Nvidia Invests $4 Billion in Photonics to Maintain AI Lead

Nvidia announced on Monday that it will invest $2 billion each in Lumentum and Coherent. Both firms are developing photonics technologies for data centers, such as optical transceivers, circuit switches, and lasers, which facilitate high-speed data transmission over long distances. Following Nvidia's 2020 acquisition of networking hardware company Mellanox to enhance NVLink and increase data transfer between GPUs, these technologies are expected to improve energy efficiency, data transmission speeds, and bandwidth in future AI data centers.

For Lumentum, this non-exclusive multi-year agreement includes "purchase commitments worth billions of dollars and access to future capacity for advanced laser components," as well as support for research and manufacturing expansion. Coherent's agreement is described similarly, containing "billions in purchase commitments along with access and capacity rights for future advanced laser and optical network products."

The development of agent-based AI, such as Anthropic's Claude Cowork and Microsoft's Copilot Tasks, has driven bandwidth demands in AI data centers to support rapid multi-task execution. Photonics may provide the solution—fiber optics can support higher bandwidth and lower latency than copper cables while consuming less power.

Nvidia is not the only organization focusing on photonics. Last month, DARPA released a request for research proposals to advance photonics computing, particularly related to AI applications. Nvidia's competitor AMD also acquired silicon photonics startup Enosemi last year, stating it would "accelerate" optical innovation within its AI systems.