The Mag 7’s Chip Strategies Just Got Serious, and Nvidia’s Once‑Clear Lead Looks More Contested

Nvidia (NASDAQ:NVDA | NVDA Price Prediction) is the undisputed AI chip king, and its heft lead probably won’t be given up anytime soon, even as more tech titans spend on their own custom silicon efforts. Just because there is a growing number of players entering the AI chip chat does not mean that it’s time to hit the worry button, even if demand shifts gears from training to inference and ASICs start to gain real steam.

Until Vera Rubin sales come up short of expectations (I just don’t see this happening as Blackwell and Vera Rubin sales look to eclipse $1 trillion next year), it’s looking like Nvidia will continue to lead the way, even as the rest of the Mag Seven look to take control of the silicon layer.

As the tech titans look to allocate even more resources towards designing even more efficient training and inference chips, while Elon Musk looks to take things a step further by getting into the business of fabs with the Terafab, the questions linger as to how the AI chip landscape will shift heading into the 2030s.

Google TPUs are hard to ignore any longer

Any way you look at it, custom silicon and sovereign foundries might be the way of the future. But even if Google TPUs continue to make big leaps, the total addressable market (TAM) might still be so underestimated that Alphabet (NASDAQ:GOOGL), Nvidia, and the rest of the Mag Seven might all stand to win from AI chips as the demand stays off the charts.

If the next couple of years see millions of Optimus humanoid robots being produced, while fleets of autonomous vehicles hit the roads, all while AI data centers of the future continue to hog much of the chips, DRAM, optical cable, and NAND out there, perhaps there’s no reason why Nvidia can’t continue to win big, even as some customers become less dependent on Nvidia GPUs. As Nvidia takes more control of the “five-layer” AI cake, perhaps the next wave of gains in Nvidia will lie above and beyond the hardware itself. 

TPU momentum is really picking up

Of course, Alphabet stands out as an Nvidia rival that’s on the ascent. Google’s TPUs are incredibly well optimized, and if Gemini winds up being one of the few frontier models that ends up controlling market share in the distant future, perhaps it’s the model that acts as the main attraction while TPUs act as the powerful, but efficient engine underneath the hood.

Also, after Blackstone (NYSE:BX) inked a deal to start a joint venture with Google for TPU cloud access, questions linger as to just how sizable the opportunity could be as Google unleashes its custom silicon after keeping it behind closed doors for quite some time.

It’s an exciting time in the AI chip race, but it might be a bit too early in the game to write off Nvidia just because of its size and the meteoric gains that now lie behind it. Even as Google and Nvidia clash, I see no reason why both firms can’t win, given there’s no shortage of AI demand out there.

What about the rest of the Mag Seven?

Given how fast things are moving in AI and chips, it’s hard to overlook the rest of the Mag Seven and some of the smaller names out there, like Cerebras (NASDAQ:CBRS), which are hungry to disrupt. Apple (NASDAQ:AAPL) has the edge with AI on the edge with its M-series and A-series custom silicon. Add the secretive Project Baltra server chip into the equation, and it’s anyone’s guess as to how Apple could disrupt the business of AI chips.

Going down the rest of the list, Meta Platforms (NASDAQ:META) has MTIA, Amazon (NASDAQ:AMZN) has Trainium and Inferentia, Microsoft (NASDAQ:MSFT) has Maia, and Tesla (NASDAQ:TSLA) is taking it a step further with Terafab. Could these Mag Seven names have an opportunity to get a piece of the glorious high-margin growth that Nvidia is enjoying? I guess it all comes down to whether Nvidia can keep making chips that are worlds better. If it can, there are still plenty of wins for the GPU titan to be had.