TECH NEWS – OpenAI CEO Sam Altman wants to purchase 100 million AI GPUs, which would require staggering amounts of value and power!
The hype surrounding AI is unstoppable. Despite the fact that we’ve heard that Microsoft, Google, and Meta are building large-scale AI clusters, organizations are still reporting shortages in computing capacity. OpenAI seems to be in a similar situation. Sam Altman claims to have big plans, one of which is to acquire one million AI chips. While this goal may seem ambitious, nothing seems impossible for OpenAI and its elite AI team.
Altman is known for his impressive investment track record. A few months ago, he was traveling the world raising funds to build his network of chips, but the project has disappeared. Even if OpenAI needs the processing power of 100 million AI chips, the company would require trillions of dollars, nearly equaling Nvidia’s current market capitalization. Therefore, acquiring such a large number of AI chips seems impossible for the time being. However, given that GW-level AI clusters are becoming more common, we can’t rule it out completely.
we will cross well over 1 million GPUs brought online by the end of this year!
very proud of the team but now they better get to work figuring out how to 100x that lol
— Sam Altman (@sama) July 20, 2025
Assuming each chip requires 750 W of power, 100 million AI GPUs would represent a 75 GW cluster, equivalent to 75% of the UK’s total network capacity. Unless OpenAI can get nuclear power somehow, it will still need 75 GW. That alone makes Altman’s target audacious. The industry believes that accumulating AI clusters could lead to AGI (general AI), so companies are focused on continually and significantly increasing the number of AI chips.
Companies are undoubtedly engaged in a race for AI infrastructure and are prepared to spend hundreds of billions of dollars to win. Spending on AI by big tech companies is steadily increasing, and it’s safe to say Nvidia will make a significant contribution.
Source: WCCFTech




Leave a Reply