be adversely affected,” OpenAI wrote in the document.

An OpenAI spokesperson told CNBC the language is “a standard legal risk factor disclosure” and that “similar language has been in place for years.”

What the Document Reveals

The filing-style document includes sections titled “Risks Related to the Transaction” and “Risks Related to our Business.” Beyond the Microsoft relationship, OpenAI flagged several other concerns:

  • Compute costs: As of December, OpenAI had roughly $665 billion in estimated compute spend commitments through 2030. The company noted that “our compute requirements are dynamic and may expand significantly.”
  • Chip supply chain: OpenAI specifically named Taiwan Semiconductor Manufacturing Company as a potential single point of failure, citing the risk that regional conflict between China and Taiwan could cause “severe disruptions.”
  • Litigation: The document detailed three separate lawsuits from co-founder Elon Musk or his company xAI, plus at least 14 lawsuits from ChatGPT users or family members alleging the product contributed to mental illness, suicide, or death.

Microsoft has invested $13 billion in OpenAI since 2019 and held a 27% diluted stake in the for-profit entity valued at $135 billion as of October 2025. But the relationship has grown more complicated: Microsoft added OpenAI to its list of competitors in its 2024 annual report, and OpenAI has since expanded to other cloud providers including CoreWeave, Google, and Oracle.

Why Agent Builders Should Pay Attention

The disclosure matters beyond OpenAI’s investor relations. Any company running autonomous agents on GPT-4, o-series, or future OpenAI models inherits a layered concentration risk: their agents depend on OpenAI’s APIs, which depend on Microsoft’s Azure compute, which depends on TSMC’s chip fabrication.

OpenAI’s $665 billion compute commitment through 2030 signals the scale of infrastructure required to keep those APIs running. If Microsoft were to reprioritize Azure capacity, renegotiate terms, or simply fail to scale fast enough, the downstream effects would hit every agent workflow built on OpenAI’s stack.

The company reported $13.1 billion in 2025 revenue and has surpassed $25 billion in annualized revenue as of February 2026, according to Benzinga. ChatGPT now has 900 million weekly active users. That growth makes the infrastructure dependency more consequential, not less.

For agent builders evaluating their own supply chain risk, the document is a rare moment of candor from the largest foundation model provider about the fragility of its own infrastructure stack.