Introduction
Sam Altman, the CEO of OpenAI, has sparked a debate about the usage of the world’s limited DRAM supply. According to him, running AI on home PCs is a poor use of this limited resource. Instead, he argues that memory should be concentrated in data centers, where it can serve users around the clock.
The Problem with Home PCs
Altman’s argument is that home PCs, such as those running on Intel Core i9 processors or equipped with Nvidia GeForce RTX 3080 graphics cards, waste the world’s limited DRAM supply. These PCs often have high-capacity RAM, such as the Corsair Vengeance LPX 64GB DDR4 RAM, but they are not utilized efficiently. Most of the time, these PCs sit idle, and the memory is not being used to its full potential.
Data Centers: The Better Option
In contrast, data centers are designed to run 24/7, and they can make better use of the limited DRAM supply. Companies like Google, Amazon, and Microsoft have invested heavily in building massive data centers, such as the Google Cloud Data Center or the Amazon Web Services Data Center, which are equipped with servers like the Dell PowerEdge R750 or the HP ProLiant DL380. These data centers can provide AI services to users around the clock, making better use of the available memory.
Example Use Cases
For instance, the Samsung Galaxy S25 can be used to access AI services provided by data centers. The Apple Watch Series 10 can also be used to control smart home devices, which are connected to the internet and can access AI services. Moreover, the iPad Air M2 can be used to run AI-powered applications, such as those using the Apple Neural Engine.