Refresh
Intel’s surrendering to Nvidia: Does it matter anyway?
The amount of buzz Nvidia’s investment in Intel will generate will likely hit fever pitch within 24 hours because of Nvidia rather than Intel. After all, Intel’s adoption of AMD’s Radeon GPU for the ill-fated range of Core i5 (8305G) and Core i7 (8705G) CPUs back in 2018 barely produced a blip back then.
The Intel/Nvidia announcement came on the day that Huawei announced that it will launch a supercluster with 1-million NPU and produced a roadmap showing a server CPU with 256 cores and 512 threads. Of course, Huawei is just one of the many challengers that could jeopardize both x86 and Nvidia’s hegemony on training and inference. AI has blown the market wide open. Welcome to the jungle!
A pragmatic Intel licks its wounds
The last few quarters have been a rollercoaster ride from hell for Intel; the graph of its stock price over the past 12 months delivers a magnificent illustration of that. The Nvidia partnership, the third in less than 60 days, injects some much-needed capital and helps Intel refocus its resources on its core, x86 computing. A resurgent Intel will force AMD to juggle resources accordingly, both in consumer and business markets. It provides Nvidia with a missing piece of its data center puzzle, an x86 answer to AMD’s Instinct APUs. Helped by the US government and Nvidia, this 2026 Intel could prove too much even for AMD and Lisa Su.
AMD is going to be very, very nervous
Share prices of AMD fell more than 4% pre-market after the announcement by Intel/Nvidia of their partnership. In comparison, Nvidia’s stock rose by 3%, a $120 billion increase, more than Intel’s entire market capitalization and 24x bigger than the $5 billion investment pledged by Nvidia.
AMD has been successful in taking on Intel in the x86 enterprise market but with Nvidia’s backing, that’s going to be trickier. Intel will be Nvidia’s favored x86 CPU partner inside datacentres and elsewhere. (Nvidia of course has its own Arm-based Grace CPU). Nvidia is also likely to help Intel prop up its integrated GPU to fend off AMD’s APUs. The proverb, “The enemy of my enemy is my friend”, never sounded so true.
A couple of notes from Desire Athow at TechRadar Pro:
I think that the White House, at the highest level, encouraged Nvidia to assist Intel. Let’s be in no doubt here. Intel needs Nvidia more than Nvidia needs Intel – its market capitalization, $4.14 trillion, is almost 40x that of its prestigious Santa Clara-based peer. The US Government took a 10% stake – worth almost $9 billion – in the US chipmaker back in August just days after Softbank purchased $2 billion in shares, making it back then the 6th largest shareholder with just under 2% of Intel share base. Within weeks therefore, almost a fifth of Intel’s stock has changed hands. The company, famous for its jingles and blue men ads, is seemingly too big to fail.
It’s also funny to see Jensen wearing a suit rather than his customary leather jacket. This is definitely an investor call.
Intel’s Tan doesn’t rule out Nvidia’s using Intel Foundry for this partnership.
Nvidia’s investment in Intel, Huang says, is a major vote of confidence in the future of Intel. Jensen also mentioned earlier that the Trump administration wasn’t involved in this partnership, but that the administration is very supportive.
Neither will say which process node — or which foundry — the new chips will be fabbed on. However, this will be a tens of billions of dollar product partnership between the two companies, which is definitely a lift for Intel.
Another question about whether Nvidia is going to lean into Intel’s foundry business. Both Huang and Intel CEO Lip-Bu Tan praise TSMC’s foundry business. But this announcement is going to stay focused on datacenter and client CPU integration.
Cramer follows up with a question about Intel Foundry services and whether Nvidia will move over to Intel from TSMC. Huang is non-commital, but given the amount of investment involved, Huang says that Nvidia will be a major customer of Intel going forward.
Take that as you will, but if Intel 18A is as good and viable as Intel claims it to be, I can’t see how Nvidia won’t get first crack at those nodes.
OK, so Jim Cramer at CNBC is asking for the business/market impact of this announcement. Nvidia CEO Jensen Huang brings up the integrated RTX GPU chiplet into Intel SoCs, which will be very competitive with AMD’s APUs found in thin and light gaming laptops as well as mini-PCs and gaming handhelds.
Hey folks, this is John Loeffler, US Computing Editor, here to bring you all the news and reaction I can muster to the Nvidia x Intel press conference this afternoon. We’ve also got some insights from my colleagues at TechRadar Pro regarding the new data center x86 products, which I’ll be bringing to you shortly.
We’re about to start, so stay tuned for the latest as it happens.
