Snapdragon 8 Elite Gen 6 Pro may grow on 2nm

Documents reportedly detailing Qualcomm’s next flagship mobile processor, internally identified as SM8975, have surfaced online. The chip is expected to use TSMC’s N2P 2nm-class process, yet its die could be larger than the previous generation’s: approximately 134 mm², compared with 126.2 mm² for its predecessor.

That expansion suggests Qualcomm may be adding cache, logic, or dedicated accelerators rather than simply shrinking the existing design. The move would give the company another way to separate its premium Pro model from the standard version, alongside clock speeds and GPU capabilities.

Snapdragon 8 Elite Gen 6 Pro specifications

The leaked documents reportedly point to the following configuration:

• Process: TSMC N2P, 2nm-class
• Die size: approximately 134 mm²
• CPU: Oryon cores in a 2+3+3 arrangement
• GPU: Adreno 850 with 18 MB GMEM
• Memory: LPDDR5X and LPDDR6 support
• Cache: 16 MB of shared L2 cache and 8 MB of system cache

The CPU would retain Qualcomm’s Oryon architecture, with two large cores, three performance cores, and three efficiency cores. The leak does not reveal clock speeds, leaving that detail for Qualcomm’s expected autumn presentation.

Recommended reading

Huawei Nova 16 SE may pack an 8,500mAh battery

The graphics configuration is more distinctive. The Pro model is said to use an Adreno 850 with 18 MB of GMEM, while the standard SM8975 could receive a simpler Adreno 845 with 12 MB of graphics memory.

AI Frame Fusion and LPDDR6 support

The Pro version reportedly adds two dedicated Matrix ALU units linked to Qualcomm’s AI Frame Fusion feature. According to the leaked description, the technology is intended to support both upscaling and intermediate-frame generation, helping games appear smoother without consuming battery as quickly.

The leak suggests these Matrix ALU units and AI Frame Fusion will remain exclusive to the Pro model. The standard chip is expected to omit them, continuing Qualcomm’s familiar split between a higher-priced flagship configuration and a more widely deployed version.

Memory support is also expected to differ. The Pro chip could support both LPDDR5X and the newer LPDDR6, while the standard model would remain limited to LPDDR5X.

A larger die is not necessarily unusual when a mobile processor is being redesigned for heavier graphics and on-device AI workloads. In Qualcomm’s case, the extra silicon may also help distinguish the Pro model without relying solely on higher frequencies. The company is competing with MediaTek’s Dimensity chips in the premium Android market, where phone makers weigh peak performance against thermal limits, battery life, and local AI capabilities.

Qualcomm has scheduled its Snapdragon Summit in Maui for September 22–24, 2026. The event is where the company would typically disclose final clock speeds, the complete lineup, and commercial names.