Goldman Sachs has cautioned that Nvidia’s multibillion-dollar investments in customers like OpenAI and Intel could be distorting the company’s top line. In a recent note, analyst James Schneider said Nvidia’s equity financing to customers is beginning to resemble “circular revenue,” where funds flow out as investments and return through purchases of Nvidia GPUs, effectively boosting reported growth. He added that when equity investment comes from a supplier, additional scrutiny is warranted, particularly if the same money recirculates through commercial agreements.
Beyond the accounting optics, Goldman Sachs frames Nvidia’s checks as strategic, not merely financial. The bank argues the deals deepen lock-in to Nvidia’s CUDA software stack, expand hardware pull-through, and shape how the Artificial Intelligence market evolves. Schneider said the transactions signal Nvidia’s confidence in the scale of the opportunity and reinforce its market leadership. A prominent example cited in the note is Nvidia’s arrangement with OpenAI, where Nvidia expects significant revenue in 2026 and a portion of profits is slated to be reinvested back into the business, underscoring how intertwined the commercial and funding relationships have become.
The bank also highlights a shifting customer mix for Nvidia’s datacenter business. While Big Tech buyers such as Alphabet, Meta, and Amazon remain the most reliable demand base, Goldman Sachs expects a growing share of revenue to come from Artificial Intelligence startups and government programs. Those cohorts, the note warns, are less stable and depend heavily on external capital, which increases the risk that spending could prove volatile if funding conditions tighten or priorities change.
Goldman Sachs estimates that OpenAI will require substantial capital in 2026 to run and train its Artificial Intelligence systems, with needs stretching beyond near-term compute to longer-term obligations, including Nvidia-backed data centers and planned Stargate infrastructure. The bank expects the funding to be pieced together from OpenAI earnings, Nvidia’s investment, and additional equity or debt, yet still sees a cash shortfall even after available reserves. With thinner margins than hyperscalers and greater exposure to market cycles, startups like OpenAI may find it harder to sustain steady infrastructure spend. Schneider concludes that the sustainability of OpenAI’s outlays will increasingly hinge on new equity and debt raises, a dynamic that could test the durability of Nvidia’s customer-driven growth.
