Disclosures on Interconnected AI Deals Draw SEC’s Attention

Key Facts

Bloomberg Law reported that increasingly interconnected deals between major technology companies and AI developers have drawn the attention of the U.S. Securities and Exchange Commission, with a focus on accounting disclosures. The issues include how to record revenue when a supplier is also a customer of an AI provider, and whether technology companies should include certain related entities on their balance sheets. The scrutiny concerns how these relationships are presented in financial statements, not an announcement of a formal investigation or an allegation of wrongdoing. That puts the quality of disclosures about commercial and financing relationships in focus as AI infrastructure deals expand. Melissa Ramenbor, deputy chief accountant in the SEC’s Division of Corporation Finance, said disclosures concerning relationships between providers and suppliers may extend across multiple parts of financial statements.

The significance of the supplier-customer relationship goes beyond how the parties to a deal are classified, as it may affect how revenue, obligations and related entities are presented in financial statements. Ramenbor said accounting disclosures about relationships between providers and suppliers “may include a range of financial statement line items.” The breadth of disclosure means investors may need to assess these relationships across more than one line item when evaluating the companies involved. The issue carries particular weight when deals to sell services or equipment overlap with agreements to purchase computing capacity. According to the report, the SEC’s questions focus on the accounting judgments involved in recording revenue for suppliers that also buy services from AI companies, and on the threshold for consolidating related entities on balance sheets.

The scrutiny comes as cloud hyperscalers and AI developers increasingly rely on off-balance-sheet entities to finance access to server farms, Bloomberg Law reported. These structures provide financing for infrastructure and access to computing capacity through arrangements that may not appear in the same way as direct assets and liabilities. The question of whether related entities should be consolidated on balance sheets is therefore tied to how investors understand the financing used to expand infrastructure. The report points to growing reliance on these entities without specifying the scale of the financing or which companies are subject to a particular review. The accounting focus remains on whether links between companies and their suppliers and customers are clearly presented in financial statements.

CoreWeave’s disclosures for the quarter ended September 30, 2026, show the degree of revenue concentration at a computing infrastructure provider: about 30% of revenue came from its largest customer and 24% from its second-largest customer. No other customer accounted for 10% or more over the same period. These figures show significant reliance on a limited number of customers, an important context for evaluating supplier-customer relationships in the computing sector. The company said OpenAI committed to pay up to about $6.5 billion through May 31, 2031. Meta also initially committed to pay up to about $21.0 billion under an agreement covering access to computing capacity through December 20, 2032, with an option extending through April 10, 2032.

Intel’s disclosures offer another example of how accounting classification changes when an entity becomes an external customer. The company recorded revenue from Altera as a customer of $126 million in the third quarter of 2026 and $446 million year to date. In the corresponding periods of 2025, Altera had contributed revenue of $367 million in the third quarter and $1.1 billion year to date within Intel’s consolidated results. Intel attributed the change to Altera becoming an external customer after it was deconsolidated in the third quarter of 2025. The example highlights how a change in a company’s relationship with a related entity can alter how revenue is presented between periods, making disclosure of the relationship’s nature and timing important to understanding the figures.

The regulatory backdrop also includes other reviews in the semiconductor and AI sectors. Nvidia disclosed in its annual report that it had received broad and continuing requests for information from competition authorities in the United States, the European Union, the United Kingdom, China and South Korea. According to the company’s disclosure, the requests covered its sales, supply allocations, AI partnerships, and agreements with customers and suppliers, and it expected to receive additional requests. On October 30, SEC Director of Enforcement David Woodcock said the agency was examining disclosures by companies operating in AI amid investor concerns about a potential bubble fueled by circular deals. He added that the SEC’s role was to ensure investors received a complete picture of material risks. That puts corporate disclosures at the center of the debate over interconnected deals and the financing of sector expansion.

The scrutiny of disclosures coincided with a decline in major technology stocks during the week covered by Reuters on October 30, after companies revealed heavy cash spending to expand their AI businesses as interest rates were rising. That raised questions about the sustainability of AI spending. In a report published by Reuters on November 9, banks and credit investors showed greater caution toward AI-linked debt deals, with some banks saying they would hold back from financing additional deals even when backed by guarantees from major technology companies. Reuters said the market decline erased $2.5 trillion from the Nasdaq 100’s market capitalization on Thursday and pushed the index into a correction. These developments reflect the sensitivity of infrastructure financing to market volatility and the assessment of customer obligations.

Looking ahead to financing, Goldman Sachs Research forecast that global debt issuance by cloud hyperscalers would reach $420 billion in 2027, an increase of more than 60% from its estimate of $250 billion for fiscal 2026. At the same time, lenders and investors are testing cash flows linked to AI infrastructure under economic slowdown scenarios and scrutinizing contract terms. Their review includes service delivery obligations, operational timelines and the circumstances that could allow major customers to reduce or terminate their commitments. The scale of expected financing and the terms of contracts keep disclosures about relationships among providers, customers and related entities relevant to the assessment of these deals. Accounting and regulatory questions continue alongside funders’ search for greater clarity on cash flows and contractual obligations.