NXP Semiconductors (NXPI) Technology Leadership Forum 2026 summary

Market environment and inventory trends

• Market conditions have improved significantly over the past year, with book-to-bill ratios above one across all end markets and distribution inventory back to target levels at 11 weeks.

• Some large tier 1 automotive customers are holding low inventory levels, relying on suppliers to manage stock, which could lead to future supply challenges if demand spikes.

• Inventory is currently at 156 days, above the 110-day target, with plans to reduce it by year-end; no intention to hold inventory on behalf of customers due to the build-to-order model.

• Lead times are extending in certain areas, providing better visibility, but forecasting varies widely by customer.

• Distribution channel health is strong, with detailed visibility into customer backlogs and shipments.

Automotive and regional dynamics

• Automotive business turned the corner after Q3 last year, with all geographies and product categories up in the most recent quarter.

• Chinese automakers are leading in EVs, with domestic sales offset by exports, while European automakers face challenges but maintain strength in luxury segments.

• Growth drivers in auto include SDVs, radar, and BMS, with early-stage customer engagements for next-gen products like S32N and S32K5, expected to generate revenue from 2027 onward.

• AI interest in automotive is emerging, particularly for in-cabin applications using distilled large language models, though revenue impact is still a few years out.

Physical AI and software opportunities

• Physical AI is a major focus, with both industrial and automotive applications; industrial opportunities have a shorter revenue cycle than automotive.

• Kinara NPU offers high performance (40 TOPS) and is seeing rapid growth in the opportunity pipeline, now at $1.5 billion.

• Software monetization is advancing in automotive through the CoreRide platform and MotionWise middleware, with ongoing investment in model distillation tools for industrial markets.

• Most industrial AI models are third-party, not developed in-house.