Chat with this content: Summary, Analysis, News...
Share on Reddit
Coherent Corp.: From Optical Components to AI Infrastructure Giant
I. Introduction & Episode Teaser
Picture this: March 2025 at NVIDIA’s GTC conference in San Jose. Jensen Huang is on stage, not just talking about faster GPUs, but about something even more fundamental: silicon photonics networking switches. The kind of technology you need if you want “AI factories” to scale—connecting vast numbers of GPUs with the bandwidth and efficiency those systems demand.
On NVIDIA’s slide of ecosystem partners—TSMC, Browave, Coherent, Corning Incorporated, Fabrinet, Foxconn, Lumentum, SENKO, SPIL, Sumitomo Electric Industries, TFC Communication—one name sticks out for how unfamiliar it is to most people. Coherent Corp. Not a household brand. But suddenly, it’s sitting in the inner circle of companies building the optical backbone of the AI era.
By early January 2026, Coherent’s market cap is roughly $30 billion, nearly doubling over the prior year. The stock has run from a 52-week low of $45.58 to highs approaching $200, powered by a surge of AI-driven demand that has re-rated the business from a relatively under-the-radar photonics supplier into critical infrastructure.
So here’s the deceptively simple question driving this whole story: how did a Pittsburgh-area company—founded in 1971 to make specialty crystals for industrial lasers—end up as one of the key enablers of the AI boom?
The answer is less about a single invention and more about a decades-long build. A relentless, disciplined roll-up of adjacent technologies that culminated in three pivotal moves: the Finisar acquisition in 2019, the Coherent Inc. mega-merger that closed in 2022, and a $1 billion strategic investment from Japanese auto giants that validated the next platform the company is building. Together, they turned Coherent into a vertically integrated photonics powerhouse—positioned right where multiple megatrends collide.
The financials show what happens when that strategy meets the right moment. Fiscal 2025 revenue hit a record $5.81 billion, with non-GAAP gross margin of 37.9% and non-GAAP earnings per diluted share of $3.53—up 23% year over year. Management headed into fiscal 2026 expecting continued growth, with AI infrastructure becoming a meaningful tailwind rather than a nice-to-have.
This is the story of one of the most successful technology roll-up strategies of the past two decades: a company that kept buying, integrating, and compounding capabilities—and then woke up to find the world suddenly needed exactly what it had spent fifty years assembling.
II. The II-VI Foundation: Building from the Periodic Table (1971–1990s)
In 1971, a young electrical engineer named Carl Johnson ran into a very specific problem: he needed a high-quality cadmium telluride crystal for his PhD work at the University of Illinois. Not the kind you could just order from a catalog. So he tracked down a tiny, two-person crystal-growth outfit outside Pittsburgh. The crystal grower was James Hawkey. Hawkey’s material ended up enabling the measurements Johnson was trying to make—studying the electro-optic properties of cadmium telluride using far-infrared lasers and spectroscopy.
Johnson wasn’t new to hard problems. He earned his bachelor’s degree at Purdue, his master’s at MIT, and his doctorate in electrical engineering at the University of Illinois. After finishing his PhD in 1964, he did a quick tour through some serious industrial R&D: two years on Bell Labs’ technical staff, then five years at Essex International, an automotive products company, where he served as director of research and development. But by 1971—still in his late 20s—Johnson wanted to build, not just research. He founded II-VI, and he’d run it for more than 35 years.
Even the name told you what kind of company this was going to be. “II-VI” wasn’t branding. It was chemistry. The company began by producing cadmium telluride—cadmium from group II of the periodic table, tellurium from group VI. In other words: the identity of the business was the materials themselves. Johnson and Hawkey were materials people first, and everything else came second.
Their partnership worked because their strengths were different and perfectly matched. Hawkey led crystal growth and equipment innovation. He was the kind of builder who could wire electronics, do electrical work, handle plumbing, weld, construct—whatever the job demanded. Johnson would analyze the crystals Hawkey grew, study the results, and propose the next round of experiments. It was an iterative loop: grow, measure, improve, repeat.
Hawkey’s path into this world wasn’t formal. After high school, he couldn’t afford college. He learned his skills through work, persistence, and self-study. In the late 1960s, he got a shot working with a professor at Carnegie Tech, growing bulk gallium arsenide crystals. That experience—hands-on, practical, unusually deep for someone without a degree—became the foundation of what he brought to II-VI.
The early years were, frankly, shaky. In 1972, before the company had real revenue, Johnson and Hawkey missed several payrolls for themselves. But they never missed payroll for their two shop-floor employees. Eventually, their luck turned with a late-1972 uptick in sales—driven by just-in-time wins in shaping, polishing, and thin-film coating cadmium telluride parts, plus demand for optics and Pockels cells. That choice—employees first, founders last—didn’t just keep the lights on. It helped set a tone that would stick.
On the technical side, nothing was smooth. They blew up a few vertical Bridgman furnaces. Nobody was hurt, but you can’t build a manufacturing business when every third run takes your equipment offline. They eventually fixed the issue by changing how they synthesized the material. Hawkey knew how to work around problems. Johnson got exceptionally good at rebuilding. If a furnace failed, Johnson could have it running again in about 24 hours—roughly the time it took Hawkey to prep the next load.
