A recent report by The Wall Street Journal, cited by Wccftech, alleges that Nvidia artificially sustained high demand for its GPUs by investing in AI startups required to purchase its hardware. The claims emerged on November 13, 2026, just after the departure of longtime CEO Jensen Huang, and suggest that the strategy was designed to offload an oversupply of chips from TSMC and maintain market momentum.
The report describes a pivotal August 2026 meeting between Huang and senior finance executives, including CFO and acting CEO Colette Kress. Kress reportedly raised concerns about the scale of Nvidia’s $250 billion agreement with OpenAI and proposed scaling back on circular financing arrangements—where Nvidia funds clients who then buy its products. Huang allegedly responded aggressively, calling Kress a “deer in the headlights” and pushing instead for deeper investment in small-scale cloud providers.
Pushing chips through startup pipelines
According to the WSJ, Nvidia faced an “avalanche” of GPUs from TSMC with no guaranteed buyers. As major AI hyperscalers like Google, Amazon, and Microsoft increasingly shifted toward custom silicon, demand for off-the-shelf Nvidia GPUs risked declining. To counter this, the company allegedly directed capital toward AI startups and cloud providers with the implicit—or in some cases explicit—expectation that they would use the funds to build data centers using Nvidia hardware.
This approach effectively recycled Nvidia’s own capital back into GPU sales, creating a self-sustaining demand loop. While such financing models are not illegal, they blur the line between organic market demand and engineered scarcity, especially when the recipients lack independent purchasing power.
Financial risks and market distortions
Kress reportedly warned that many of these startups were not financially viable and could become “a financial dead end”—unable to sustain operations or repay investments. This raises questions about the long-term stability of such arrangements and whether they mask underlying weaknesses in Nvidia’s sales pipeline.
The strategy may have temporarily inflated demand metrics, supporting Nvidia’s market valuation and supply chain leverage. But it also exposed the company to concentration risk: if these startups fail or reduce spending, the glut of unsold GPUs could re-emerge, potentially destabilising pricing and investor confidence.
Geopolitical and supply chain implications
The UK, UAE, and US tech sectors rely heavily on access to high-performance computing hardware, and Kazakhstan is increasingly investing in AI infrastructure and data centres. Artificial scarcity in GPU supply affects all three regions differently: UK startups face higher entry barriers; UAE sovereign funds may overpay for inflated-capacity projects; and Kazakhstani firms could experience delayed deployment of critical AI tools.
TSMC’s role as the sole manufacturer adds another layer of dependency. With no alternative foundry capable of matching its 3nm and 2nm process nodes at scale, any disruption in the Nvidia-TSMC pipeline—or revelations of market manipulation—could trigger broader supply chain recalibrations.
What it means
If verified, these allegations suggest that reported demand for AI hardware may not reflect real-world adoption but rather financial engineering. For founders and IT leaders, this underscores the need to assess supply chain transparency and avoid overreliance on single vendors. Businesses should diversify compute strategies—considering alternatives like AMD Instinct, Intel Gaudi, or cloud-based inference services—especially when planning long-term AI infrastructure.
Regulators in the US and EU may also take interest, as antitrust authorities have previously scrutinised practices that distort market signals. While no formal investigation has been announced, the timing of the report—following a leadership change—could prompt internal audits or whistleblower cooperation.
The situation highlights a broader tension in the AI race: the gap between perception and reality. As governments and investors pour capital into AI, they must distinguish between genuine technological progress and financially amplified signals designed to sustain valuations.