On-Device AI Silicon: Apple, Qualcomm, MediaTek NPUs — 2026-10-24
The global semiconductor market faces a severe contraction as memory shortages and rising RAM prices drove smartphone shipments down 22% in Q3 2026, while PC shipments fell 21%. Despite this downturn, the battle for on-device AI supremacy intensifies, with Qualcomm’s new Snapdragon 8 Elite Gen 6 achieving significant NPU efficiency gains and Apple, Qualcomm, and MediaTek all pivoting to TSMC’s 2nm process to run 30-billion-parameter models locally.
On-Device AI Silicon: Apple, Qualcomm, MediaTek NPUs — 2026-10-24
Top developments
Qualcomm’s Snapdragon 8 Elite Gen 6 delivers major NPU efficiency gains Counterpoint Research reported on October 14, 2026, that the new Snapdragon 8 Elite Gen 6’s Hexagon NPU runs 35% faster than its predecessor and is 33% more efficient per watt for AI workloads. This chip is designed to support on-device execution of 30-billion-parameter Mixture-of-Experts (MoE) models, marking a shift toward heavier local AI computations.
Smartphone and PC shipments plummet due to memory shortages IDC data indicates that global smartphone shipments fell 22% to 255.3 million units in Q3 2026, a record drop attributed to soaring memory costs. Similarly, Counterpoint Research notes that global PC shipments dropped 21% in the same quarter due to RAM price hikes. These supply-side pressures are impacting the adoption rate of new AI-capable devices despite strong technical advancements in NPUs.
2nm process battle defines the next generation of mobile AI Qualcomm, Apple, and MediaTek are all transitioning their flagship chips to TSMC’s 2nm process to handle the thermal and power demands of on-device AI. While specific 2026 shipment shares for these new chips are not yet finalized, the consensus is that the 2nm node is critical for fitting large language models into mobile form factors without overheating.
Local view
Chinese media highlights the shift from cloud to edge AI Weibo discussions and financial analyses emphasize that the "edge-side AI" (端侧AI) trend is accelerating, with users recognizing that running large models locally on devices like smartphones and glasses offers better privacy and faster response times than cloud-dependent solutions. The narrative suggests that while the market is cooling, the technological imperative for local AI remains the primary driver for future hardware cycles.
Context & numbers
- Smartphone Shipments: Global smartphone shipments fell 22% in Q3 2026 to 255.3 million units, with IDC forecasting no recovery before 2028 due to persistent memory cost pressures.
- PC Market Contraction: Global PC shipments declined 21% in Q3 2026, driven by significant increases in RAM and component prices.
- AI PC Penetration: Counterpoint Research projects that AI Advanced PCs will account for approximately 59% of global shipments in 2026, a sharp increase from 39% in 2025, indicating that despite volume drops, the mix of devices is shifting heavily toward NPU-equipped hardware.
On the radar
- Chip Stock Volatility: Major semiconductor stocks including AMD, Intel, and Marvell saw significant declines (up to 8%) in late October 2026 following warnings from chipmaking equipment suppliers about supply chain bottlenecks affecting AI data center spending.
- Memory Shortage Duration: Industry analysts warn that the current memory shortage, which has crippled smartphone unit growth, is expected to persist well into 2027, potentially delaying the full market saturation of new AI smartphone models.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