The first products were straightforward and essential: lenses, windows, and mirrors for CO2 lasers. Then came a major milestone. II-VI went public in 1987, and Johnson later said the IPO proceeds enabled the company to expand its zinc selenide manufacturing capacity. The timing, though, was brutal: the offering launched five days before the October stock market crash.
Still, II-VI came through it. The company remained profitable and grew revenue every year through its first 15 years. And it wasn’t just business momentum—it was engineering momentum. II-VI’s work with zinc-, cadmium-, and gallium-based compounds helped improve the materials that controlled and directed laser energy. Over that same arc, the power generated by some lasers climbed from about 300 watts in 1971 to about 5,000 watts by 1987—a glimpse of how quickly the whole field was advancing, and how closely II-VI was tied to that progress.
Even the headquarters told you something about the company’s DNA. The Saxonburg site had lived multiple lives. It was once property of KDKA, the world’s first commercially licensed radio station, and some original structures still stood. In the 1940s, the site expanded under the Carnegie Institute of Technology and the U.S. Nuclear Regulatory Commission into a 400 MeV synchrocyclotron research facility. II-VI acquired the property around 1978 and kept expanding it over the years.
From radio pioneer, to nuclear research, to photonics manufacturing—it felt like a physical metaphor for what Johnson was building: an institution designed to accumulate capability over time.
And that’s the key investor takeaway from the “foundation” era. Johnson wasn’t building II-VI to flip it or optimize for short-term optics. He was building a materials science engine—patient, methodical, and built to last. That patience would become a competitive advantage, because it created the one thing the company would lean on for decades: the ability to keep adding new technologies without losing its core.
III. The Acquisition Machine Begins: Strategic Expansion (1990s–2010s)
Coming out of the IPO, Johnson still owned a large stake in II-VI and carried the company into the 1990s. This is the decade where the story changes. II-VI stops being “just” a great materials business and starts behaving like a builder of platforms. And the tool it reaches for, again and again, is acquisition.
But it didn’t start as some grand roll-up thesis. It started as a response to pressure.
By the early 1990s, nearly half of II-VI’s sales came from overseas customers, and a big share of that business ran through Japan. For years, II-VI had enjoyed a strong position there with limited competition, helped by a distributor, Material Technology Trading Corp., which II-VI partly owned. Then the environment shifted. Japan’s Ministry of International Trade and Industry began pushing Japanese companies to get more involved in industrial lasers, including advanced optical lenses—exactly the kind of products II-VI sold.
Competition forced evolution. II-VI’s answer was to broaden what it could do, and to do it faster than it could by R&D alone. So it began buying capabilities that sat right next to its core.
In the 1990s, II-VI started stacking acquisitions. It bought Litton Systems’ silicon carbide group from Northrop Grumman. It acquired Virgo Optics in 1995 and Lightning Optical Corporation in 1996. Those two deals were especially telling: Johnson didn’t just add revenue, he combined them into something new—VLOC, a division focused on one-micron solid-state lasers.
The Virgo/Lightning story captures the playbook in miniature. In 1995, II-VI bought Virgo Optics, based in Tampa, Florida, from Sandoz Chemical. A year later it added Lightning Optical in nearby Tarpon Springs, then merged the two into VLOC in New Port Richey. This wasn’t collecting trophies. It was assembling a capability. That expansion helped drive II-VI to $52.7 million in revenue in 1997—nearly triple what it had been just three years earlier.
Then came the kind of deal that changes a company’s trajectory: Laser Power Optics in 2000.
There aren’t many moments in corporate history when you get to buy your largest, strongest competitor. For II-VI, Laser Power was that moment. The two companies had been running into each other across global markets for a decade. In the late 1990s, Laser Power ran into trouble, its board opted to sell, and II-VI pounced. The acquisition strengthened II-VI’s position in infrared optics, taking it from a leader to the leader.
Laser Power was a roughly $31 million revenue business founded by a former II-VI employee. It made infrared optics for CO2 lasers used by industrial and military customers—and about half its sales came from military contracts. That military footprint was a big part of the appeal. Laser Power made missile domes that served as infrared “eyes,” window assemblies for tanks and helicopters, and coatings for space-based lasers. It was exactly the kind of adjacent market II-VI wanted more exposure to, especially since military sales were under 10 percent of II-VI’s own revenue at the time. Johnson was determined to land it.
And critically, this is where execution mattered as much as strategy. Fran Kramer—then CEO—backed the deal and, in the years that followed, led the integration and worked through the synergy opportunities. The company wasn’t just buying assets; it was proving it could digest them.
Through the 2000s and into the 2010s, the acquisition list kept growing: Laser Power Optics in 2000; Marlow Industries in 2004; HIGHYAG (a majority stake in 2007, then the remainder in 2013); Photop Technologies in 2010; Oclaro’s optics coating facility in Santa Rosa in 2012; Anadigics and EpiWorks in 2016; and Integrated Photonics in 2017.
Marlow, in particular, shows how II-VI defined “adjacent.” In late 2004, II-VI agreed to acquire Marlow Industries, a maker of thermoelectric products used for cooling and power generation. It wasn’t optics—but it solved a problem that optics-heavy systems always run into: thermal management. The deal was valued at about $31 million, for a business doing roughly $26 million in sales. Another bolt-on, but one that expanded what II-VI could offer system designers.
