Snapdragon 4 Gen 6 leak suggests on-device AI is coming to budget phones

A leaked diagram suggests Qualcomm's Snapdragon 4 Gen 6 chip will bring a dedicated AI processor to budget Android phones under $300, potentially enabling on-device AI features for the first time in this price tier.

What happened

NotebookCheck obtained a confidential diagram sheet for the unannounced SM8475 chipset, which it claims is the Snapdragon 4 Gen 6 processor. According to the leak, the SoC is built on TSMC's 4nm process and uses two Kryo Gold CPU cores (Cortex-A78) and six Kryo Silver cores (Cortex-A55). It also reportedly includes an Adreno 600-series GPU and a dedicated Hexagon DSP and AI processor, a feature not available on previous Snapdragon 4 series chips.

Other leaked specifications include support for LPDDR5 RAM, UFS 3.1 storage, a Release 17 5G modem with sub-6GHz, and USB 2.0. The CPU setup is described as pedestrian for 2026, but the GPU is expected to see more significant gains, continuing from the Gen 5's roughly 77% GPU boost over its predecessor.

Why AI builders should care

For developers building AI-powered apps, the addition of a dedicated AI processor on a budget chip is a meaningful shift. If validated, the Snapdragon 4 Gen 6 could bring significant AI and machine learning upgrades to sub-$300 phones. This means on-device AI features like real-time image processing, voice recognition, and lightweight ML models could run locally without cloud offload, even on entry-level devices.

Currently, on-device AI is largely reserved for flagship and upper-midrange chips. A budget SoC with dedicated AI hardware would expand the addressable market for AI-driven applications, especially in regions where sub-$300 phones dominate.

Practical implications

If the leak holds, developers targeting budget devices may be able to run more AI tasks locally, reducing reliance on network connectivity for features like smart camera enhancements, face detection, and app usage prediction. The Hexagon DSP and AI processor could handle these workloads efficiently, even with the modest CPU cores.

For product teams, this means planning for a baseline AI capability on devices that previously lacked dedicated hardware. However, performance will likely be limited compared to flagship chips, so lightweight model optimization (quantization, pruning) will remain important.

Caveats

All details are based on a leaked diagram sheet obtained by NotebookCheck and reported by Android Authority. Qualcomm has not confirmed any specifications or availability. The CPU setup is mild, and there is no change in manufacturing process from the Gen 5. GPU and AI gains are anticipated but not quantified. Until official benchmarks or product announcements emerge, treat these claims as speculative.