Michael Hurlston, CEO of Lumentum, has issued a warning regarding a significant supply deficit in Indium Phosphide (InP) lasers, essential components for AI data center infrastructure. According to Hurlston, the supply gap for these critical optical communication elements is currently more severe than the widely reported shortage of storage chips. Despite operating five dedicated InP wafer fabrication plants, Lumentum reports that its product shipments remain more than 30% below current customer demand, highlighting a deepening bottleneck in the physical layer of artificial intelligence scaling.
The Electro-absorption Modulated Laser (EML) Bottleneck
The primary pressure point in the optical supply chain is concentrated within the EML (Electro-absorption Modulated Laser) segment. These lasers are vital for high-speed data transmission over fiber optic networks, which are increasingly replacing copper connections within GPU clusters. Analysts, including the prominent market commentator "White Hair Stock God" Serenity, suggest that this shortage is likely to spread across the broader InP laser ecosystem as Nvidia, AMD, and various hyperscale cloud providers compete for limited resources.
- Production Deficit: Shipments are trailing demand by over 30% as of August 2026.
- Critical Material: Indium Phosphide is the foundational compound semiconductor for nearly every laser used in modern AI data centers.
- Strategic Investments: Major players like Nvidia have recently committed billions to secure long-term supply agreements and future capacity.
Investment Implications and Infrastructure Impact
The persistent scarcity of InP lasers has refocused market attention on "bottleneck investments" within the optical communication laser supply chain. While the market has historically focused on GPU compute power, the ability to utilize that power is increasingly dependent on interconnect efficiency. Serenity noted that this environment creates unique value for companies controlling the "chokepoints" of AI infrastructure, particularly those involved in 1.6T optical modules and Co-packaged Optics (CPO) architectures.
"Between the two of us, I don't think we can service the demand that Nvidia and others are now putting on us to solve this resistance problem in the data center", Hurlston stated, referring to the combined capacity of Lumentum and its primary competitors.
The current supply-demand imbalance is expected to persist as AI clusters scale, shifting the industry's focus from chip design to manufacturing execution. While companies are expanding domestic manufacturing, such as Lumentum's new facility in North Carolina, significant new capacity is not expected to reach the market until 2028, suggesting a multi-year period of tight supply for the optical components that power the AI boom.
Frequently Asked Questions
Quick answers to the most common questions about this topic.
For educational purposes only. Nothing here constitutes financial or investment advice. Crypto markets are highly volatile — always DYOR before making any decisions.
Fact-checked per our editorial policy. We maintain strict independence from advertisers. Spot an error? Let us know.