The result of all of this breadth was real financial momentum. Over a decade, II-VI grew annual revenue from $18.6 million in 1994 to $150 million in 2004. And importantly, it did it while widening its foundation—setting itself up to keep buying, keep integrating, and keep moving up the value chain.
One of the most consequential expansions came later, with VCSELs—vertical-cavity surface-emitting lasers that became important for 3D sensing. The acquisitions of Avalon Photonics, Anadigics, and EpiWorks helped II-VI ramp VCSEL capability and production, a step that would matter more and more as high-volume sensing and data applications grew.
During all of this, leadership transitions were handled with the same deliberate, engineered feel as the acquisitions. Johnson served as CEO from 1985 to 2007, then became chairman until 2014. Francis Kramer, who had been president since 1985, became CEO in 2007 and later took the chair role in 2014. In 2016, Vincent D. “Chuck” Mattera, Jr. became the third CEO, with Kramer continuing as chair.
Looking back, what stands out isn’t just that II-VI bought a lot of companies. It’s that the deals were rarely random. They were evaluated not only on revenue, but on fit—how they strengthened vertical integration from materials, to components, to more complete solutions. This wasn’t empire-building. It was the construction of a photonics stack, one acquisition at a time—quietly compounding capabilities that, later, the world would suddenly find indispensable.
IV. The Finisar Deal: Betting Big on Optical Communications (2019)
By late 2018, II-VI had already proven it could buy and integrate. But the next move wasn’t another bolt-on. It was a swing that would change what the company was.
On November 9, 2018, II-VI announced it would acquire Finisar in a deal valued at $3.2 billion. It took time—nearly a year, plus regulatory work—but on September 24, 2019, II-VI closed the acquisition. Overnight, the company went from a specialty materials and components player to something much closer to a full-stack photonics force.
The deal structure tells you how serious it was. Finisar shareholders received a mix of cash and II-VI stock—on average, $15.60 in cash plus 0.2218 shares of II-VI per Finisar share, depending on each holder’s election. The transaction was fully taxable to Finisar shareholders. II-VI financed the acquisition with a combination of term loans—$1.9 billion raised through Term Loans A and B—and about $1.1 billion of its stock. Management also laid out a clear integration target: $150 million in run-rate synergies over the following three years.
Finisar was worth all that complexity because it wasn’t just another optics company. Founded in 1988 and headquartered in Sunnyvale, Finisar was a global leader in optical communications—components and subsystems sold into networking equipment makers, data center operators, telecom service providers, and also into consumer electronics and automotive. It had R&D, manufacturing, and sales spread around the world, and it had spent decades living at the intersection of bandwidth demand and high-volume manufacturing.
One of Finisar’s defining strengths was VCSELs—vertical-cavity surface-emitting lasers—and it had credibility here early. In 1997, Finisar launched the first VCSEL-based transceiver products. Then it caught the late-1990s optical telecom moment perfectly: in 1999 it went public, raising nearly $150 million, and its value more than quadrupled on its first day of trading at the height of the optical telecommunications bubble.
That timing mattered, because the bust that followed was brutal. Finisar survived—and used the downturn to buy. It acquired companies and capabilities including passive components maker New Focus; laser diode manufacturer Genoa; Honeywell’s internal VCSEL manufacturing capability; and Infineon’s 10 Gb/s product portfolio. In other words, Finisar didn’t just endure the cycle. It came out more integrated.
By 2010, Finisar had shipped its 100 millionth VCSEL die. By 2018, that total had climbed to 300 million. By then, it was generating annual revenue well over $1 billion and employed around 12,000 people.
It was also still expanding. One of the biggest moves was investing in the former Texas Instruments facility in Sherman, Texas, to support VCSEL production for 3D sensing applications like facial security scanning. Finisar wasn’t standing still—and that was exactly why II-VI wanted it.
Strategically, II-VI framed the merger as a way to address a set of emerging markets all at once: 5G, 3D sensing, cloud computing, electric and autonomous vehicles, and advanced microelectronics manufacturing. The combined company would be one of the world’s largest optics and photonics players, with roughly 25,000 employees across 70 locations globally, including about 12,000 in China. It would also have five gallium arsenide fabrication facilities and combined annual revenue around $2.5 billion.
But deals this global don’t close on ambition alone. The Finisar acquisition faced regulatory scrutiny, particularly in China. II-VI obtained antitrust clearance from the State Administration for Market Regulation (SAMR), with a key condition: II-VI agreed it would operate Finisar’s wavelength selective switch business separately for a period expected to be three years.
And now to why this deal ended up being a hinge moment in the company’s history. Finisar brought II-VI exactly what it needed to ride the next wave: high-volume 3D sensing VCSELs, an indium phosphide (InP) platform, and deep experience integrating InP lasers with optics and electronics inside transceivers.
Think of it this way: before Finisar, II-VI sold the ingredients. After Finisar, it could deliver the finished product—up the stack, closer to the customer, and directly into the pipes that moved data through the world’s data centers.
