A new report has raised questions about whether NVIDIA helped keep GPU prices high by financing AI companies that purchased its chips, potentially allowing the semiconductor giant to clear excess inventory while artificially inflating demand. The allegations come amid mounting scrutiny of the semiconductor giant following the abrupt ouster of longtime NVIDIA CEO Jensen Huang.
Huang’s 33-year tenure at the company he co-founded came to an abrupt end on November 12, following revelations about multiple extramarital affairs, including one with a woman suspected by U.S. authorities of working for Chinese intelligence and helping circumvent the ban on top-end chip sales to China. NVIDIA appointed longtime CFO Colette Kress as interim CEO while its board searches for a permanent successor. Previously little-known details about Huang’s leadership and business practices have begun emerging in the aftermath.
NVIDIA Allegedly Used Financing Deals to Sustain Demand for Its GPUs
-
Image by GameRant | Source: Nvidia -
Image via Nvidia -
In the latest example of that trend, a November 13 investigation by The Wall Street Journal revealed that NVIDIA had explored providing billions of dollars in financial guarantees to startups that otherwise could not afford its chips. The arrangements primarily targeted startups in the cloud provider space, as well as data center developers, trying to make long-term customers out of businesses with limited purchasing power. Some of the deals were spearheaded by NVIDIA Vice President Nico Caprez, Huang’s son-in-law, who joined the company in early 2024 as a corporate development manager.
Caprez and fellow VP Raj Mirpuri reportedly began promoting these credit-support arrangements in spring 2026, attracting interest from dozens of companies. NVIDIA’s backing allowed prospective customers to secure loans at favorable rates, while the chipmaker proposed receiving a share of their future revenue in return. The strategy raised (even) more questions about circular financing in the AI space and related industries, as the practice of a supplier effectively funding demand for its own products is not considered economically sustainable.
NVIDIA CEO Reportedly Feared an ‘Avalanche’ of Unsold AI Chips
The WSJ report also describes how Huang expressed concerns about NVIDIA’s existing supply commitments during an August 2026 meeting with senior finance executives. Major cloud providers were exploring alternative processors, potentially reducing their reliance on NVIDIA just as enormous quantities of chip wafers and memory the company had committed to buying (largely from TSMC) were approaching delivery. Huang reportedly characterized the incoming supply as an “avalanche.” This apparent risk of excess AI hardware commitments contrasts with conditions in the consumer market, where the RTX 5090 graphics cards and other modern GPUs had become increasingly difficult to find, partly because AI buyers were competing for available inventory.
NVIDIA’s Financing Strategy Faced Internal Pushback
The financing plans faced resistance from Kress, with the now-interim-CEO questioning how NVIDIA selected borrowers, valued its guarantees, and calculated potential returns. During the August meeting, Huang reportedly disparaged her as “a deer in the headlights.” Kress also helped reduce a proposed financial backstop for a major OpenAI data-center project in Ohio, with NVIDIA subsequently disclosing a $105 billion cap on its guarantee exposure. By late August, the company had paused some financing proposals and introduced greater executive and board oversight. Huang nevertheless continued emphasizing NVIDIA’s pricing power, reportedly warning cloud customers in October against concentrating too much inventory among a small number of buyers. Meanwhile, growing AI infrastructure demand contributed to PC memory prices reaching new highs during 2026 as manufacturers prioritized more profitable server products.
Gaming GPU Prices Are Still Spiraling Out of Control as of Late 2026
Whether NVIDIA’s financing arrangements indirectly made gaming GPUs more expensive is inconclusive as of yet. TechSpot’s November 2 analysis found that graphics card prices rose approximately 35% between August and October 2026, based on a 10-country sample. In the U.S., the cheapest RTX 5080 was listed for $2,200 against a $999 launch MSRP, while the RTX 5090 reached $9,400 compared with its original $1,999 price.
