Lumentum CEO on business: Very much levered to the hyperscalers

Key Concepts

• Optical Connectivity: The use of light as the fundamental mechanism for data transmission within high-performance computing environments.
• Compute Clusters: Massive groupings of processors (GPUs/CPUs) that work in parallel to train and run AI models.
• Hyperscalers: Large-scale cloud service providers (e.g., Google, Microsoft) that build and operate massive data center infrastructures.
• Inference and Training: The two primary stages of AI development; training involves building the model, while inference involves the model performing tasks.
• Connectivity Fabric: The physical and logical infrastructure that links compute resources within a data center to ensure high-speed data transfer.

---

1. Lumentum’s Role in the AI Ecosystem

Lumentum, led by CEO Michael Hurlston, operates as a critical infrastructure provider for the AI revolution. The company specializes in optical technology, which serves as the "backbone" for modern data centers. By using light to transmit data, Lumentum enables the high-speed connectivity required to link massive clusters of compute power provided by companies like NVIDIA, AMD, and Intel.

2. Market Dynamics and Growth Drivers

• Data Center Dominance: Hurlston confirmed that the data center is the primary driver of Lumentum’s growth. The company’s business is specifically tailored to support the infrastructure that powers AI models.
• Hyperscaler Dependency: Lumentum’s revenue is heavily leveraged toward "hyperscalers"—the massive tech companies building the world’s largest data centers. Hurlston noted that their business is less focused on traditional data center real estate providers (like Equinix) and more focused on the high-performance hardware that resides inside those facilities.
• Stock Performance: Despite a recent market correction (a 30% imbalance in demand leading to a significant single-day share price drop), the company has seen a 1,300% increase in share value over the past 12 months, reflecting the massive surge in AI-related capital expenditure.

3. Strategic Monitoring of AI Developers

Lumentum maintains a close watch on the activities of AI model developers such as OpenAI and Anthropic. While Lumentum is "one step removed" from these companies—meaning they do not sell directly to them—they monitor these firms because:

• Demand Forecasting: The spending plans and development roadmaps of OpenAI and Anthropic dictate the hardware requirements of the hyperscalers (Google, Microsoft, etc.).
• Infrastructure Scaling: As these AI companies develop more complex models, the demand for training and inference capacity increases, which in turn drives the need for Lumentum’s optical connectivity solutions.

4. Technical Methodology: Connecting Compute

The core value proposition of Lumentum is the "connectivity fabric." As AI models grow in size, they require thousands of GPUs to work in unison. Traditional copper-based connections cannot handle the bandwidth and speed requirements of these massive clusters. Lumentum’s optical solutions allow for:

• High-Speed Data Transfer: Utilizing light to move data between compute nodes with minimal latency.
• Scalability: Enabling the creation of massive, interconnected clusters that are essential for modern AI training and inference tasks.

5. Synthesis and Conclusion

Lumentum occupies a vital niche in the AI supply chain. By providing the optical "plumbing" that allows massive compute clusters to function, the company has become a direct beneficiary of the AI infrastructure boom. While the company faces short-term market volatility, its long-term growth is intrinsically tied to the continued expansion of hyperscale data centers and the increasing computational demands of AI developers like OpenAI and Anthropic. The transition toward light-based connectivity is not merely an upgrade but a fundamental requirement for the next generation of AI compute.