Jensen Huang has identified a new critical bottleneck in artificial intelligence: the speed at which data moves between chips and servers. As AI chatbots and agents become faster and more responsive, the infrastructure that connects them is struggling to keep pace. This insight has prompted Nvidia to make aggressive strategic investments in companies developing photonics technology, a solution that uses light instead of traditional electrical signals to transfer data more efficiently.

What Is Silicon Photonics and Why Does It Matter?

Silicon photonics is a technology that leverages light to move data between chips and across data centers, dramatically improving the speed and efficiency of data transfer. Unlike traditional copper-based connections, photonics can handle vastly larger amounts of information with less power consumption and heat generation. As AI systems grow more complex and demand faster responses, the bottleneck has shifted from raw computing power to the infrastructure connecting that power.

"The amount of silicon photonics technology capacity that we need is substantially higher than the world has today," said Jensen Huang, Nvidia Founder and CEO.

Jensen Huang, Founder and CEO at Nvidia

This statement, made at a conference in March, underscores the urgency of the problem. Huang's recognition of this constraint has shaped Nvidia's investment strategy, directing billions toward companies positioned to scale photonics production and deployment.

How Is Nvidia Addressing the Data Transfer Challenge?

• Coherent Investment: Nvidia invested $2 billion in Coherent and formed a strategic agreement giving Nvidia future access and capacity rights to Coherent's advanced laser and optical networking products. As of March 31, Nvidia's stake in Coherent was just under 4%, and the company's stock has climbed approximately 110% in 2026.
• Lumentum Partnership: Nvidia announced a $2 billion investment in Lumentum Holdings in March, with a similar strategic agreement. Lumentum's total revenue nearly doubled in the quarter ended March 28, rising from $425.2 million to $808.4 million, with its systems segment growing 121%.
• Nokia Collaboration: In 2025, Nvidia and Nokia partnered to accelerate AI networking infrastructure development. Nvidia invested $1 billion, acquiring a 2.9% stake in Nokia as of March 31. Nokia's silicon photonics technology integrates complex optical subsystems into silicon chips, improving performance and reducing data movement costs.

These three investments total $5 billion and represent Nvidia's bet that photonics will become essential infrastructure for the next generation of AI systems. Each company brings different strengths: Coherent specializes in laser and optical networking, Lumentum focuses on components and systems for data centers, and Nokia brings decades of telecommunications expertise combined with emerging AI infrastructure capabilities.

What Are the Financial Implications for These Companies?

All three companies have experienced significant stock appreciation in 2026, driven by optimism about AI infrastructure demand. Coherent reported revenue surged 27% to $1.8 billion in its most recent quarter, with earnings per share increasing 55% to $1.41 and data center and communications revenue climbing 41%. However, this segment accounts for 75% of Coherent's total revenue, creating concentration risk if data center demand slows.

Lumentum's growth has been even more dramatic, with total revenue nearly doubling and its components segment growing 77% to $533.3 million. The company expects revenue in the current quarter to range from $960 million to $1 billion. Yet Lumentum faces intense competition, including from Coherent, and its forward price-to-earnings ratio of 50 suggests high expectations with limited room for error.

Nokia's cloud and AI operations revenue rose 49% in the first quarter of 2026, and its stock has climbed over 100% this year. The company is transitioning from its legacy telecommunications identity to position itself as an AI infrastructure provider, a transformation that still requires successful execution.

How Does This Shift Reflect Broader AI Infrastructure Trends?

For years, energy and memory stocks have been the primary beneficiaries of AI's explosive growth, as demand for computing power outpaced supply. Huang's identification of the data transfer bottleneck signals a maturation in how the industry thinks about AI infrastructure. Rather than focusing solely on raw computational capacity, attention is shifting to the systems that connect that capacity and enable rapid data movement.

This represents a significant opportunity for companies with photonics expertise, but it also highlights the complexity of building AI infrastructure at scale. Each component of the system, from chips to cooling to data transfer, must evolve in concert. Nvidia's strategic investments suggest the company believes photonics will be as critical to future AI systems as GPUs (graphics processing units) have been to current ones.

Beyond Nvidia's direct investments, the broader market is taking notice. Coherent's addition to the S&P 500 in March reflects growing recognition of photonics companies' importance to the technology sector. As AI applications demand faster response times and more efficient data movement, the companies solving these infrastructure challenges are likely to see sustained demand.

The convergence of Huang's public statements about photonics constraints and Nvidia's substantial financial commitments to three key players suggests this is not a speculative trend but a deliberate strategic pivot. Whether these companies can scale production to meet Nvidia's expectations, and whether they can maintain profitability amid intense competition, will shape the next phase of AI infrastructure development.