🎧 Prefer to listen?
I’ve been watching the AI infrastructure space for a while now, and something just happened that deserves way more attention than it’s getting. Yesterday — June 16, 2026 — a company called Coherent Corp broke ground on a $650 million expansion of its semiconductor facility in Sherman, Texas. Jensen Huang, the CEO of NVIDIA, was literally there with a shovel. And if you’re wondering why the CEO of a trillion-dollar AI company is flying to a small town an hour north of Dallas for a groundbreaking ceremony, that’s exactly the right question.
What Coherent Actually Does (In Plain English)
You know how your Wi-Fi slows down when too many devices are connected? Imagine that problem, but across a data center with thousands of AI chips all trying to talk to each other at the same time. Copper cables — the ones most data centers still use — physically can’t carry data fast enough when you’re linking hundreds of GPUs across multiple racks.
That’s where Coherent comes in. They make the lasers and optical components that move data between AI chips at the speed of light. Not figuratively — literally at the speed of light. Their specialty is something called indium phosphide (InP), a compound semiconductor that can emit and modulate light for data transmission.
Think of it this way: if NVIDIA’s GPUs are the brains of AI, Coherent’s photonics are the nervous system. Without them, those brain cells can’t communicate.
Why Jensen Huang Showed Up in Person
The timing isn’t random. NVIDIA just announced its Vera Rubin Ultra NVL576 system — a beast that links 576 GPUs across eight racks, operating as a single system. At that scale, copper is dead. You need optics, and you need them manufactured at a volume that doesn’t exist yet.
Coherent runs the world’s first 6-inch InP wafer fab. Most of the world’s InP production is still stuck on 3- and 4-inch wafers. Moving to 6-inch roughly quadruples the usable area per run, which means more lasers, more transceivers, more optical modules — all the components that plug into NVIDIA’s networking switches and move data across the data center floor.
“This site will support more than 550 direct jobs — and thousands of jobs, direct and indirect,” said Coherent CEO Jim Anderson at the groundbreaking.
The Money Behind It
Here’s where it gets interesting for anyone paying attention:
- NVIDIA invested $2 billion in Coherent — a multiyear strategic partnership covering R&D, manufacturing expansion, and a multibillion-dollar purchase commitment for optical products
- $50 million CHIPS Act grant from the U.S. Department of Commerce
- $17 million from the Texas CHIPS program and Sherman Economic Development Corporation
- $650 million total expansion cost for the new facility
That’s not a speculative bet. That’s NVIDIA putting real money into a supply chain they depend on. When the company that makes the world’s AI chips invests $2 billion in a photonics supplier, it’s because they’ve done the math and they need those lasers at scale.
Can You Actually Invest?
Yes. Coherent trades on the NYSE under the ticker COHR. As of today, it’s trading around $414 — up roughly 65% year-to-date after jumping 8.2% on the groundbreaking news. The stock is near its 52-week high, so this isn’t a “buy the dip” situation. It’s a “the market is starting to understand what this company actually is” situation.
Some numbers to know:
- EPS guidance for the current quarter: $1.52–$1.72
- Adjusted EPS last quarter: $1.41 (beat estimates)
- Free cash flow: negative $383 million (as of March 2026) — heavy capex phase as they’re building out capacity
- Operating margin: under pressure from the expansion
The negative free cash flow is worth noting. They’re spending aggressively to build capacity, which is exactly what you’d expect from a company that just secured $2 billion from NVIDIA and a $50 million federal grant. They’re not hoarding cash — they’re scaling.
Why This Matters If You Use AI Tools
If you’ve ever used ChatGPT, Claude, or any AI image generator, you’ve benefited from the infrastructure layer that companies like Coherent build. Every time you send a prompt, that data travels through optical networks — lasers, transceivers, fiber optic cables — to reach the GPU clusters that process your request, then travels back.
The AI tools you use every day (and there are a lot of them) are only as good as the infrastructure underneath them. When that infrastructure scales, AI gets faster, cheaper, and more capable. When it bottlenecks, everything slows down.
This is the picks-and-shovels play of the AI gold rush. While everyone’s debating which AI chatbot is best or whether AI agents will take their job, the companies building the physical backbone of AI are quietly locking in billions in contracts.
The Bigger Picture: AI’s Supply Chain Is Being Rebuilt in America
Coherent’s expansion is part of a larger trend. The CHIPS and Science Act — the $50 billion federal program to bring semiconductor manufacturing back to the U.S. — is funding projects like this across the country. But this one is different because it’s not about logic chips (the ones Intel and TSMC make). It’s about the compound semiconductors that enable AI networking — a supply chain that’s been thin for years.
“AI is the ultimate general-purpose technology,” Huang said at the event. “Because intelligence is fundamental — the ability to process information, to reason and solve problems — it affects every single industry.”
He’s not wrong. And the fact that NVIDIA is vertically integrating its optical supply chain — investing $2 billion in a supplier rather than just buying components on the open market — tells you everything about where they think demand is heading.
What I’d Watch Next
If you’re tracking this space, here are the signals:
- Coherent’s next earnings — watch for revenue growth from the NVIDIA partnership
- NVIDIA’s Vera Rubin launch — the NVL576 system is the first product that requires Coherent’s scale
- CHIPS Act disbursements — more grants = more capacity = more revenue for Coherent
- Competitor moves — companies like II-VI (now part of Coherent) and Lumentum are also in the photonics space
This isn’t financial advice — I’m a blog that teaches you how to use AI tools, not a brokerage. But if you’re wondering why your AI automations are getting faster and cheaper, part of the answer is being built right now in Sherman, Texas.
The Bottom Line
Coherent Corp is the company most people have never heard of that makes the technology everyone’s AI depends on. NVIDIA just bet $2 billion that this dependency is only going to grow. The stock is already up 65% this year, but the expansion just started.
Whether you invest or not, understanding the infrastructure layer of AI helps you make smarter decisions about which tools to use, which skills to learn, and where the industry is actually heading — not where the hype says it’s going.
The light that carries your AI prompts is being manufactured in a small Texas town. And that town just got a lot more important.
Want to stay ahead of AI infrastructure shifts like this? Start here — I cover the tools, trends, and tactics that actually matter for non-technical people building with AI.