V. The Coherent Inc. Mega-Merger: Creating a Photonics Giant (2021–2022)
If the Finisar deal was transformative, the Coherent Inc. acquisition was historic. On July 1, 2022—after a year-and-a-half saga that included a protracted bidding war and an extended antitrust review—II–VI completed its $6.56 billion acquisition of laser-maker Coherent. And then, in a move that said as much about ambition as it did about branding, II–VI didn’t keep its own name. The combined company would be called Coherent.
In the press release announcing the closing, chair and CEO Vincent D. Mattera, Jr. framed the choice as more than laser jargon. Yes, “coherent” is what you want light to be in a laser. But he also pointed to the broader meaning: bringing things together. That was the real thesis of the merger—building a company that could participate across the photonics value chain, from materials and components all the way up to systems.
The target had a story worthy of the name. Coherent was founded in May 1966 by physicist James Hobart and five cofounders, and it went public in 1970.
The laser itself had only been developed a few years earlier—at Bell Laboratories in 1958—and it was obvious to some people that it wasn’t just a lab curiosity. Hobart was one of them. He was fascinated by the idea of putting lasers to work on factory floors—cutting, welding, and manufacturing tasks that conventional tools struggled with. But at Spectra-Physics, where he worked, he couldn’t get management to commit to the industrial direction he believed in. So he left and built it himself.
Coherent began with a shoestring: about $10,000 that Hobart and his associates scraped together. The earliest version of the company looks like the beginning of a garage-startup myth, because it basically was. They needed a 220-volt outlet to power the laser they wanted to build, and the only place they could find one was a laundry room in Watson’s house. So next to a washer and dryer, they built their first product—using a piece of rain gutter as a key component in the prototype.
The methods were crude. The results weren’t. Just four months after founding the company, Coherent showed its first laser at the Westcon trade show—now refined into a shiny, telescope-like device. More importantly, manufacturers paid attention.
That first laser, built in the summer of 1966, became the first commercially available carbon dioxide laser. Boeing was the first customer. Hobart personally installed the system in January 1967 and, during the process, accidentally burned a hole through a sport coat hanging on a chair. He paid to replace it. The customer stayed happy. Coherent had liftoff. After later receiving capital from the Rockefeller family, the company became publicly owned in 1970.
Fast-forward to the deal that brought Coherent into II–VI’s orbit, and the tone is very different: not laundry rooms and rain gutters, but a full-on public bidding war. Coherent had initially agreed in January 2021 to be acquired by Lumentum for $5.7 billion. Then MKS Instruments and II–VI entered the fray with higher bids.
By mid-March, II–VI put forward an offer valued at $7.01 billion, and Coherent accepted it on March 17. Lumentum came back one last time with a $7.03 billion bid on March 22. Two days later, II–VI publicly said it was holding the line at $7.01 billion. Coherent ultimately stuck with II–VI’s offer and agreed to pay Lumentum the $217.6 million termination fee specified in their merger agreement.
The merger terms were straightforward in concept, heavyweight in size: each share of Coherent common stock converted into the right to receive $220 in cash and 0.91 shares of II–VI common stock.
The strategic rationale was the cleanest part of the story. The combined company would be more evenly distributed across the value chain: materials, components, subsystems, systems, and service. II–VI argued that its scale—especially where materials expertise mattered—was complementary to Coherent’s scale in laser systems. Put the two together, and the pitch was you could serve customers in strategic markets with more of the stack under one roof. With more than 28,000 associates across 130 locations, the company would serve four end markets—industrial, communications, electronics, and instrumentation—which II–VI described as together representing a fast-growing $65 billion total addressable market.
The new company took the Coherent name and traded under the ticker symbol COHR on the Nasdaq. Management again leaned into the dual meaning: not just coherence as a property of laser light, but coherence as “bringing things together”—diversity of thinking distilled into unity of action around mission, vision, and values.
Operationally, the combined business was organized into three segments. Dr. Giovanni Barbarossa continued as chief strategy officer and led the materials segment. Sunny Sun continued to lead the networking segment. And Dr. Mark Sobey—formerly Coherent’s COO and executive vice president—became president of the lasers segment.
The leadership transition that followed set up the company’s current chapter. In November 2021, Francis Kramer moved from chair to Chair Emeritus, and Mattera became chair and CEO. In February 2024, Coherent announced that Mattera would retire and remain in place until a successor was named. In early June 2024, Jim Anderson was named CEO.
Anderson arrived with a resume that read like a signal flare about where Coherent intended to compete next. He had been CEO of Lattice Semiconductor, and before that served as senior vice president and general manager of AMD’s computing and graphics business group. He holds an MBA and a master’s in electrical engineering and computer science from MIT, a master’s in electrical engineering from Purdue, and a bachelor’s in electrical engineering from the University of Minnesota.
His pay package made headlines, too. Anderson was reported as the highest-paid CEO in the U.S. for 2024, with total compensation of $101,497,009, according to an early Equilar study. He was entitled to a $1.06 million base salary, but since he joined in June 2024 he received $81,538 in cash salary. He also received a $500,000 signing bonus, and the bulk of the compensation came as more than $100 million in stock awards. Among the 100 highest-paid CEOs, he was the only one whose compensation crossed nine figures.
For investors, the takeaway isn’t the headline number—it’s what it implies. Bringing in a leader with deep semiconductor and computing experience suggested Coherent saw its future increasingly tied to the same datacenter and infrastructure spending cycles that drive chip companies. In other words: the Coherent name wasn’t just a rebrand. It was a statement that the company planned to be central to the next era of computing hardware—optics included.
