[ad_1]
Wall Street is abuzz with excitement around artificial intelligence. At the center lies Nvidia , which designs and sells graphics processing unit (GPU) chips that are used in data centers and go on to power AI applications. The blowout earnings it reported in late May came on the back of a huge forecast beat driven by strong AI chip demand , and triggered a wave of buying into the stock. But Jordan Cvetanovski, chief investment officer and portfolio manager at Pella Funds Management, told CNBC Pro this week that he wouldn’t buy Nvidia as it’s simply too expensive right now. Instead, he would buy Taiwan’s TSMC and Dutch firm ASML . Both are tied to how well Nvidia does: The U.S. chipmaker depends on TSMC to manufacture its GPUs. TSMC, in turn, uses machines made by ASML to make the most advanced semiconductors. Why Nvidia is a ‘poor investment’ Cvetanovski said Nvidia is likely to capture a large portion of the initial wave of investments into GPU stacks in data centers, and relies on an extreme positive outlook for this area. “However, there are already other players with GPU alternatives and some of the very large technology companies like Apple have been investing in their own solutions. The point is that competition in GPUs will eventually intensify and margins and revenue share that Nvidia enjoys will diminish,” he said. Investors would be paying for more than 40 times EBITDA (earnings before interest, taxes, depreciation and amortization) by the end of this year, with markets pricing in around 20% growth for the next decade, said Cvetanovski. “This is a hefty valuation with very little margin for even the smallest slippage,” he said. “AI is a great story, Nvidia is a great story but a poor investment due to its valuation.” ASML versus TSMC While ASML and TSMC will both benefit from AI, one has an edge over the other, according to Cvetanovski. “So we think that a lot of the positivity is priced into Nvidia, however, we think buying TSMC and ASML is actually a much better way — to play it to invest in the space now given the valuation differential, but also the different potential there,” Cvetanovski said. He said TSMC offers a more diversified exposure to “general advancement in global technology.” “TSMC is arguably a great story at a great valuation making it a very strong investment for us. Even before the Nvidia upgrade, TSMC was very cheap, it has almost the same EBIT margins as Nvidia, expected to grow at high teens vs Nvidia at high 20s but at a fraction of the valuation,” he said, adding that it’s still cheap even now. “TSMC won’t care which GPU wins in the long term as it will supply all GPU suppliers,” he added. He said both TSMC and ASML have become monopolies in their respective areas: TSMC in leading-edge chips — “they are the only ones that can really make this on a grand scale” — and ASML in extreme ultraviolet lithography (EUV) machines, in which it has 100% market share. Cvetanovski described ASML as a “keystone in all technology.” “The whole advancement in technology is relying on them,” he said, referring to its EUV machines. “ASML does not abuse its monopolistic position and will not take advantage through price, but it will certainly enjoy volume growth and stability in its order book over time. Although it has done well it is still undemandingly valued on a FCF basis relative to its visibility and growth potential,” he said. If he had to pick one stock, however, he would give TSMC more upside — around 30% from its price today, while he believes ASML is already fairly valued. “TSMC is a better investment in terms of returns. We think it’s cheaper, and offers a better return potential than ASML assuming there are no risks in any way from anything negative happening,” said Cvetanovski. “However, we think although TSMC is a leader in what they do, ASML is one of the best companies in the world — without ASML there’s no TSMC, without TSMC there’s no Apple, without Apple, there’s no iPhone.” According to FactSet, analysts covering both TSMC and ASML give them around 10% upside each on average.
[ad_2]
