In the last quarter, NVIDIA has poured at least $6.5 billion into a variety of photonics technology R&D ventures, ramping up its efforts to overcome key bottlenecks in AI applications. Photonics technology leverages optical signals for data transmission, boasting significantly higher efficiency compared to the current standard of electrical signal transmission. The substantial energy consumption linked to electrical signal transmission poses a significant barrier to the widespread adoption of AI on a grand scale. Since early March, NVIDIA has funneled a total of $2 billion into three photonics technology firms—Lumentum, Coherent, and Marvell—and has invested $500 million in Corning to collaborate on developing advanced optical interconnect solutions. Moreover, NVIDIA took part in the $500 million Series E funding round for Ayar Labs, a burgeoning startup in the optics field. Analysts highlight that by embracing photonics technology, NVIDIA can scale up its AI infrastructure, sidestep the high energy consumption issues tied to the persistent use of copper cables and electrical signal transmission, guarantee ongoing technological advancements, and break through bottlenecks in computing power expansion and performance enhancement. Photonics technology has the potential to replace traditional copper cable-based electrical signals for data transmission between graphics processing units, memory, network chips, servers, and data centers. In the long run, photonics technology is poised to play an ever-more crucial role in AI infrastructure. NVIDIA's next-gen rack-level AI solutions will lean more heavily on optical interconnect technologies to tackle the exponential growth in bandwidth. Presently, NVIDIA has already integrated some photonics technologies into its networking products and rolled out related tools to facilitate AI computing centers in connecting millions of graphics processing units across different locations, substantially cutting down on energy consumption and operational costs.