The four largest hyperscalers (owners of large data centers) have said they have plans to spend more than $650 billion this year building out their artificial intelligence (AI) infrastructure. Let's look at several AI stocks investors can buy that are likely to benefit from this huge spending spree.
Chips, foundries, and memory makers
The most direct beneficiaries of the AI data center build-out are chip and memory makers. Nvidia's (NVDA 1.46%) graphics processing units (GPUs) are the main chips used to power AI workloads, and its CUDA software platform, which is where most foundation AI code has been written, continues to give it a wide moat in training. It is also a leader in AI inference, although Advanced Micro Devices (AMD 3.29%) has carved a niche and has made some recent large deals with OpenAI and Meta Platforms (META 2.43%) in this area.
Image source: Getty Images.
Broadcom (AVGO 1.32%), meanwhile, is seeing a huge opportunity with custom AI ASICs, which are hardwired chips built for specific tasks. The company helped Alphabet (GOOGL 1.45%) (GOOG 1.45%) create its Tensor Processing Units (TPUs), which it is using to power most of its internal data center infrastructure. This is a big driver in and of itself, but it is also helping other companies, including OpenAI, develop their own custom chips, which should help fuel growth for many years.
Taiwan Semiconductor Manufacturing (TSM 4.55%) is also a great option, as it has a virtual monopoly on manufacturing advanced logic AI chips. This also gives it pricing power.
NYSE: TSM
Key Data Points
Since AI chips require high bandwidth memory (HBM) for optimal performance, Micron Technology (MU 2.98%) is another smart way to play the AI infrastructure boom. HBM is a special form of DRAM, and given that it needs upward of 3 times the wafer capacity as regular DRAM, the entire DRAM market is now in short supply, pushing up prices. Micron is one of the three big DRAM memory makers, along with Korean companies Samsung and SK Hynix, and the company has seen its revenue and gross margin soar. Most importantly, it is now able to lock in long-term HBM commitments, making its business much less cyclical than in the past.
Not to be overlooked, with the rise of agentic AI, central processing units (CPUs), which act as the brains of a computer, are becoming increasingly important. AMD is the leader in this space and has a nice opportunity. Arm Holdings (ARM 4.26%) and Intel (INTC 5.22%) are other potential beneficiaries.
NASDAQ: MU
Key Data Points
The hyperscalers
When the largest tech companies are planning to spend the equivalent of the gross domestic product of a relatively large country on AI infrastructure this year, you can bet they are expecting a strong return on their investment. Three of those big spenders are in the cloud computing space, with Alphabet, Amazon (AMZN 1.44%), and Microsoft (MSFT 0.48%) planning to spend big on AI this year. I like all three stocks, as their cloud revenue is only going to accelerate from here.
Meanwhile, all three companies are doing a great job of incorporating AI into their core businesses to help drive growth. Alphabet is using its Gemini model to drive search queries and growth, while Microsoft's AI copilot assistants have been powering its enterprise software growth. Meanwhile, Amazon is using AI and robotics in its e-commerce operations to drive efficiencies and increase its operating income.
The fourth large hyperscaler spending big is Meta Platforms. The social media giant has embedded AI into its recommendation algorithm to boost user engagement, while also creating AI-powered tools to help advertisers create more enticing campaigns. This is leading to both more ad impressions and higher ad prices, driving Meta's revenue growth.
NASDAQ: META
Key Data Points
Energy pipelines
AI uses a lot of energy, so one conservative and under-the-radar way to play the AI boom is with pipeline company Energy Transfer (ET +0.03%). The company has strong natural gas assets in the Permian Basin, which is one of the cheapest sources of natural gas in the country. As a result, it is in the midst of a lot of high-return, natural-gas-related projects, both directly and indirectly, centered around AI data center power needs. With a cheap stock price and over 7% yield, this is an alternative way to play the AI build-out.
