The semiconductor trade looks all but over, with investors rushing to take profits across a wide range of names that have led the broad tech sector higher through the year. Of course, the AI boom is still on. In fact, not that much has changed regarding demand for AI and the path forward for next-generation AI data centers. Still, there’s this feeling of unease when it comes to the semi stocks, which have been making a boatload of profits in these earlier days of the great AI infrastructure buildout.
The cash is flowing out of the pockets of the hyperscalers and a whole wide range of firms that are using tokens to unlock value within the enterprise. As the chips get faster and more efficient while the price of tokens goes down, there are going to be interesting shifts across the scene as perhaps the greatest buildout of all-time moves ahead. Indeed, semis have a reputation for cyclicality.
Why are semis so heavily out of favor? And why’s it worth braving?
They boom explosively and go bust just as brutally. And with all the comparisons to the AI bubble, questions linger as to whether things are going to end in tears as they did a quarter of a century ago. I don’t have a crystal ball, and while there are strong cases for both sides, I do think that the current climate looks far more sober than the one leading up to the tech bubble bust of 2000-01.
The technology is actually starting to make money. And while not every firm is making good use of their tokens, I do think that will change in due time.
As for what wins in the next stages of AI, I think it’s the firms innovating at the application layer with agentics and workflow automation. Add robotics and consumer agents into the equation, and perhaps there is a chance that hyperscalers aren’t just going to stop spending in three or so years from now when more AI compute is finally online. Will hyperscalers grow content? Or will the upgrade cycle pave the way for more demand for chip stocks each and every year?
The case for the latter has not been shot down yet, even though you’d think it was, given the recent selling activity in the semis. Like it or not, it might not yet be curtains for semis as a painful digestion phase strikes. I think there will eventually be an opportunity as semi stocks all collectively plunge for those willing to pick and choose winners. And who knows? Perhaps the broad industry will win if the AI buildout requires constant CapEx for many years, or more than a decade, to come.
Which of the semis is intriguing on the dip?
In short, the demand story isn’t broken. But semi stocks have run so far that there doesn’t need to be any such negative shift to spark a sell-off. SK Hynix (NASDAQ:SKHY), Taiwan Semiconductor (NYSE:TSM | TSM Price Prediction), and Nvidia (NASDAQ:NVDA) have been major multi-baggers in recent years. And a period of consolidation or selling, I think, is absolutely normal, healthy behavior.
Whether you choose to play the big chokepoint in the AI buildout with SK Hynix, the fab side with Taiwan Semiconductor, or the best-in-breed GPU maker with Nvidia, I do think that the following trio is worthy of the watchlist as semi valuations come in.
In my view, Taiwan Semiconductor stands out as the best value because it’s got Nvidia’s business, custom silicon business, exposure to edge AI, physical AI, and, perhaps most importantly, it’s at the frontier of chip manufacturing. It’s in a unique spot, and there is no AI boom without the name, given a stark lack of alternatives at such a critical moment in the AI infrastructure boom.
So many companies depend on the company that it’s ridiculous. And with more diversification across customers and AI themes (edge versus cloud), it certainly stands out as one of the best ways to cover most bases in this AI revolution. Wherever the AI boom spreads and whoever designs that leading inference chip, Taiwan Semiconductor is bound to win. Indeed, if Taiwan Semi goes down, so too does the entire sector pretty much. If DRAM demand collapses due to some algorithmic efficiency breakthrough, SK Hynix could come under pressure.
The case for Taiwan Semiconductor
If Nvidia loses its lead in the AI race or if it can’t keep up in the shift to inference, the shares could take a hit. But if Taiwan Semi takes a hit? More than a handful of companies will also feel the heat. It’s literally the company that cannot afford to fumble in the slightest. Given its exceptional track record of operational excellence, I don’t expect it to.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks