6.1 Hardware product iteration

6.1.1 Device matrix expansion direction

Product Line	Core positioning	Technology exploration direction
N series	Main flagship	3D light field display optimization/DePIN computing power sharing
Lite series	Emerging market entry models	Simplify sensor configuration/lower the threshold for distributed participation
Pro series	Professional creator tools	Integrated LiDAR/support 4K 3D shooting
X series	Concept technology verification platform	Concept technology verification platform

6.1.2 Technical pre-research list

a. Computing power network protocol: Design an asynchronous computing framework for mobile phone clusters to allow idle chips (such as NPUs during night charging) to participate in distributed training tasks (2026 laboratory verification).

b. Security evolution: Explore the lightweight implementation of anti-quantum encryption algorithms on mobile terminals to deal with computing power threats after 2030 (long-term tracking).

c. Research on lightweight model architecture: Develop a dynamic sparse model architecture that is adapted to mobile hardware, and achieve 50% reduction in parameter volume by 2026 while maintaining 90% of the original model performance.

d. Multimodal federated learning protocol: Design cross-device and cross-modal federated learning protocols to support joint training of mobile phone sensor data (such as gyroscopes, light sensing) and text/images.

e. Crowdsourcing data quality game mechanism: building an incentive compatibility system based on blockchain.

f. Edge-end real-time cleaning engine: use mobile phone NPU to accelerate the cleaning process.

g. Fragmented data transaction agreement: Determine the minimum transaction unit (including metadata hashing and terms of use).

h. Dynamic pricing model: Combining oracle price feeding and historical trading model (similar to Uniswap V3 centralized liquidity).

i. Spatial computing middleware: Dynamic physics engine supports real-time simulation of ten thousand-level rigid bodies.