
Credit: Michael Hurlston(CEO Lumentum) and Anna Tutova(Founder AI Crypto Minds) at RAISE AI Summit in Paris.
The optics giant’s stock has surged 637% in a year, but a recent pullback has Wall Street asking: Is the AI infrastructure trade cooling, or is this the best buying opportunity of the decade?
When Michael Hurlston took the helm at Lumentum in February 2025, the company was a respected but relatively obscure player in the telecom optics space. Eighteen months later, Lumentum (NASDAQ:LITE) commands a $59 billion market cap, its stock has returned 637% over the past year, and it just earned a seat in both the S&P 500 and the Nasdaq-100. The reason? Michael Hurlston saw something the rest of the market didn’t: the moment when copper inside AI data centers would start to hit a hard physics wall, and light would have to take over.
“I’ve been a CEO now at three different companies,” – Michael Hurlston said to me during the RAISE AI Summit in Paris. “This is the third time I’ve been a CEO of a public company. That’s very rare because, quite often as a CEO, you’re either fired or you retire.” He pauses, then adds: “Previously, I was in a semiconductor company, Synaptics. And then I joined this as my second optics company because I saw a lot of opportunity. I thought that the market was shifting from ethernet connectivity, copper connectivity to optics.”
That shift is now the defining investment thesis of the AI infrastructure boom and it’s happening faster than almost anyone predicted.
From Telecom to Hyperscalers: The Customer Base That Changed Everything
For decades, Lumentum’s business revolved around the usual suspects: AT&T, Verizon, Deutsche Telekom, British Telecom. These carriers built the internet backbone with fiber optics, and Lumentum supplied the laser chips that made it possible. Then AI happened.
“Our optical industry was very much focused on telecom,” Michael Hurlston explains. “Anything that needs to move over a long distance typically travels across light. And what you see now in the data centers is, with the speeds that they’re trying to incur when you’re moving data around the data center, you have the need now for more and more optics.”
The customer list tells the story. Google, Meta, Nvidia, Amazon, the hyperscalers now driving the world’s AI buildout, have become Lumentum’s core market. And their demands are on an entirely different scale.
“The volumes that they’re asking us to generate are much, much larger,” Michael Hurlston says. “Our whole shift really has been toward these hyperscalers and away from the telecommunications type of companies.”
The numbers back it up. Lumentum’s quarterly revenue has tripled from roughly $300 million to over $800 million in just four to five quarters. Gross margins have expanded from 33% to approximately 48%. And the company is, by its own account, sold out through 2028.
The $2 Billion Nvidia Bet and Why Coherent Is a Customer, Not Just a Rival
In March 2026, Nvidia made a $2 billion strategic investment in Lumentum, a move designed to lock in supply of the laser chips that power its next-generation AI systems. It was a watershed moment and one that came with some irony. Lumentum had previously lost a bid to Coherent in 2021, yet now finds itself as Nvidia’s preferred partner in the co-packaged optics race.
“Nvidia is a key customer of ours,” Michael Hurlston says. “They are taking the lead in co-packaged optics where they actually have the technical ability to bring the optical engine onto their silicon substrate. So, they’ve got a big lead in this area, and we’re partners with them. They invested $2 billion in the company and have really, really been a strong partner.”
But the relationship with Coherent, often framed as Lumentum’s chief rival, is more nuanced than the headlines suggest.
“Coherent actually, it turns out in this industry, you have coopetition,” – Michael Hurlston says. “Coherent is actually one of our largest customers, and they’re a very, very good customer of ours. I think this market is so big, what we’re talking about here with optical scale-up and this replacement of copper wires in the backplane of a server system. The market is so big, it’s going to take multiple people to serve it.”
He draws a sharp distinction between the two companies’ strategies: “We think of ourselves as components-first, where we design these laser chips, and that’s our kind of bread and butter. But, Coherent, I think, thinks of themselves as a transceiver company, systems manufacturer first, and they’re one of the best in the world in terms of systems manufacturing. So, we supply components into their systems, play to our strength, and they very much play to their strength in the industry as well.”
The Copper Question: Jensen Huang’s Warning and Why Michael Hurlston Agrees With Him
At Nvidia’s GTC conference, CEO Jensen Huang put it bluntly: “You use optics wherever you must and you use copper wherever you can”. For a company whose entire valuation thesis rests on replacing copper with light, it could have been a devastating soundbite. Michael Hurlston’s response? He agrees.
“Look, a really good question,” – he says when asked about Jensen Huang’s comment. “I think that I actually agree with what Jensen said. Optics is fundamentally more expensive. So, if you can get copper to work, companies want to do that. That makes a lot of sense.”
But here’s the catch, and it’s the entire investment case in a nutshell.
“What’s happened, though, is the speeds inside this backplane are increasing,” Michael Hurlston explains. “And there’s a graph that shows as speeds increase, how copper, the distance over which signals can travel on copper, falls off. So, if we’re talking about 1.6 terabits per second, the distance over which signals can travel on copper is maybe one to two meters. And if you look in the backplane of a rack, there’s many, many lanes that are longer than one or two meters.”
The implication is clear: copper isn’t going away, but it’s reaching its physical limits just as AI clusters demand ever-higher bandwidth. For Lumentum, that means an entirely new addressable market, wiring the backplane of a rack, that didn’t exist for optics companies before.
“It’s not like copper is going to go away tomorrow. There’s going to be a coexistence of optics, of copper. But for the optical industry, it’s a whole new opportunity. We’ve never participated in wiring the backplane of a rack. And so, for us, it’s a very, very significant thing.”
