Nvidia may have enjoyed a red-hot run on the back of the artificial intelligence craze, but many other tech stocks have also benefited. As with any investment theme, diversification is required. Veteran tech investor Paul Max told CNBC Pro that even if Nvidia is a “great story,” it’s “too risky to be the only one” when it comes to AI. Tech giants and semiconductor stocks have drawn investors’ attention, given the chips and other systems needed to expand early AI-stage infrastructure. CNBC Pro talks to fund managers to find the best alternatives to Nvidia that investors can consider. Semiconductors Jordan Kwietanowski, portfolio manager at Sydney-based Pella Funds Management, named Taiwan’s TSMC and Dutch firm ASML as two stocks to buy. Both are listed in the United States. Cvetanovski, who manages the Pella Global Generations Fund, says TSMC is a particularly good way to play growth in tech and AI. “TSMC is one of the cheapest technology companies out there,” he said. “Because there is no TSMC without TSMC, there is no Nvidia without TSMC,” Cvetanovski said. ASML has a monopoly on EUV lithography machines, which are needed to make advanced processor chips. The three stocks are interdependent: The American chipmaker relies on TSMC to make its graphics processing units. TSMC, in turn, uses machines made by ASML to make advanced semiconductors. Cvetanovski believes that competing with advanced microdevices such as players Despite this, Nvidia will continue to dominate the chip industry. Ray Wang, principal analyst and founder of Constellation Research, also named TSMC, saying it “always wins.” He also believes that AMD “will come close.” “The Data Center Play Data centers are also poised to benefit from AI, which has a lot of power in their applications. Cvetanovski singled out Vertio as a beneficiary. “AI needs more data centers, all of that demand will require more data centers, more sophisticated cooling systems and all kinds of other things that go along with AI investment,” he said. Cvetanovski said Vertiv is experiencing 10% to 12% growth and further growth is “still not a big obstacle for them over the next three to five years.” Super Microcomputer Max, co-chief investment officer at Harvest Portfolio Management, says Super Microcomputer is his favorite alternative AI stock to play right now. He said that since AI is still in the infrastructure building phase, AI products are only coming in 2025 or 2026. “So with this thesis, Supermicro builds custom servers that AI users use, so it uses Nvidia’s chips in its servers and [Super Micro] People like Microsoft sell their servers, and then they put them in a data center. And so SuperMicro has done a great job of shifting its focus to AI users. The stock has risen astronomically over the past year, but Meeks is confident it and Nvidia will continue to beat earnings estimates. “And as long as they continue to beat analyst numbers, you know the stock should go up … because Wall Street is what … you do the opposite of what you expect,” he told CNBC Pro last week, as That will be infrastructure plays, according to Amazon, Alphabet, Meta and Microsoft, Macs, because the cloud data centers — owned by these tech giants — will be needed to train and run AI models. Shaping the Internet The current phase of “The people who made the initial money were companies like Cisco because they provided networking for the Internet. At that time in the late 90s, the largest market cap in America was Cisco because it was plumbing the Internet. Many investors want to buy new exciting ideas, but in this case, “the strong get stronger,” he said. “You want to stay with the established companies because they will benefit the most from AI because they can spend money,” he said, adding that building AI infrastructure is extremely expensive. Companies like AMD, Amazon and Meta — he has some Also named startups. Next Nvidia may be a few companies that specialize in the tensor processing unit space, he said, which are application-specific integrated circuits for Google’s AI accelerator, neural network machine learning. for, using Google’s own TensorFlow software, he explained. One name he highlighted was OpenAI CEO Sam Altman’s TPU startup Tigris. He also named Groq. Dia, a Google-funded startup that develops custom AI chips for driving models.
