A fresh leak is stirring up excitement around Intel’s next big move in high-performance APUs, and it could mark one of the most surprising partnerships in years. According to the information shared, Intel’s upcoming Serpent Lake chips may arrive with a major twist: an integrated Nvidia GeForce RTX-class GPU chiplet designed to go head-to-head with AMD’s Strix Halo and future Zen 6 “Halo”-style APUs.
What makes Serpent Lake so interesting is the rumored split of responsibilities inside the package. On the CPU side, Intel is said to lean on its Titan Lake architecture, featuring upgraded Griffin Cove performance cores (P-cores) alongside Golden Eagle efficiency cores (E-cores). There’s also talk that the SoC tile, which includes low-power efficiency cores (LP E-cores), may use Titan Lake-derived technology as well—suggesting Intel is aiming for a strong blend of performance, efficiency, and platform-level integration.
The bigger headline, though, is the graphics approach. Instead of relying solely on Intel’s own integrated GPU designs, Serpent Lake is rumored to include an Nvidia GPU chiplet built on the RTX Rubin architecture (or something closely related). If accurate, that would be a major shift for integrated graphics—and a direct attempt to challenge the growing market for “gaming-capable” and creator-focused APUs that don’t need a separate discrete graphics card.
Manufacturing details also hint at why Intel might pursue this route. The integrated RTX GPU chiplet is said to be built on TSMC’s N3P process node, a modern 3nm-class technology that could enable higher performance and improved efficiency—two key factors for thin-and-light laptops, compact desktops, and high-end mobile workstations where power and heat constraints matter.
Memory support could be another pillar of the Serpent Lake strategy. The leak suggests these APUs may support up to 16X LPDDR6, which would be a meaningful upgrade for bandwidth-sensitive workloads such as gaming, 3D rendering, AI tasks, and content creation. Faster memory is often one of the most important ingredients for integrated graphics performance, so pairing next-gen LPDDR6 with an RTX-derived iGPU could be a deliberate move to close the gap with lower-end discrete GPUs.
There’s also an intriguing connection to an older rumor: Serpent Lake may effectively be a reworked version of a previously discussed concept sometimes referred to as “Nova Lake-AX.” That earlier idea was framed as an answer to AMD’s Strix Halo, featuring a CPU configuration around 8 P-cores and 16 E-cores paired with a large Intel iGPU. The new twist is that Serpent Lake could keep a similar big-APU blueprint, but swap in Titan Lake CPU cores and replace Intel’s integrated graphics with an Nvidia RTX-based chiplet instead.
If these details hold up, Intel Serpent Lake could become one of the most ambitious APU designs the PC space has seen in years: next-generation Intel CPU cores, an Nvidia RTX-class integrated GPU chiplet, advanced TSMC process technology, and LPDDR6 memory support—all targeting the same performance territory that AMD’s halo-class APUs are pushing into.
For now, it’s still leak territory, and key specifics like final core counts, exact GPU specifications, and real-world performance targets remain unknown. But the overall direction is clear: Intel appears to be gearing up for a serious fight in the premium APU market, and partnering with Nvidia for integrated RTX graphics would be a bold way to do it.
