China’s Optical Supply Chain Powers AI Scaling

Read original on Pandaily

#optical-transport#16t-optics#in-p-substratesai-optical-transport-infrastructurelightcountingchinasilicon-photonicscpo

AI clusters need more than GPUs—China’s optical supply chain may determine how quickly they can scale.

30-Second TL;DR

What Changed

Seven Chinese vendors rank among LightCounting’s global top ten optical-module suppliers.

Why It Matters

AI infrastructure builders may face fewer constraints from compute chips alone and increasingly need to manage optical-network availability. Supplier concentration in China also makes geopolitical exposure and export controls important factors in data-center planning.

What To Do Next

Ask your infrastructure team to map 1.6T optical-module dependencies and qualify at least one alternative supplier for each critical link.

Who should care:Enterprise & Security Teams

Key Points

•Seven Chinese vendors rank among LightCounting’s global top ten optical-module suppliers.
•The 1.6T upgrade cycle is shifting attention beyond pluggable modules to complete optical transport systems.
•Silicon photonics, InP substrates, and co-packaged optics are becoming strategic parts of the AI networking stack.
•China’s supply-chain depth may influence the cost, availability, and deployment speed of AI data-center networks.

Deep Insight

AI-generated analysis for this event — not the original article.

Enhanced Key Takeaways

•Chinese optical manufacturers are increasingly vertically integrating by producing their own laser chips and TOSA/ROSA components to mitigate geopolitical supply chain risks.
•The transition to 1.6T modules is driving a shift toward EML (Electro-absorption Modulated Laser) and thin-film lithium niobate (TFLN) technologies to manage power consumption at higher data rates.
•Major Chinese suppliers are aggressively expanding manufacturing footprints in Southeast Asia, specifically Vietnam and Thailand, to circumvent US trade restrictions and maintain access to Western hyperscaler markets.
•The integration of CPO (Co-Packaged Optics) is being prioritized by Chinese firms to reduce the 'energy-per-bit' bottleneck, which is currently the primary constraint for scaling GPU clusters beyond 100,000 units.
•Chinese optical firms are heavily investing in AI-driven automated assembly and testing lines to improve yield rates for complex 1.6T and 3.2T optical engines.

Competitor Analysis

Cost Structure

Chinese Optical Vendors: Highly Optimized/Aggressive
Western Optical Vendors (e.g., Coherent, Lumentum): Premium/High-Margin
Silicon Photonics Maturity: Rapidly Scaling

Supply Chain

Chinese Optical Vendors: Vertically Integrated
Western Optical Vendors (e.g., Coherent, Lumentum): Outsourced/Hybrid
Silicon Photonics Maturity: R&D Focused

1.6T Readiness

Chinese Optical Vendors: High (Volume Production)
Western Optical Vendors (e.g., Coherent, Lumentum): High (Volume Production)
Silicon Photonics Maturity: Advanced

Market Focus

Chinese Optical Vendors: Hyperscale Data Centers
Western Optical Vendors (e.g., Coherent, Lumentum): Telecom & Enterprise
Silicon Photonics Maturity: Data Center Interconnect

Feature	Chinese Optical Vendors	Western Optical Vendors (e.g., Coherent, Lumentum)	Silicon Photonics Maturity
Cost Structure	Highly Optimized/Aggressive	Premium/High-Margin	Rapidly Scaling
Supply Chain	Vertically Integrated	Outsourced/Hybrid	R&D Focused
1.6T Readiness	High (Volume Production)	High (Volume Production)	Advanced
Market Focus	Hyperscale Data Centers	Telecom & Enterprise	Data Center Interconnect

Technical Deep Dive

1.6T modules utilize 8x200G PAM4 signaling, requiring significant improvements in DSP (Digital Signal Processor) power efficiency and thermal management.
Silicon Photonics (SiPh) platforms are moving toward monolithic integration, combining lasers, modulators, and detectors on a single die to reduce coupling losses.
InP (Indium Phosphide) substrates remain the gold standard for high-speed lasers due to superior gain characteristics compared to GaAs or Silicon-only solutions.
CPO architectures move the optical engine closer to the ASIC/GPU, reducing the electrical trace length and significantly lowering power consumption compared to traditional pluggable optics.

Future ImplicationsAI analysis grounded in cited sources

Chinese optical firms will capture over 60% of the global 1.6T module market by 2027.

Aggressive pricing strategies combined with rapid manufacturing scale-up in Southeast Asia provide a significant competitive advantage over Western incumbents.

CPO adoption will become the standard for AI clusters exceeding 100,000 GPUs by 2028.

The power-per-bit limitations of traditional pluggable optics make them unsustainable for the thermal envelopes of next-generation AI supercomputers.

Timeline

2022-05

LightCounting reports Chinese firms capturing significant market share in 400G/800G optical modules.

2023-11

Major Chinese optical vendors announce mass production capabilities for 800G modules.

2024-09

Industry shift toward 1.6T standards accelerates, with Chinese firms unveiling prototypes at major trade shows.

2025-06

Chinese optical suppliers begin large-scale deployment of Southeast Asian manufacturing facilities to bypass trade barriers.

2026-03

First commercial shipments of 1.6T optical modules from top-tier Chinese suppliers to global hyperscalers.

2022-05
LightCounting reports Chinese firms capturing significant market share in 400G/800G optical modules.
2023-11
Major Chinese optical vendors announce mass production capabilities for 800G modules.
2024-09
Industry shift toward 1.6T standards accelerates, with Chinese firms unveiling prototypes at major trade shows.
2025-06
Chinese optical suppliers begin large-scale deployment of Southeast Asian manufacturing facilities to bypass trade barriers.
2026-03
First commercial shipments of 1.6T optical modules from top-tier Chinese suppliers to global hyperscalers.