The trillion dollar stock club is a pretty exclusive group. There have only been 10 US-listed stocks that have exceeded this valuation level, and there are a few more within striking distance.
However, one stock that is a bit far is Advanced microdevices (NASDAQ: AMD). AMD has a market cap of $330 billion, so it may not be on investors’ radar for moving into the $1 trillion valuation range. However, AMD hardware is starting to become more popular in the world of artificial intelligence (AI), and it may get there faster than many think.
How fast can AMD get to $1 trillion? Well, if his predictions come true, he could be there in just four years.
Nvidia (NASDAQ: NVDA) is the undisputed king of graphics processing units (GPUs). GPUs are well-suited for AI workloads because they can process multiple calculations in parallel. At the start of the AI boom in 2023, Nvidia’s GPUs, control software, and other hardware that supports them were far superior to AMD’s. As a result, Nvidia products became the preferred option, while AMD products became only an alternative.
However, these trends are changing. AMD has made massive improvements to its control software, ROCm. It saw a 10x increase in downloads year-over-year in November 2025. That’s a big deal because it shows that developers are exploring their hardware. This could be a signal that AMD products are starting to become a viable alternative and could be poised to steal market share from Nvidia.
There is only so much money available to build data centers. Computing hardware can account for nearly half the cost to build them, and while Nvidia’s products are the best, they’re not cheap.
There are no prices available for flagship data center GPUs, and these figures are estimates from reports. Nvidia’s Blackwell B200 GPU can cost between $30,000 and $50,000 per chip, depending on options. Its AMD rival, the MI350, costs $25,000. This allows AI hyperscalers to get more money for cloud GPUs, but it remains to be seen whether this cheaper price is worth it in terms of performance.
However, AI hyperscalers may not have a choice if they want to use AMD chips anymore. During Nvidia’s Q3 results, the company announced that it was “sold out” of cloud GPUs. While this may seem like a good problem to have, the bigger problem is that customers can turn to AMD products as an alternative if they can’t get the computing power they need from Nvidia. If AMD products can deliver similar results at a lower price, this could create a scenario where more customers choose AMD hardware in the future.