VI. The AI Moment: 800G Transceivers and Datacenter Dominance
If you want to understand why Coherent’s stock nearly doubled in 2025, you need to understand transceivers. They’re the small modules that turn electrical signals into light, shoot that light down fiber, then convert it back again—so data can move through a datacenter at mind-bending speeds. When AI workloads surged, these little boxes became the toll booths on the new optical highways. Demand didn’t just rise. It snapped upward.
The speed ladder here is a good proxy for how quickly the world changed. About 20 years ago, the leading edge was 10G. Now, more than half of Coherent’s datacom revenue comes from 200G and higher data-rate transceivers. With AI and machine learning driving the need for bandwidth, 800G transceivers are now shipping in production. And the first 1.6T transceivers are expected to ship in the next few years. Coherent believes that, within five years, the opportunity for 800G and 1.6T transceivers will be larger than all other datacom transceivers combined—largely because AI systems are forcing networks to scale faster than anyone planned. Coherent says it already has a complete portfolio aligned to those AI/ML requirements.
What makes this moment especially powerful for Coherent is that transceivers are not one technology. They’re an assembly of lasers, optics, electronics, packaging, and manufacturing discipline. This is where vertical integration stops being a buzzword and starts being an advantage you can feel.
For shorter links—under 100 meters, including Level 0 interconnects and some Level 1—transceivers typically use VCSELs, or vertical-cavity surface-emitting lasers. These are built on a gallium arsenide (GaAs) platform and are generally the lowest-cost, lowest-power option for short-reach connections, which is why they’re the lasers of choice in that range. Coherent has been seeing significant demand for VCSEL-based transceivers for AI/ML applications. It runs multiple 6-inch GaAs VCSEL fabs in the U.S. and Europe, and it’s among the highest-volume VCSEL manufacturers in the world, shipping into both datacom and consumer applications. Its 100G-per-lane VCSELs are already in production to support 400G and 800G transceivers.
For longer distances—Level 1 switching beyond what VCSELs can support, and for telecom access—the industry shifts to single-mode devices, typically built from indium phosphide (InP). Coherent has multiple InP fabs in the U.S. and Europe, and says its InP platform is one of the few that has been field-proven at scale: more than two hundred million datacom lasers deployed over the last two decades. The company also says those lasers have been qualified and deployed by virtually every network OEM and “web 2.0” player in the world.
CEO Jim Anderson highlighted this AI-driven pull in datacenters alongside the rollout of new optical networking products.
AI datacenter demand remained a core growth engine. Coherent reported that full-year revenue in this area jumped 61% year over year, and Q4 fiscal 2025 grew 38% year over year. During the quarter, Coherent began shipping early batches of its 1.6T transceivers and expected shipments to ramp through the rest of the year. It also advanced its 3.2T transceiver line, built around its own 400G-per-lane EML technology—something it positioned as a key differentiator. In parallel, Coherent scaled up continuous-wave laser output, a key input for co-packaged optics. And it pointed to its vertically integrated indium phosphide manufacturing capacity—now tripled versus the prior year—as a lever to reduce costs and improve supply chain reliability.
The quarter also marked the first revenue from Coherent’s optical circuit switch (OCS) platform, which uses a non-mechanical, digital liquid crystal approach.
Optical circuit switching is another arrow in the quiver for AI datacenters, where traffic patterns and congestion increasingly become performance constraints. Management said OCS represents more than $2 billion in addressable market opportunity over the long run.
And then there’s the signal that turned heads well beyond the photonics world: NVIDIA.
In March 2025, Coherent announced a collaboration with NVIDIA on silicon photonics networking switches using co-packaged optics (CPO), part of the ecosystem NVIDIA introduced at GTC to connect “AI factories” at massive scale. “We are pleased to be NVIDIA’s collaborator on this new transceiver form factor,” Jim Anderson said. “We expect that CPO will further accelerate the expansion of optical networking in the datacenter.” NVIDIA’s Gilad Shainer, senior vice president of Networking, framed the stakes from the other side: “AI factories are growing and networking infrastructure must evolve to keep pace. NVIDIA’s collaboration with innovators, such as Coherent, on silicon photonics will propel the next generation of AI.”
Coherent was also recognized as an NVIDIA Ecosystem Innovation Partner, reinforcing that it wasn’t just selling into the trend—it was helping define the next form factor for how the trend gets built.
A lot of the near-term momentum showed up in the Communications end-market, where revenue was projected to rise 50% to $3.4 billion in 2025. Within Communications, the growth has been driven by datacom demand for high-speed optical transceivers—especially 800G and 400G modules—as hyperscalers raced to upgrade their networks for AI-driven workloads.
“More than 70% of the coherent bandwidth deployed in 2024 was in pluggables, and that percentage will continue growing as AI drives datacenter interconnect bandwidth,” said Dr. Scott Wilkinson, lead analyst at Cignal AI.
VII. Silicon Carbide & Beyond: The Next Platform
While optical transceivers are powering Coherent’s growth today, the company has also been building something that could matter even more in the 2030s: silicon carbide semiconductors for electric vehicles and power electronics.
