# NVIDIA Highlights AI Hardware Bottlenecks

NVIDIA CEO Jensen Huang has identified memory and silicon photonics as the critical bottlenecks constraining AI infrastructure performance, emphasizing the need for advanced hardware components to support growing AI workloads. Significant investments and supply commitments are underway, including NVIDIA's $2 billion investment in Coherent Corp. and multibillion-dollar purchase commitments, as well as Tower Semiconductor securing over $1 billion in photonics contracts for 2027. Procurement professionals and government contractors should prioritize engagement with key suppliers like Micron Technology, Coherent Corp., Tower Semiconductor, and Lumentum Holdings to address these emerging hardware demands and support AI deployment initiatives.

- **Why this matters:** Memory and silicon photonics are now recognized as the primary hardware constraints limiting AI scalability, signaling procurement shifts toward these specialized components.
- Agencies and contractors involved in AI infrastructure should evaluate supplier capabilities and contract opportunities with leading vendors investing heavily in these technologies.
- The substantial contract values and investments indicate growing market demand and potential for long-term procurement partnerships in AI hardware supply chains.
- Organizations should consider supply chain resilience and capacity access when planning AI-related procurements to mitigate bottlenecks in data center hardware.

**Jurisdictions:** federal
**Industries:** Information Technology
**Topics:** Artificial Intelligence
**Published:** September 20, 2026

### Vendors
- NVIDIA (prime contractor)
- Micron Technology (supplier)
- Coherent Corp. (supplier)
- Tower Semiconductor (supplier)
- Lumentum Holdings (supplier)

### Key Quotes
> NVIDIA is investing $2 billion in Coherent alongside a multibillion-dollar purchase commitment and future access to its capacity.
> — Jason Bodner, Co-founder of MoneyFlows

> AI's constraint right now is not appetite. It is the physical hardware that moves data around inside a data center, and NVIDIA named the two pinch points: memory and silicon photonics.
> — Jensen Huang, CEO of NVIDIA

### Sources
- [NVIDIA Just Named AI's Next Bottleneck—And These 3…](https://www.inkl.com/news/nvidia-just-named-ais-next-bottleneck-and-these-3-stocks-sit-right-in-it) - inkl