NVIDIA Launches Revenue-Sharing GPU Framework as Burry Opens Short at $198

NVIDIA has introduced a new AI compute framework allowing cloud providers to deploy advanced hardware through revenue-sharing and credit-support agreements, targeting startups, enterprises, and regional AI operators. Simultaneously, investor Michael Burry disclosed a direct short position on NVIDIA stock at approximately $198.09 per share, intensifying debate over the sustainability of AI-driven valuations.

The new business model is designed to lower capital barriers for partners seeking to build large-scale, multi-tenant AI infrastructure. Under the arrangement, NVIDIA earns revenue both from hardware sales and from cloud usage generated by supported capacity. The company described the shift as a response to the industry moving from model training toward continuous, always-on token production, which demands higher infrastructure utilization at scale.

Two early participants highlight the initiative's scope. Sharon AI has announced plans to deploy up to 40,000 Grace Blackwell GB300 GPUs under the framework. Firmus, meanwhile, is developing a major campus in Indonesia projected to support approximately 170,000 GPUs and 360 megawatts of power capacity.

By helping partners accelerate deployment, NVIDIA aims to increase utilization rates across its hardware ecosystem while reducing dependence on a small number of hyperscale cloud customers. Analysts have noted this approach could also diminish the strategic significance of individual deals with major platforms.

On the opposing side, Burry — best known for predicting the 2008 housing market collapse — has positioned against NVIDIA alongside bearish bets on Tesla, Applied Materials, Caterpillar, and the iShares Semiconductor ETF (SOXX). He characterized the semiconductor index as representing a 'pure form of overvaluation' that is 'rarely seen,' according to Business Insider.

Burry's thesis centers on what he views as excessive capital spending by hyperscalers, rapid hardware obsolescence risk, and the possibility that AI demand growth could decelerate. He has drawn comparisons between current conditions and prior technology bubbles.

Supporters of NVIDIA counter that inference demand is accelerating enterprise adoption and reinforcing the company's dominant market position. NVDA shares were trading near $195 at the time of publication, placing the chipmaker's market capitalization at approximately $4.77 trillion.

The divergence between bulls and bears on NVIDIA has widened on Wall Street. Proponents argue the company sits at the center of a multi-year AI expansion cycle, while Burry's position reflects a scenario in which market expectations have outpaced underlying economic fundamentals.

Upcoming earnings reports, cloud partnership expansions, and infrastructure deployment progress are expected to be key indicators for both camps. Any material slowdown in enterprise spending, GPU utilization rates, or demand for next-generation platforms — including Blackwell and Rubin-based systems — could increase volatility in NVDA shares. Conversely, continued adoption growth may reinforce the bull case heading into the next several quarters.