NVIDIA’s next-generation “AI PC” SOC is said to feature Arm-based Blackhawk processors, Blackwell RTX GPUs, and LPDDR6 memory.
NVIDIA’s next-gen SOC could shake things up in the AI ​​computing segment if these rumored specs are correct, Blackhawk Arm CPU cores, Blackwell RTX GPUs and LPDDR6 memory all on board in the same package
We’ve seen several rumors and reports recently that NVIDIA is potentially planning to enter the client PC space with its own SOC. There are reports that these SOCs will also open the door to new consoles, primarily handhelds at first. Just yesterday, the CEOs of NVIDIA and Dell teased the venture, and it looks like we have a new rumor that could give us an early hint at what these SOCs would look like.
https://x.com/XpeaGPU/status/1793469879334687122
The details come from @XpeaGPU, who we quoted in a previous article talking about a potential handheld and console SOC designed by NVIDIA based on the same architecture. It looks like NVIDIA is trying to incorporate several next-gen IPs into one package, which seems to be the way to go when it comes to modern PC designs. We’ve seen it with Apple and now with Intel’s Lunar Lake, which offers multiple IP addresses on the same memory die in a package, creating a very efficient and powerful solution in a small footprint.
https://x.com/Kepler_L2/status/1793523940968820899
For its next generation SOC, NVIDIA is reportedly looking to use a 3nm process node, reportedly from either Intel or TSMC, and an advanced packaging solution that will combine Arm-based Cortex X5 BlackHawk CPU cores and Blackwell RTX GPU cores. It is likely that NVIDIA will use BlackHawk CPU cores as they are reportedly working with MediaTek to create these SOCs.
MediaTek is said to be using the same Arm architecture for its upcoming Dimensity 9400 chip, and will reportedly take on Apple’s A17 Pro architecture and Qualcomm’s Oryon architecture.
The use of a Blackwell RTX GPU is also not that far-fetched, as MediaTek previously confirmed in a press release that it will be using the latest RTX & AI GPU IP within its Dimensity SOC, so we might get a modified version of it. designed and further optimized by NVIDIA. NVIDIA “AI PC” SOCs are also said to include LPDDR6 memory as an in-package design. LPDDR6 will have a speed of 10,667 MT/s from scratch and would reach up to 14,400 MT/s during its life cycle, as recently revealed.
Another indication of this development comes directly from NVIDIA CEO Jensen Huang, who says that the recent introduction of Microsoft’s Copilot+ PCs has paved the way for opportunities in the client segment. Jensen says that Copilot+ PCs are great at supporting ARM, and Qualcomm’s Snapdragon X processors are proof of that, giving NVIDIA an incentive to try and do something even better in this segment.
Today, as you saw in yesterday’s build, which I thought was really excellent, Satya announced a next-generation PC, the Copilot+ PC, which runs fantastically on the NVIDIA RTX GPUs that come in notebooks. But it also supports ARM beautifully. And so it opens up opportunities for system and PC innovations.
Jensen Huang – NVIDIA CEO (Q1 2024 Earnings Pick)
Once again, it’s too early to tell what the final design and specifications will look like for NVIDIA’s “AI PC” SOC for client PCs, but the company has a proven track record of incorporating high-end GPU architectures alongside ARM processors for data centers (Grace), so we’re looking forward to , until NVIDIA brings the same combo on the client side.