The global semiconductor industry is undergoing a quiet yet profound paradigm shift—from pure-play logic chip competition toward heterogeneous integration centered on novel displays and power devices. Against this backdrop, West Lake Yanshan Technology’s recent launch of an 8-inch GaN-on-silicon Micro LED IDM (Integrated Device Manufacturer) production line in Deqing, Zhejiang, represents more than a technical milestone; it signals China’s strategic pivot to bypass traditional CMOS foundry ecosystems and build a new semiconductor architecture anchored in vertical integration.
Claimed as the world’s first 8-inch GaN-on-Si Micro LED full-chain IDM line, the facility spans epitaxial growth, chip fabrication, and advanced packaging—enabling closed-loop production from silicon-based GaN wafers to finished Micro LED devices. Compared to conventional 4-inch sapphire substrates, the 8-inch silicon platform reportedly improves light extraction efficiency by over 15% while significantly reducing cost per unit area. Industry estimates suggest GaN-on-Si Micro LED wafers could be 30–40% cheaper than sapphire-based alternatives—a critical advantage for high-resolution AR/VR, automotive displays, and ultra-HD large screens.
Yet the real significance lies not in specs but in strategic alignment. Though a startup, West Lake Yanshan is backed by Haiwang Capital—a vehicle linked to China’s National Integrated Circuit Industry Investment Fund (Big Fund II)—and operates in tacit synergy with BOE and BYD. BOE, the world’s largest LCD panel maker, has accelerated its Micro LED and OLED investments, with pilot lines entering customer validation in 2025. BYD, through its semiconductor arm FinDreams, is rapidly scaling automotive-grade GaN power modules; its GaN-based inverters already appear in premium EV models as of 2024. While no formal alliance exists, their supply chains, equipment validation protocols, and application scenarios are deeply complementary.
This “display + power + mobility” triangle is forging a new IDM model—one that transcends single-device manufacturing and instead uses end applications as the anchor for backward integration across materials, epitaxy, chips, and packaging. This stands in stark contrast to TSMC’s foundry-centric approach, which relies on standardized process nodes serving diverse clients. The Chinese model sacrifices universality for system-level optimization. In my view, this approach holds distinct advantages within China’s industrial context: clear policy support, massive domestic demand, and limited immediate reliance on cutting-edge logic nodes.
Geopolitics further amplifies this divergence. U.S. semiconductor export controls continue tightening, especially on advanced lithography and EDA tools. But Micro LED and GaN power devices largely operate on mature or semi-mature nodes (e.g., 180nm to 90nm), where domestic equipment penetration is relatively high. Chinese toolmakers like Naura and Advanced Micro-Fabrication Equipment (AMEC) already supply key MOCVD and etch systems. This enables China to construct a largely “de-Americanized” production line without needing ASML’s EUV or Applied Materials’ most advanced PVD tools.
Challenges remain formidable. Mass transfer—the process of relocating millions of micro-scale LEDs onto backplanes—still suffers from yield bottlenecks. Industry-wide yields hover below 60%, far short of the 85% threshold for commercial viability. West Lake Yanshan claims breakthroughs in laser lift-off and self-aligned bonding, but independent verification is lacking. Moreover, IDM models demand sustained capital intensity at a time when investor appetite favors AI accelerators and HBM memory. Global Micro LED funding fell 22% year-over-year in 2025, reflecting this bias.
Here, BOE and BYD may prove decisive. BOE brings display backplane integration expertise; BYD contributes automotive qualification channels and co-design experience in power electronics. If both integrate West Lake Yanshan into their supply ecosystems, they can share R&D risk and accelerate iteration cycles. I predict that within two years, China will launch its first mass-produced automotive HUD or AR-HUD system based on GaN-on-Si Micro LED—co-developed by this triad.
The implications extend far beyond displays. This IDM experiment embodies an alternative semiconductor development path: one that sidesteps Moore’s Law extremes in favor of heterogeneous integration and system-level efficiency, and replaces globalized fragmentation with controlled vertical ecosystems. As supply chains fracture worldwide, such “localized closure” strategies may become templates for others.
But a critical question lingers: while the world pours capital into AI compute, can China’s bet on this non-mainstream track demonstrate commercial viability within five years? Or will it become another technology island driven more by geopolitical anxiety than market logic? The answer will shape who controls the next interface between humans and machines.