AI's Next Bottleneck: Indium Phosphide
💡InP prices are surging and lead times stretch to 40 weeks—the hidden constraint behind faster AI networking.
⚡ 30-Second TL;DR
What Changed
Two-inch optical-communications indium phosphide substrates rose from about $800 to $2,300–$2,500 in 18 months, with urgent spot orders above $3,000.
Why It Matters
The shortage could raise the cost and delay the deployment of 800G, 1.6T, and faster optical links that connect AI accelerators. AI infrastructure companies may need to secure materials earlier and diversify suppliers rather than treating photonic components as a routine procurement item.
What To Do Next
Audit your 800G/1.6T optical-module bill of materials and request 2027 capacity commitments from at least two qualified InP suppliers before expanding AI clusters.
Key Points
- •Two-inch optical-communications indium phosphide substrates rose from about $800 to $2,300–$2,500 in 18 months, with urgent spot orders above $3,000.
- •Fourth-quarter substrate and epitaxial-wafer prices may rise by more than 10%, while some suppliers are booked through 2027 or 2028.
- •Nvidia invested $2 billion each in Coherent and Lumentum and is using long-term purchasing and capacity agreements.
- •Estimated 2026 demand is 2.6–3.0 million two-inch-equivalent wafers versus only about 750,000 units of effective compliant capacity.
- •Indium supply depends on lead, zinc, and tin mining byproducts, while crystal growth, qualification, and supplier concentration make rapid expansion difficult.
🧠 Deep Insight
AI-generated analysis for this event.
🔑 Enhanced Key Takeaways
- •Indium Phosphide (InP) is critical for 800G and 1.6T optical transceivers because its direct bandgap allows for high-speed modulation and lower signal attenuation compared to Silicon Photonics in long-reach applications.
- •The supply chain bottleneck is exacerbated by the 'wafer breakage' rate in InP manufacturing, which is significantly higher than Silicon due to the material's inherent brittleness and lower thermal conductivity.
- •Geopolitical export controls on Gallium and Germanium, implemented by China in 2023, have created a 'scarcity mindset' among InP manufacturers, leading to aggressive inventory stockpiling that further inflates spot prices.
- •Major telecommunications and data center operators are shifting toward 'Thin-Film Lithium Niobate' (TFLN) as a potential alternative to InP for specific high-bandwidth modulators to mitigate long-term supply chain risk.
- •The transition from 4-inch to 6-inch InP wafer production is currently stalled because the crystal growth process for larger diameters suffers from high defect densities, limiting the yield of high-performance laser diodes.
📊 Competitor Analysis▸ Show
| Feature | Indium Phosphide (InP) | Silicon Photonics (SiPh) | Thin-Film Lithium Niobate (TFLN) |
|---|---|---|---|
| Primary Use | High-speed Lasers/Modulators | Integrated Transceivers | High-bandwidth Modulators |
| Bandwidth | Extremely High | Moderate to High | Ultra-High |
| Cost | Very High (Rising) | Low (CMOS compatible) | Moderate |
| Scalability | Low (Fragile wafers) | Very High (12-inch wafers) | Moderate |
🛠️ Technical Deep Dive
- InP utilizes a direct bandgap semiconductor structure (1.35 eV at 300K), enabling efficient photon emission for laser diodes (LDs) and photodetectors (PDs).
- Current industry standard for high-speed interconnects relies on Multi-Quantum Well (MQW) structures grown via Metal-Organic Chemical Vapor Deposition (MOCVD) on InP substrates.
- The lattice matching requirement for epitaxial growth on InP substrates is extremely stringent, requiring precise control of Indium and Gallium ratios to prevent dislocation defects.
- Thermal management in InP-based lasers is a limiting factor; the material's low thermal conductivity necessitates advanced packaging techniques like Aluminum Nitride (AlN) submounts to dissipate heat during 1.6T operation.
🔮 Future ImplicationsAI analysis grounded in cited sources
⏳ Timeline
Weekly AI Recap
Read this week's curated digest of top AI events →
👉Related Updates
AI-curated news aggregator. All content rights belong to original publishers.
Original source: 虎嗅 ↗

