AMD vs. Broadcom: The AI Chip Race Is Big Enough for Both. Here's the Better Buy.

Nvidia remains the AI infrastructure market leader, and the chipmaker isn't going to give up its crown anytime soon. However, this is a large and rapidly growing market that is big enough for more than one winner. Two other AI semiconductor stocks that look poised to perform well over the coming years are Advanced Micro Devices (NASDAQ: AMD) and Broadcom (NASDAQ: AVGO).

While I like both stocks, I think Broadcom is the better buy. Let's take a closer look at both to find out why.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

AMD has long played second fiddle to Nvidia in AI. The reason is that most foundational AI code was written for Nvidia's CUDA software platform and optimized for its graphics processing units (GPUs), while AMD's ROCm software greatly lagged. This made Nvidia's GPUs the standard for training AI models, and essentially kept AMD out of the market.

However, AMD and its GPUs are much better positioned for the inference market. It has greatly improved its software platform over the past few years, and inference is much less technically demanding and much more about fast memory access. AMD's chiplet design allows its GPUs to be packaged with more high bandwidth memory (HBM), making them more suitable for inference.

On top of that, the company has acquired memory optimization company MEXT and inference chipmaker Taalas to help differentiate its offering. It's also formed a partnership with Cerebras to offer a combined system using Cerebras' high-end chips for the decode phase of inference and its GPUs for the pre-fill phase.

AMD has signed two large partnerships with OpenAI and Meta Platforms that should generate around $100 billion each in revenue over the next few years. It also has GPU inference deals with Anthropic and Microsoft in place. In addition to its GPU opportunity, AMD is also a leader in server central processing units (CPUs). This market is growing rapidly with the rise of agentic AI, and it is projected to be worth $220 billion over the next few years.

With hyperscalers (large data center owners) looking to save costs on their AI infrastructure spending, they have been increasingly turning toward Broadcom to help them develop custom AI chips. The company is a leader in ASIC (application-specific integrated circuit) technology, providing essential IP and services to turn a customer's design into physical chips that can be manufactured at scale.