This didn’t appear overnight. Coherent expanded its silicon carbide position through acquisitions including INNOViON, Ascatron, and a set of patents from GE—moves that steadily broadened its capabilities in SiC substrates.
Silicon carbide is a wide-bandgap semiconductor that can operate at higher voltages, temperatures, and frequencies than traditional silicon. That combination is exactly what you want in power electronics—the hardware that converts and controls electricity in electric vehicles, industrial systems, and renewable energy infrastructure.
Then, in late 2023, Coherent got a major validation of the bet. The company announced that DENSO CORPORATION and Mitsubishi Electric Corporation agreed to invest a combined $1 billion in Coherent’s silicon carbide business, a transaction that came out of the strategic review process Coherent had announced in May 2023. Under the terms, DENSO and Mitsubishi Electric each invested $500 million for a 12.5% non-controlling stake, leaving Coherent with 75%.
Coherent closed the investment in December 2023. To do it, the company separated the silicon carbide business and contributed it into a new subsidiary that would operate the unit going forward.
The partnership wasn’t just capital. As part of the transaction, the silicon carbide business entered into arm’s-length, long-term supply arrangements with DENSO and Mitsubishi Electric to support demand for 150 mm and 200 mm SiC substrates and epitaxial wafers. “We are excited to expand our strategic relationships with DENSO and Mitsubishi Electric to capitalize on the significant demand for silicon carbide,” said Dr. Vincent D. Mattera, Jr., Chair and CEO, Coherent. He described the relationship with two leaders in SiC power devices and modules as “the best path forward” to maximize shareholder value and position the business for long-term growth.
The market backdrop helps explain why. Estimates put the SiC total addressable market at growing from $3 billion in 2022 to $21 billion in 2030—roughly a 28% compound annual growth rate. Coherent framed the deal as an extension of more than two decades of work in SiC materials, and pointed to aggressive recent investment to scale manufacturing of 150 mm and 200 mm substrates. That effort tied back to a longer arc: a 10-year, $1 billion commitment announced in August 2021, supported by capital spending and R&D over the following two years.
From the investors’ side, the logic was straightforward. “Through this strategic relationship with Coherent, we will secure a stable procurement of SiC wafers, which are critical for battery electric vehicles, and contribute to the realization of a carbon-neutral society by promoting the widespread adoption of BEVs in all regions around the world,” said Shinnosuke Hayashi, President & COO, Representative Member of the Board at DENSO. Dr. Masayoshi Takemi, Executive Officer, Group President, Semiconductor & Device for Mitsubishi Electric, said, “We are pleased that this investment has been successfully completed. Going forward, we will further strengthen our collaboration with Coherent, leveraging their capabilities in development and manufacturing of SiC substrates, to achieve solid growth of our SiC power device business and contribute to a more sustainable world through decarbonization.”
The structure here is the part worth lingering on. Coherent didn’t sell the SiC business outright. It set it up as a separately capitalized subsidiary, kept 75% ownership, brought in $1 billion of growth capital from strategic partners, and paired that capital with long-term supply agreements. That’s a tight package: keep the upside, share the funding burden, and turn two investors into committed customers.
VIII. Financial Performance & Market Position
By fiscal 2025, the strategy was showing up cleanly in the results. Revenue hit a record $5.81 billion. GAAP gross margin was 35.2%, but GAAP earnings were weighed down enough that the company posted a net loss of $0.52 per diluted share. On a non-GAAP basis, the picture looked very different: gross margin expanded to 37.9%, and non-GAAP net income reached $3.53 per diluted share.
CFO Sherri Luther put the year in plain terms: “In fiscal 2025, in addition to strong revenue growth, we achieved gross margin expansion of 358 basis points on a year-over-year basis.”
You could also see the momentum building quarter by quarter. In Q1 fiscal 2025, revenue was $1.35 billion with non-GAAP gross margin of 37.7% and non-GAAP EPS of $0.74. Q2 stepped up to $1.43 billion, with non-GAAP gross margin of 38.2% and non-GAAP EPS of $0.95. Q3 climbed again to $1.50 billion, with non-GAAP gross margin of 38.5% and non-GAAP EPS of $0.91.
That third quarter became a good snapshot of what was changing at Coherent. Revenue was up 24% year over year. GAAP gross margin improved to 35.2%, and while GAAP EPS was still a loss at $0.11, it was meaningfully better than the year before. Non-GAAP EPS, at $0.91, was up sharply year over year—evidence that the earnings power of the combined company was starting to show through.
Q4 kept the story going. Revenue came in at $1.53 billion, up 16% year over year and modestly above consensus estimates. The mix mattered: Networking revenue jumped 39% year over year, offsetting declines in Materials and Lasers. Non-GAAP gross margin was 38.1%, up 220 basis points year over year. Non-GAAP operating income rose 35.8% to $275 million, with an operating margin of 18%.
Fiscal 2026 started the way you’d want a new “AI infrastructure” narrative to start: Q1 revenue was $1.58 billion, with GAAP gross margin of 36.6% and GAAP net income of $1.19 per diluted share. And importantly, Coherent said its AI datacenter business grew 23% year over year in the quarter—an early sign that the demand tailwind it had been describing was translating into real dollars.
