InterviewAlly

Funding & Valuations

NVIDIA Invests $4B in Optical Interconnects

NVIDIA announced a $4 billion strategic investment split between Lumentum and Coherent, securing the optical interconnect supply chain essential for next-generation AI data center infrastructure.

March 9, 2026 · 6 min read · Source: CNBC

NVIDIA · Photonics · AI Infrastructure · Supply Chain · Data Centers

Optical fiber cables and photonics components representing NVIDIA's AI data center infrastructure investments

NVIDIA announced a landmark $4 billion investment in optical interconnect technology, committing $2 billion to Lumentum and $2 billion to Coherent. The strategic funding, announced on March 2, 2026, secures critical supply chains for photonics components essential to scaling next-generation AI data centers and managing the exponential growth in AI computing infrastructure.

The Investment Structure

Rather than outright acquisitions, NVIDIA established multi-year strategic partnerships with both companies, backed by multi-billion dollar purchase commitments. The agreements include dedicated R&D resources focused on NVIDIA's requirements, expanded U.S.-based manufacturing facilities, long-term supply security for optical components, and technology co-development between NVIDIA engineers and photonics specialists.

Both Lumentum and Coherent will use the funding to build out their U.S.-based manufacturing capabilities, aligning with broader reshoring trends in the semiconductor and photonics industries. The investments also include capacity rights agreements, giving NVIDIA future access to production output as demand for optical interconnects accelerates.

The stock market reacted enthusiastically: Lumentum shares jumped approximately 12%, while Coherent stock surged roughly 15%, reflecting investor confidence in the optical interconnect sector's growth trajectory.

Why Photonics Matter for AI

As AI models grow larger and training clusters expand to tens of thousands of GPUs, data centers face unprecedented networking demands. Traditional copper-based interconnects struggle with power consumption, bandwidth bottlenecks, latency, and scalability constraints. Optical interconnects, which transmit data as light through fiber optic cables, overcome these limitations by delivering higher bandwidth at lower power consumption, improved energy efficiency, greater network resiliency, and support for massively scaled AI training clusters.

"Optical interconnect technology and package integration are critical for the continued scaling of AI factories, improving the energy efficiency and resiliency of large-scale AI networks." — NVIDIA

Strategic Rationale and Supply Chain Security

This investment underscores NVIDIA's recognition that GPUs alone don't define competitive advantage in AI infrastructure. The interconnects linking GPUs together are equally critical. By securing supply chains for photonics technology, NVIDIA ensures it can build larger, more efficient AI clusters while reducing data center operating costs through energy efficiency.

The investment also reflects geopolitical considerations. By backing U.S.-based photonics manufacturers, NVIDIA reduces dependence on international supply chains for critical AI infrastructure components, supporting reshoring initiatives and economic resilience. Lumentum brings expertise in optical modules and transceivers used in high-speed networking, while Coherent specializes in coherent optical technologies for long-distance, high-capacity data transmission.

Industry Context: The Interconnect Arms Race

NVIDIA is not alone in recognizing the importance of networking infrastructure. Google has invested heavily in custom networking chips, Meta built its own interconnect specifications, and Microsoft is developing specialized data center architectures. NVIDIA's $4 billion commitment signals that optical interconnects represent the future of AI infrastructure — and that securing supply chains is as strategic as chip design itself.

What This Means for Engineers

For engineers working in AI infrastructure, cloud computing, or data center design, this investment validates a critical career path: infrastructure engineering is becoming as important as algorithm development. Understanding networking, photonics, and data center architecture provides a competitive advantage in the AI economy. If you're preparing for technical interviews at infrastructure-focused roles, InterviewAlly can help you practice system design questions covering distributed computing, networking, and scalability — the exact domains where photonics is revolutionizing how AI systems operate.