The artificial intelligence (AI) infrastructure boom has created some massive winners, and it’s likely to keep minting winners well into the future. AI is perhaps the biggest technological shift the world has seen, and right now it’s a race to see which companies will win. So if you think AI data center spending is set to soon peak, I’d think again.
Two of the companies that have been leading the AI charge are Nvidia (NASDAQ: NVDA) and Taiwan Semiconductor Manufacturing (NYSE: TSM). Both stocks have outperformed over the past year, but one looks better positioned for the long term.
Will AI create the world’s first trillionaire? Our team just released a report on the one little-known company, called an “Indispensable Monopoly” providing the critical technology Nvidia and Intel both need. Continue »
It’s hard to overstate how dominant Nvidia has been over the past several years. The company has seen parabolic revenue growth and managed to garner about a 90% market share in the graphics processing unit (GPU) space, which are the chips that have been fueling the AI revolution.
Nvidia also didn’t stumble into its role of being the AI infrastructure leader. This was a carefully orchestrated move that was set into motion well before AI became mainstream. It built a free software platform (CUDA) and seeded it into the places where early AI research was being done, and smartly acquired a conflicted data center networking company (Mellanox) that was ahead of its time.
Nvidia has shown an ability to move to where the ball is going before it is even passed. This is why it has been a market winner and why it will continue to be one. Its “acquisitions” of Groq and SchedMD are the latest examples of this. Its licensing of Groq’s technology gives it a more compelling solution for AI inference that it can plug into its CUDA ecosystem. SchedMD, meanwhile, provides it with an important software element that can be critical with agentic AI.
TSMC has ingrained itself as one of the most important players in the AI value chain. Its scale and technological expertise have given it a near monopoly in the manufacturing of advanced chips. This includes GPUs, AI ASICs (application-specific integrated circuits), high-performance central processing units (CPUs), and other logic chips.
This essentially positions TSMC as the arms dealer in the AI infrastructure race. If a company wants its advanced chip designs to be manufactured at scale, it needs to go through TSMC. It’s basically the only option right now to get these chips manufactured at high yields with few defects. Consequently, chip designers don’t just book floor space; they enter a multiyear technological marriage with TSMC where architectural roadmaps and capacity commitments are co-designed years before a single chip is even produced.