Nvidia has announced a new initiative that allows fast-growing startups to access its compute power in exchange for a share of their future profits, marking a significant shift in how AI-oriented companies can obtain critical GPU resources. The program, revealed on Thursday, targets cloud-based AI firms, model builders, and other enterprises, offering them token credits to power their development in return for both product and cloud revenue sharing with Nvidia [1].
Two initial partners have been named in the program: Sharon AI, which will deploy up to 40,000 Nvidia GPUs, and Firmus Technologies, an Australian AI infrastructure company building a data center in Batam, Indonesia. This data center is expected to scale to 360 megawatts and house up to 170,000 Nvidia GPUs, providing potential access to more than 200,000 GPUs through the partnership [1].
The move underscores the growing importance and scarcity of compute power for AI startups, with GPUs being compared to oil and reportedly tied to futures contracts due to fluctuating costs and availability issues. The sector has seen a trend of AI firms entering revenue and equity-sharing agreements with chipmakers to address liquidity challenges, with OpenAI previously engaging in similar deals involving partners such as Amazon and AMD [1].
Additionally, Nvidia recently announced plans to raise debt, with sources indicating the amount could reach at least $20 billion. The proceeds are intended for general corporate purposes, including repayment and refinancing of existing debt [1].
CONCLUSION
Nvidia's new revenue-sharing program represents a strategic move to solidify its position as a key intermediary in the AI ecosystem, providing startups with much-needed GPU access while securing a share of their future growth. The initiative highlights the ongoing challenges around compute power scarcity and the evolving financial arrangements within the AI sector.