Intel’s 15th Gen Arrow Lake processors will be a true “next-gen” lineup with a new CPU, GPU, and an Angsltrom-class 20A (~2nm) process node. The high-performance H-series mobility lineup will allegedly pack up to 14 cores (6P + 8E) and a GT2 iGPU based on the Battlemage graphics architecture (successor to Alchemist used by Arc A-series GPU). According to previous rumors, the graphics tile on Arrow Lake-H will feature a massive 320 EU iGPU with 2,560 shaders.
Meteor Lake is expected to feature up to 128 EUs based on the Xe-LPG architecture. Going from 128 to 320 within one generation is unlikely, especially if the process node is retained. Intel’s original slides indicate the use of the TSMC N3 process for the Arrow Lake iGPU tile, but plans may have changed, and by changed I mean downgraded.
Arrow Lake will feature a heterogeneous CPU consisting of six P-cores based on the Lion Cove core architecture and eight E-cores leveraging the Skymont architecture. These updates are expected to offer substantial performance uplifts over the Redwood Cove and Crestmont cores utilized by Meteor Lake. Read more on that here:
The CPU tile on Arrow Lake will be fabbed on the Intel 20A 2nm-class process, making it one of the most advanced in the industry. With RibbonFET (GAA) transistors and PowerVia backside delivery, it is expected to counter TSMC’s N2 (2nm) process nodes at every step.
The iGPU tile will be fabbed on TSMC’s N3 (3nm) or N4 (4nm) process node. The rumor mill claims that the 15th Gen Arrow Lake mobile processors will pack up to 320 EUs or 2,560 cores. That’s 2.5x more than Meteor Lake. It’s not entirely impossible, just unlikely. I’d bet on a more reasonable 192 EU part with 1,536 shaders. Either way, Intel’s targeting RTX 3050/RTX 4050 mobile performance. They are the crux of the “gaming laptop market,” accounting for most mobile GPU sales.