The Pullback Paradox: Sold Out Through 2028, Stock Down 29.5% From Highs
Here’s where the story gets interesting for investors. Despite being sold out through 2028, despite tripling revenue and expanding margins, Lumentum’s stock has pulled back sharply from its all-time high of $1,085.68 to trade around $765 as of July 20, a decline of roughly 29.5%.
“Our stock hit over $1000 a share. Now it’s trading closer to $700, and that’s largely due to the macro,” Michael Hurlston says. “I think there’s a lot of concern in the investor community around AI and around the AI hardware trade. Of course, now we have more around oil and the conflict in Iran. All of these macro trends are affecting our stock price, as well as every other company in the industry.”
But he doesn’t mince words about where he thinks the stock is headed: “I won’t be surprised if our stock goes well above $1000 a share in the very near future. I think our numbers will end up being very, very robust, and that will only help the stock price.”
Wall Street seems to agree, at least directionally. Analysts maintain an average price target of $1104.89, with a high estimate of $1400, implying upside of roughly 50% to 90% from current levels.
The AI Bubble Debate: “We’re Actually Undervalued Relative to Our Peers”
With AI-related stocks trading at eye-watering multiples, the bubble question is unavoidable. Michael Hurlston’s counterargument is rooted in fundamentals, not hype.
“We’ve grown the company from $300 million a quarter to now a billion dollars a quarter. So, our growth has been over triple in the short period of time I’ve been with the company,” he says. “At the same time, we’ve actually seen margin expansion. Our margin has gone from 33% gross margin to something closer to 48% gross margin. All of that leads to earnings. And our earnings power has increased measurably over the four to five quarters that I’ve been with the company.”
He adds: “If you look at our stock in particular, I think we’re actually undervalued relative to our peers because of the growth, because of the earnings expansion that we’ve seen.”
The structural tailwind, he argues, is what separates Lumentum from pure-play AI trades that could collapse if data center buildouts slow. “Irrespective of the growth of data centers, there’s a technical shift, a technical shift towards optics. And so, even if the number of data centers stagnates, we don’t see the growth rate that we’ve seen over the past four or five years in terms of the build-out, we still have a growth driver because you have the shift inside the data center from copper to optics.”
No Need to Buy Growth But They’re Shopping for Tech
Unlike his previous company, Synaptics, which was “built on M&A,” Michael Hurlston says Lumentum is in the enviable position of not needing acquisitions to grow. The organic demand is simply too strong.
“In the past, when we were in the telecom industry and serving those telecom customers, the growth rate was much lower. So, the company needed to acquire to grow,” he says. “But right now, we’re in a position where we don’t need that.”
That said, the company is selectively hunting for complementary technologies to widen its hyperscaler product portfolio. “An obvious one for us would be additional components that go around the co-packaged optics and near-packaged optics,” Michael Hurlston says. “If you look at CPO and NPO, lasers are key. But there’s also photonics ICs. We make a photonic IC, but I wouldn’t say we’re the best. There are photodiodes that receive the information on the receive side, we would like to get into that market if we could. There are laser drivers and trans-impedance amplifiers that go into that optical engine.”
The Automation Gap and Why It Might Not Matter
In an industry where manufacturing efficiency can make or break margins, Michael Hurlston is surprisingly candid about Lumentum’s automation shortcomings.
“In general, I’d say we’re behind on the use of AI as a company,” he admits. “Interestingly enough, as you said, we’re building all this infrastructure for AI, we just hired somebody that’s going to really be a champion internally to look for these various use cases where we can employ AI.”
He also acknowledges lagging behind Coherent on factory robotics. “If you look at the best-in-class optics companies, we talked about Coherent. Coherent does a terrific job employing automation, which means robotics, inside their factory. We have some. We definitely use robotics in some parts of our operation, but I think there’s an opportunity for us to use more.”
The reason for the gap? Unlike semiconductor manufacturing, where processes are standardized across foundries like TSMC, optical manufacturing is highly customized.
“In the optical industry, there’s a lot of customization that goes on. Our factory is probably different from Coherent’s factory, and therefore we have to customize the robotics,” Michael Hurlston explains. “Our biggest problem is probably invention. How do we invent the right robots? How do we make the right robots?”
What Comes Next
With earnings due August 11 and the stock bouncing off recent lows, all eyes are on whether Lumentum can deliver the kind of numbers that justify a return to four-digit territory. For Miachel Hurlston, the path forward is clear: deepen hyperscaler partnerships, expand the component portfolio, and let the physics of data center bandwidth do the rest.
“Nvidia is an important partner for us because they’re a great optical engineering company, and they’re teaching us ideas, giving us ideas that we can use not just for our laser chips, but for other products we might do. The same is true of Google,” he says. “We learn a lot from them. We really learn how to modify and adapt our products to what they need. And by having these close relationships, it’s really enabled us to grow our business pretty successfully.”
In an AI market obsessed with the next model release or GPU architecture, Lumentum’s story is a reminder that the most durable investment themes often hide in the supply chain: in the laser chips that no one talks about but everyone needs. Michael Hurlston, for his part, seems content to let the skeptics sell.
After all, when you’re sold out through 2028, you don’t have to convince anyone. You just have to ship.
Lumentum (LITE) closed at $765.55 on July 20, 2026. The company reports fiscal Q4 2026 earnings on August 11.
Benzinga Disclaimer: This article is from an unpaid external contributor. It does not represent Benzinga’s reporting and has not been edited for content or accuracy.
Login to comment