Nvidia is throwing beleaguered rival Intel a lifeline in a surprising partnership with significant ramifications across the semiconductor industry. The deal announced Thursday has two main parts: Nvidia will purchase $5 billion into Intel common stock, joining Japanese conglomerate Softbank and the U.S. government as new investors in the iconic semiconductor firm attempting a revival under CEO Lip-Bu Tan. But, crucially, they’re cozying up on the product side for both data centers, where Nvidia dominates the market for high-powered AI chips known as GPUs, and personal computers, where Intel maintains a strong foothold with its traditional computer processors called CPUs. Shares of Club name Nvidia rose nearly 3% Thursday, while Intel’s stock soared more than 25%. The disproportionate market reaction makes sense, considering the scale of Intel’s problems — both financially and fundamentally — means this partnership improves its future prospects a lot more than it does for Nvidia. Still, count this as an Nvidia win, too. “What this does is cement Nvidia as being the company that has, basically, a soup-to-nuts offering. You go all the way from the PC to the highest end,” Jim Cramer said Thursday morning. “We were all talking about Nvidia and China yesterday,” Jim added, in reference to reports Wednesday that the Chinese government was banning companies from using Nvidia’s AI chips. “Now we should be talking about Nvidia as being the colossus.” NVDA INTC 5Y mountain Nvidia’s five-year stock performance versus Intel. Data centers For data centers, Intel will develop custom CPUs to pair with Nvidia’s AI infrastructure products anchored by its cutting-edge GPUs. Nvidia does make its own CPU for those products already, but they take advantage of a different architecture than the CPUs that Intel has long sold into data centers. Nvidia has been using the Arm instruction set to build CPUs, while Intel and Advanced Micro Devices use what’s called x86. Historically, x86 is the go-to architecture for data centers CPUs and they still have a lot of market share. Arm-based CPUs are the newcomer, with Amazon Web Services’ in-house Graviton chips and Nvidia’s Grace CPUs leading the charge. In simple terms, Arm is seen as a simpler, more power-efficient architecture than x86, which is why it’s been used widely in chips for mobile phones. By contrast, x86 is viewed as better equipped for tasks where performance is the priority. No doubt, performance matters a great deal for AI computing. Nvidia is still planning to develop Arm-based CPUs despite the Intel partnership, Nvidia CEO Jensen Huang said on a Thursday afternoon call with reporters. For example, Nvidia was already readying its second-generation Arm-based CPU for release next year as part of its Vera Rubin AI platform. Huang said that Vera CPU is still full-systems go, and he added there are other Arm-based products that Nvidia is developing to serve markets such as robotics. Instead, he said, the Intel partnership will allow Nvidia to better serve customers who are entrenched in the x86 ecosystem as a result of tighter integration between Nvidia’s GPUs and Intel’s CPUs inside data center server racks. Analysts and investors are worried about the implications for Arm Holdings , which licenses the Arm instruction set to customers. The stock is down more than 3% on Thursday afternoon, even after Huang’s comments. Time will tell on what, exactly, it means for Arm. But it is fair to argue that, in the future, some of the money that may have gone to Arm for Arm-based CPUs will now go to Intel for x86-based CPUs. Nvidia once tried to buy Arm Holdings, but that deal was thwarted by regulators. Jim said this arrangement with Intel is better for Nvidia than outright owning Arm. Wall Street also is trying to figure out what the partnership could mean for AMD, which is both trying to compete with Nvidia in the booming AI chip market and, in recent years, has successfully grabbed a lot of share from Intel in the x86 CPU market. AMD’s stock is down about 2% Thursday. Personal computers On the other side of the deal, Intel will integrate Nvidia’s graphics cards into system-on-chips, or SoCs, that Intel designs for personal computers. These SoCs contain many of the key ingredients for computers — including the traditional processor, the graphics processor and memory — all onto a single piece of silicon. These Intel-plus-Nvidia SOCs will power “a wide range of PCs that demand integration of world-class CPUs and GPUs,” Nvidia said in a press release. Nvidia already has a presence in the PC market, but it’s primarily concentrated in higher-end products targeted toward gamers and workstations used by professionals — something Huang pointed out on Thursday’s call with reporters. This arrangement could lead to Nvidia growing its presence in the broader consumer PC market, where Intel is still well-regarded. There are “a 150 million laptops sold each year and we’re now going to combine the best CPU and the best GPU, so it’s really, really exciting,” Huang said. While the scale of the data center AI market dwarfs PCs, that doesn’t mean this isn’t worth pursuing for Nvidia. “I’m not saying Jensen was desperate for PC, but it just fell into his lap,” Jim said. In a note to clients, analysts at Wolfe Research makes Intel more competitive in PCs and therefore is “marginally negative” for AMD, which in 2024 derived a little over a quarter of its revenue from its PC segment; about half came from data center. In Nvidia’s most recent fiscal year, for comparison, almost 90% of its sales were from its data center segment. Could there be more? The biggest question hanging over the entire Intel-Nvidia deal on Thursday: Will Nvidia end up using Intel’s third-party manufacturing business — dubbed Intel Foundry — to make any of its products? The answer, according to Huang, is not right now. “We’ve always evaluated Intel’s foundry technology and we’re going to continue to do that,” Huang said on Thursday’s media call, in response to a question from Jim. “But today, this announcement is squarely focused on these custom CPUs. With this partnership, we’re essentially going to be a major customer of Intel server CPUs. This is the very first time.” Nvidia designs its chips then relies other companies, primarily Taiwan Semiconductor Manufacturing Co., to produce them. TMSC is the world’s dominant producer of the most advanced semiconductors, typically called “leading edge” chips. Compared with Nvidia and AMD, Intel stands out for both designing chips and manufacturing them. However, its leading-edge manufacturing prowess has fallen behind the likes of TSMC and South Korea’s Samsung Foundry. Getting Intel Foundry back on track has been both a strategic priority for Intel’s management team and the U.S. government, which, for national security purposes, wants an American company to be able to produce advanced chips on American soil. To date, though, one of the big challenges has been convincing third-party companies that compete with Intel on chip design to utilize tits manufacturing services. That is why there is so much focus on the question of whether this Nvidia partnership could, eventually, lead Nvidia to become a customer of Intel Foundry. Indeed, even after Huang answered his first question on Intel Foundry, he fielded multiple additional questions on the topic. Huang would go on to praise TSMC’s leadership — “You just can’t overstate the magic that is TSMC,” he said — before steering the conversation back toward the CPU partnership. As for whether Nvidia’s decision to forge closer ties with Intel was politically motivated, Huang said the Trump administration had “no involvement in this partnership at all.” Nvidia expects a “fantastic” return on its $5 billion investment, Huang said. (Jim Cramer’s Charitable Trust is long NVDA. See here for a full list of the stocks.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
