Nvidia (NASDAQ:NVDA | NVDA Price Prediction) has been a huge winner of the AI revolution, but the big question is whether the competition is finally getting a chance to catch up. It seems like every big tech firm with enough money to spend wants to build its own custom silicon, and it makes a lot of sense, especially since how much money is flowing into Nvidia’s pockets for its GPUs. With a likely inference boom on the horizon, perhaps the custom silicon is more of a pressure release than a way to move past Nvidia.
At the end of the day, hyperscalers are making real, massive strides in efficiency with their latest silicon. From Google (whose parent firm is Alphabet (NASDAQ:GOOGL)) TPUs to Meta Platforms‘ (NASDAQ:META) MTIA chips, it looks like AI innovators are about ready to move on.
But, given the magnitude of AI demand and how quickly the appetite for tokens could rise, it feels like there’s more than enough demand to go around, even as new suppliers join the chat. Despite the new options, though, firms, including the hyperscalers making custom silicon themselves, are still using Nvidia GPUs.
But what happens when the hyperscalers bridge the gap and start gravitating more towards their own silicon while potentially selling it to third parties? That’s the big question that might keep Nvidia shareholders up at night. CapEx is blasting off now, but who knows? A pause or pullback in spend could be in the future if a digestion phase is needed. Time will tell.
The semis recent volatility might introduce a new risk
As Nvidia’s multiple compresses a bit as shares drag their feet relative to its rivals in the semi scene, I do think that the case for taking a raincheck is getting stronger. Not only is Nvidia stock not as explosive as some of the other semi stocks out there, but the shares also stand to sink if the semiconductor industry as a whole collectively rolls over. Perhaps there’s a reason why bear ETFs against the semiconductors have been so popular in the past couple of weeks.
While I do see opportunity in Nvidia longer term, I cannot say that the road higher won’t be without its fair share of bumps (or perhaps even a vicious bear market drop). I have no idea when semis will fold. Perhaps a rate hike will do it. After all, there’s nothing quite like higher rates to incentivize pulling back on spend. But, regardless, I think Meta Platforms stands out as an AI chip innovator without all of the downside risks should semis decide to nosedive tomorrow.
MTIA looks seriously impressive
Of all the AI strategies, Meta’s seems the most “Mad Max,” so to speak. If it’s not reports of data centers in tents, an AI version of Mark Zuckerberg, big spend on the superintelligence team, or the rapid advancements in MTIA, it’s clear Zuckerberg is more than willing to act as a wartime CEO in this AI revolution. And I think that’s exactly why Meta will be successful in the AI race.
In any case, the six-month innovation loop with Meta’s silicon, I believe, shows just how serious Meta is in securing its custom silicon future. With a deep focus on generative AI and recommendation inference, MTIA is very application-specific. It’s custom-tailored to its business, and it might hold the keys to the future of advertising in the AI age.
In any case, things are moving fast and in a direction that I think could be most conducive to stunning ROIs. Make it for yourself, and perhaps others might come knocking. With the door open to becoming a hyperscaler willing to sell AI compute, I think the whole Meta AI story is fundamentally misunderstood. At just 20.6 times trailing price-to-earnings (P/E), shares scream deep value hiding in plain sight.
The bottom line
Where some see Meta’s AI as a bit of a jumbled mess or behind in the race, I see it as on an aggressive launch pad. If anything, Zuckerberg is a wartime CEO who’s positioned to win big in AI as he deploys the capital and makes forward-thinking decisions that I think rivals might want to copy.
Top Gaining Stocks
Top Losing Stocks