AMD to increase supply of GPU and CPU chips
AI data centers have essentially ruined gaming and the tech industry as a whole. The people who run these sites have purchased so many RAM sticks and GPUs that the prices of some individual components eclipse the cost of game consoles several times over. The problem is so bad that some manufacturers are producing older RAM models to try to meet demand. Unfortunately, whenever someone comes up with a solution, it often favors AI companies.
Recently, reports emerged that AMD plans to double its GPU and CPU production. However, these are not the chips you can use in your average computer. Instead, AMD will focus on GPUs and CPUs that power AI data centers. According to media outlets such as Yahoo! Finance, AMD CEO Lisa Su recently met with representatives from numerous companies to discuss supply chain activities and help AMD meet the demand of AI organizations. These include Foxconn and TSMC in Taiwan, as well as Samsung and SK Hynix in South Korea.
While AMD spokespeople didn’t say the company would stop producing GPUs and CPUs for the average customer, they also didn’t offer any guarantees to the contrary. This deafening silence has many concerned, as AMD’s plans may only exacerbate the damage AI hubs are doing to the computer component market. You can currently build a budget gaming PC for less than $1,000, but for how much longer if companies like AMD prioritize AI data centers?
It is not because A = B that B = A
If you’ve read the marketing for the Nvidia GeForce RTX 50 Series and AMD Radeon RX graphics cards, you’ve probably noticed that both companies are focusing on AI. Sure, most people buy RTX GPUs (and Radeon RX GPUs) for ray tracing, but these chips can still power agentic AI models. Still, AMD is not focusing on these particular components because they are not enough for AI data centers.
Even though GPUs (and, to a lesser extent, CPUs) can run AI models, this is little more than an unintended use of their capabilities. To translate data into visual information, components must perform countless calculations. GPUs are best suited for this task because they specialize in parallel processing, distributing tasks simultaneously across multiple processors to speed up the process. AI models require parallel processing to run, which is why many data centers use tons of GPUs. But they also need processors to administer these tasks.
That said, GPUs and CPUs in AI data centers are more advanced than those in an average computer, because the more power an AI model can access, the more powerful it becomes. Manufacturers can’t build these superchips without using the same components and manufacturing lines reserved for consumer GPUs and CPUs, and companies will ultimately choose the strategy that makes them the most money. It’s unfortunate for consumers, but that’s the way things are going right now.