Under the hood, Coherent now runs as three businesses. Networking sells the transceivers, optics, and semiconductor devices that go into datacenters and communications networks. Materials supplies engineered materials, optics, and specialty components that show up everywhere from industrial systems to advanced electronics. Lasers delivers the excimer, solid-state, and CO2 lasers and systems used in semiconductor equipment, display manufacturing, precision manufacturing, and research. Of the three, Networking is the biggest revenue engine—especially as datacom demand accelerates.
Management has also been shaping the portfolio to fit the moment. In August 2025, Coherent announced a definitive agreement to sell its Aerospace and Defense business to Advent for $400 million. The plan was to use the proceeds to reduce debt, which management said would be immediately accretive to EPS. The unit designs and manufactures optical and laser systems for defense applications and includes about 550 employees across 10 sites.
Debt reduction was already a focus. Luther highlighted progress on that front as well, noting $136 million of debt paydown during the third quarter of fiscal 2025 and reinforcing priorities around cash management while continuing to invest for long-term growth.
Looking ahead, management expected fiscal 2026 to be another growth year. The message was consistent: Coherent believed it was positioned for continued long-term growth by pushing market-leading photonic innovation across its core markets and moving toward the financial targets it had laid out at its Investor Day.
IX. Playbook: Lessons in Roll-up Strategy
Coherent’s journey—from a two-person crystal-growing operation to a roughly $30 billion company increasingly framed as AI infrastructure—offers a surprisingly clear set of lessons for technology investors and operators.
The first is timing. II-VI didn’t acquire like a company grabbing whatever was on sale. It bought as if it had a map of where demand was heading. Finisar, in 2019, planted it directly in the optical plumbing of hyperscale datacenters—right before AI made bandwidth the new bottleneck. The Coherent Inc. merger, completed in 2022, added laser systems depth at a moment when industrial and semiconductor manufacturing were leaning harder into photonics. And the build-out in silicon carbide was underway before the EV surge fully arrived. The common thread is that the biggest moves were made ahead of the market transition, not after it.
The second lesson is that vertical integration can be a real moat when the product is an assembly of many hard-to-source parts. Coherent controls a long stretch of the stack—from engineered materials and semiconductor platforms to devices, optics, packaging, and finished transceivers. That brings practical advantages: tighter quality control, faster iteration between design and manufacturing, and more resilience when supply chains get ugly. When customers want high-speed transceivers, Coherent isn’t simply coordinating a lineup of external vendors for the critical pieces—it can make many of them itself.
Management has been explicit about where it wants that advantage to lead. The company has said it expects continued growth in its core markets and is targeting a long-term gross margin above 42%. CEO Jim Anderson put the positioning in broad terms: “We believe there is no other company with a broader and deeper portfolio of photonic technology, expertise, and innovation.”
Third is integration. None of these deals were easy. Finisar meant fusing a materials-and-components culture with a high-volume transceiver business. The Coherent Inc. merger layered in laser systems with different customers, sales cycles, and operating rhythms. And yet Coherent has repeatedly emphasized hitting synergy goals while continuing to invest in R&D—an indication that integration, for this organization, is a developed muscle rather than a one-off project.
The scale created is massive. The combined company serves industrial, communications, electronics, and instrumentation—markets it has described as together representing a $65 billion total addressable market. It operates with more than 28,000 associates across 130 locations worldwide. That footprint is complex, but it’s also part of the point: it’s what allows the company to build, ship, and support photonics at global scale.
Finally, there’s financial discipline through growth. Coherent used debt to fund major acquisitions, but it has also kept deleveraging on the priority list. And the DENSO and Mitsubishi Electric investment into the silicon carbide unit is a good example of creative financing: bring in strategic capital at the subsidiary level, lock in long-term supply arrangements, and keep most of the upside at the parent. It’s roll-up strategy, but with a capital structure that’s trying not to become the story.
X. Bear vs. Bull Case & Investment Analysis
The Bull Case
The AI infrastructure buildout is still early. Grand View Research estimates the global AI market grows at a roughly 30% CAGR through 2033. The simple implication is that the “plumbing” keeps getting upgraded: bigger GPU clusters need more optical interconnects, and hyperscale datacenters need more transceivers. Coherent sits in a fairly rare position here—one of the few players that can manufacture these components at high volume, with meaningful vertical integration behind it.
Silicon carbide adds a second long-duration tailwind. As EV penetration rises and the broader economy electrifies, demand for SiC power electronics should keep compounding. The DENSO and Mitsubishi Electric partnerships strengthen that story in two ways at once: they inject capital and they also bring long-term supply arrangements that effectively turn strategic investors into committed customers.
And then there’s the operating leverage. Management has talked about gross margin ultimately moving above 40%. If higher-speed products like 1.6T ramp successfully and integration keeps improving efficiency, there’s a credible path to margin expansion on top of growth.
The Bear Case
The biggest business risk is concentration. Coherent disclosed that it generated “more than 10%” of revenue from its largest customer in fiscal 2023 (ended June 30) without naming them. The company’s major customers include Apple, Ford Motor Company, and Applied Materials. In a business where a small number of hyperscalers and device OEMs can account for a lot of demand, losing one major customer—or even just seeing a big customer pause spending—can show up quickly in the numbers.
