Right now, a certain type of investor conversation is taking place at trading desks and coffee shops, and it typically begins in the same manner. Is it too late now that AI stocks are at all-time highs? A truthful response is more intriguing than a yes or no. Some businesses strive for perfection in their pricing. Others appear to have been overlooked. It’s difficult to distinguish between them, and the five names that consistently appear in serious long-term portfolios aren’t the ones making waves on social media.
Nvidia is still the most obvious—almost embarrassingly so. Nearly every significant AI model on the planet uses the company’s GPUs, and its CUDA software has created the kind of barrier that rivals frequently promise to breach but seldom actually do. That lead could be extended by two or three years thanks to the upcoming Rubin chip platform, which is intended to manage the memory-intensive requirements of AI agents and lengthy prompt threads. Skeptics continue to wait for less expensive substitutes to emerge. They haven’t done so yet. There is a sense that Nvidia’s ecosystem will remain too sticky to give up even after they do.
That’s precisely why Taiwan Semiconductor is the less glamorous relative. When you stroll through Hsinchu Science Park, you’ll see factories operating around the clock, equipment that costs more than small nations, and a workforce that takes nanoscale accuracy for granted. The chips designed by Nvidia are produced by TSMC. Additionally, it produces chips for AMD, Apple, and most other well-known brands. The company’s customer list is the closest thing the semiconductor industry has to a guarantee, but geopolitical risk still looms over the stock and always will.
The alphabet is positioned oddly. For the majority of 2024, the market was concerned that ChatGPT would destroy Google Search. At least not yet, that hasn’t occurred. The company’s AI infrastructure now powers tools used by millions of people, search revenue continues to rise, and YouTube continues to be a quiet money maker. Gemini continues to feel ahead in some areas and behind in others. However, Alphabet has resources—cash, talent, and distribution—that are nearly impossible to match. Companies that can adopt new technology instead of creating it from the ground up might be rewarded in the upcoming ten years.
People are surprised by Meta. The business was dismissed a few years ago as a failing social network that was spending billions on a metaverse that no one was interested in. The AI pivot then hit the ground. Meta’s already massive advertising business became smarter and more profitable thanks to Mark Zuckerberg’s open-source Llama models, which changed how businesses think about deploying AI. A company regaining credibility through technology bets that at first appeared reckless has an almost poetic quality.
Due primarily to AWS, Amazon completes the list. Customers are still in the early stages of developing inference workloads at scale, and the cloud unit’s AI-related revenue continues to grow. Over the course of ten years, the uninteresting backend infrastructure is likely more important to shareholders than the retail business.
None of these are certain. Regulators are agitated, valuations are stretched, and there will almost certainly be at least one painful drawdown over the next ten years. However, these five continue to earn their spot on the list if the question is which AI stocks have the resilience to truly endure ten years of holding—through hype cycles, scandals, and unavoidable corrections. It’s difficult to ignore how frequently the dull response turns out to be the correct one when you watch this develop over the past two years.
