Nvidia (NVDA +1.80%) may be the king of artificial intelligence (AI) investing for now, but it doesn't mean it will stay in that position forever. Right now, it's on top because its clients buy thousands of graphics processing units (GPUs) and place them in giant data centers to produce and train AI models.
Eventually, this demand will decline as companies build out their capacity, making way for the next wave of AI investments. As a result, I think there are three companies that could be worth more than Nvidia five years from now. They are Taiwan Semiconductor Manufacturing (TSM 1.26%), Alphabet (GOOG +0.04%) (GOOGL 0.04%), and Amazon (AMZN 1.40%). Although the climb won't be easy, I think they can displace Nvidia.
Taiwan Semi will win regardless of which company sits atop the AI throne
Taiwan Semiconductor may be a surprise pick here, as it's significantly smaller than Nvidia right now (Nvidia is valued at $2.6 trillion versus TSMC's $775 billion). However, some tailwinds will push TSMC higher.
First, Taiwan Semiconductor is more diversified than Nvidia. Nvidia is basically a one-trick pony with its GPUs and products that complement them. Taiwan Semi may also be a one-trick pony because all it does is produce chips, but the use case for those chips is so wide that it ends up being diversified. For example, Apple is one of Taiwan Semi's largest customers and could see a huge demand ramp-up soon as Apple Intelligence is only available on the latest generation of iPhones.
NYSE: TSM
Key Data Points
Second, TSMC is working on its 2nm (nanometer) chips, which are capable of the same performance as its 3nm predecessors, except that they consume 25% to 30% less power. That's a huge performance gain, and it will drive the next generation of hardware used to process AI models.
However, will the next generation of AI models be run on Nvidia hardware? Not necessarily.
Apple Intelligence isn't using Nvidia GPUs. Instead, it's using Alphabet's tensor processing units (TPUs). While this may come as a surprise to some investors, it shouldn't. Nvidia GPUs are fantastic at being flexible and processing various workloads. However, there are more efficient ways to train AI models if the workload is set up properly. That's why Alphabet's TPUs can be a better choice than Nvidia's GPUs.
NASDAQ: GOOGL
Key Data Points
For most of its customers, these TPUs can only be accessed through Google Cloud. As programmers get better at setting up AI models for efficient calculations, this will shift the workloads from GPUs to hardware specifically designed for processing AI, like TPUs. So, all of the technology that TSMC is innovating may end up being used in products like TPUs instead of GPUs, which could cause Alphabet to rise and Nvidia to tumble.
But how does Amazon fit in?
Amazon and Alphabet are remarkably similar businesses
While most know Amazon as its dominant commerce platform, it's actually the largest provider of cloud computing infrastructure globally. Like Alphabet, Amazon also has purpose-built custom chips to optimize AI training.
It will benefit from the same tailwinds that will push Alphabet higher in the next few years, except that the effect may be even more profound as Amazon Web Services (AWS) makes up a higher percentage of Amazon's operating profits than Alphabet's Google Cloud.
NASDAQ: AMZN
Key Data Points
Furthermore, Amazon's commerce business is still doing well. This is another similarity to Alphabet, as Alphabet has huge growth drivers in the AI space, but its advertising business continues to print money. As Amazon's commerce business continues to grow and become more efficient, Amazon will turn into an even larger business, potentially surpassing Nvidia within five years.
While Nvidia has undoubtedly won the first round of AI investing, there are still many to go. As companies shift to load-specific hardware, I think companies like Amazon and Alphabet will succeed, bringing Taiwan Semiconductor (the maker of these chips) along with them.
