If you want to find new investment ideas and give yourself a good shot at being positioned in promising stocks, checking in on what billionaire hedge fund managers are holding is a great way to start your search.
One manager I follow is Daniel Loeb at Third Point. He has over a quarter of his portfolio in five genius artificial intelligence (AI) stocks that could be primed to soar this year. So if you're looking for investment ideas for 2026, I recommend considering this list.
Image source: Getty Images.
One sector, three categories
Third Point's concentration among these five looks like this:
- Amazon (AMZN 1.52%): 7.4% of the portfolio.
- Microsoft (MSFT 0.75%): 6.9% of the portfolio.
- Nvidia (NVDA 1.53%): 6.4% of the portfolio.
- Taiwan Semiconductor Manufacturing (TSM 4.96%): 3.7% of the portfolio.
- Meta Platforms (META 2.52%): 1.9% of the portfolio.
Though four of them are components of the "Magnificent Seven," these five diverse companies represent three distinct industries in the AI space: hardware, facilitators, and applications.
The first, hardware, includes Nvidia and Taiwan Semiconductor. Nvidia makes graphics processing units (GPUs), which have been the go-to computing unit for nearly every AI hyperscaler since the AI megatrend began to take shape in 2023. Although challengers are rising in the parallel processing world, Nvidia's chip dominance cannot be understated.
However, Nvidia only designs the hardware; it doesn't manufacture it. For that, it's heavily dependent on Taiwan Semiconductor's top-flight foundries. Without Taiwan Semiconductor, there would be no Nvidia.
NASDAQ: NVDA
Key Data Points
Both of these companies' revenues should continue rising as long as spending on data centers keeps increasing. All indications suggest that will be the case in 2026, and that spending could further increase rapidly in years to come as well.
These two are great foundational stocks to invest in AI, and they make up key parts of my portfolio. But hardware is not the only AI sector out there.
Not every company can build a multibillion-dollar data center
Amazon and Microsoft fall into the facilitator category. It's not feasible for every AI start-up to construct its own data center, and a vast number of businesses want to deploy AI solutions, but don't have the computing power in-house to do so. This is where cloud computing businesses step in. Amazon Web Services (AWS) and Microsoft Azure are the largest and second-largest cloud computing platforms right now, and their business model is simple: Build excess computing capacity and rent access to it. This has worked out well for both companies. This isn't a winner-take-all industry, either. There is plenty of room for multiple service providers, and Microsoft and Amazon have already benefited greatly from a massive increase in demand driven by AI workflows.
NASDAQ: MSFT
Key Data Points
The available amount of AI computing capacity is smaller than the demand for it, and demand levels are expected to rise over the next few years. As a result, investing in Microsoft and Amazon is a smart move.
Last are the companies developing AI applications. Third Point's exposure to that part of the industry is small, but that makes sense. So far, the best way to profit from AI investments has been to invest in hardware companies and facilitators, as AI applications haven't achieved widespread use yet. However, that is coming.
Meta Platforms is integrating generative AI solutions into its platform and using AI to improve ad conversion and the amount of time users spend on its apps. However, a true digital assistant or integration of AI into wearable glasses (like Meta is pursuing) hasn't come about yet. Developments along those lines by Meta could lead to a massive new source of profits for the company and healthy gains for its shareholders.
I think splitting the AI industry into buckets like this is a smart way to approach it as an investor, and keeping a heavy emphasis on hardware companies and facilities would be a smart move in 2026. However, having some exposure to application plays like Meta would also be a good idea, as you never know when the next big AI product will take the world by storm.