There are also execution and macro risks that come with being an industrial-scale manufacturer. Supply chain disruptions can still hit production and delivery timelines. Competition—especially from Chinese transceiver players—can pressure pricing. And if the global economy slows, customers can delay datacenter, semiconductor, and industrial capex.
Finally, the cycle cuts both ways. Semiconductor and optical markets are notoriously cyclical. Revenue can fall hard in downturns even if the long-term trend is strong, and Coherent’s expanded fixed-cost base—built through acquisitions—can squeeze margins when volumes drop.
Porter's Five Forces Analysis
Supplier Power: Low. Coherent’s vertical integration lets it produce many of its own semiconductor materials and components, reducing dependence on external suppliers.
Buyer Power: Moderate to High. Hyperscalers, large OEMs, and major auto customers have leverage on pricing and terms. Coherent’s differentiation and ability to deliver at scale offsets this, but it doesn’t eliminate it.
Threat of Substitutes: Low for the core datacenter use case—optical transceivers don’t really have a practical replacement at high speeds. In silicon carbide, substitution risk exists in some applications from gallium nitride.
Threat of New Entry: Low. The capital intensity, specialized know-how, and long customer qualification cycles create real barriers. Replicating Coherent’s integrated footprint would take years and massive investment.
Competitive Rivalry: Moderate. In photonics and lasers, Coherent competes with established companies like IPG Photonics Corp., Lumentum Holdings Inc., and TRUMPF GmbH + Co KG, especially in industrial lasers and precision manufacturing (with IPG particularly strong in high-power fiber lasers). In compound semiconductors and optical networking, rivals include Broadcom Inc., MKS Instruments, and Wolfspeed Inc. Wolfspeed is a focused silicon carbide competitor in a fast-growing category where Coherent is also investing heavily.
Hamilton Helmer's 7 Powers Framework
Scale Economies: Strong. High-volume manufacturing in VCSELs, InP lasers, and SiC substrates can translate into cost advantages.
Network Effects: Weak. This isn’t software; components don’t get more valuable because more people use them.
Counter-Positioning: Moderate. Coherent’s vertical integration is hard for less integrated competitors to match quickly without reshaping their business models.
Switching Costs: Moderate to High. Qualification and reliability requirements in datacenter and automotive markets create stickiness, even if lower-speed optical modules trend toward commoditization.
Branding: Weak. This is largely a B2B components business; brand matters less than performance, supply, and cost.
Cornered Resource: Moderate. Patents, manufacturing know-how, and customer relationships are valuable, but not fully exclusive.
Process Power: Strong. Decades of materials science and manufacturing iteration show up as repeatable execution advantages.
Key Performance Indicators to Track
For fundamental investors, two KPIs matter most:
-
Networking Segment Revenue Growth: This is the clearest scoreboard for the AI infrastructure thesis. Watch year-over-year growth to gauge share and demand as 800G and 1.6T expand, and watch quarter-to-quarter movement for near-term momentum.
-
Non-GAAP Gross Margin: This reflects product mix, pricing pressure, and manufacturing efficiency. Management’s long-term target of exceeding 42% creates a clean reference point. Sustained expansion supports the bull case; compression is an early warning for competition, mix deterioration, or execution issues.
XI. Power & Prediction: What Comes Next
The future of photonics in computing doesn’t stop at pluggable transceivers. As AI models keep scaling and training clusters move toward hundreds of thousands of GPUs, optics may get pulled closer and closer to the compute itself—potentially all the way into the server, as chip-to-chip links. That’s the logic behind co-packaged optics, and it’s why Coherent’s work with NVIDIA matters: it’s a foothold in where the architecture could be going, not just where it is today.
In Coherent’s announcement, CEO Jim Anderson put it plainly: “We expect that CPO will further accelerate the expansion of optical networking in the data center.” From NVIDIA’s side, Gilad Shainer, senior vice president of Networking, said NVIDIA’s collaborations with Lumentum and Coherent on lasers and silicon photonics would help enable the next generation of “million-scale AI.”
Then there’s the longer-dated option value. Quantum computing is still early, but many approaches depend on precision lasers for qubit control and readout. Coherent’s mix of laser systems and materials expertise is a natural fit for that world—though any meaningful revenue contribution is likely still years away.
Meanwhile, the “traditional” markets keep moving, too. EV production growth supports silicon carbide demand. Additive manufacturing continues leaning on industrial lasers. And semiconductor fabrication keeps demanding ever more sophisticated laser systems for lithography and processing. Even as AI steals the spotlight, these end markets remain important parts of what makes Coherent a diversified photonics platform.
On the near-term scoreboard, Coherent guided Q1 fiscal 2026 revenue to a range of $1.46 to $1.60 billion, with non-GAAP EPS expected between $0.93 and $1.13.
Portfolio-shaping is part of the plan as well. The $400 million sale of the Aerospace and Defense business was expected to close in Q3 2025, with proceeds directed toward debt reduction—both improving EPS and sharpening focus. With AI networking accelerating, optical circuit switching starting to contribute, and new transceiver cycles ramping, Coherent entered fiscal 2026 with multiple growth engines and a strategy that’s finally being pulled forward by the market.
XII. Recent News
Coherent opened fiscal 2026 with a strong quarter. On November 5, 2025, it reported first-quarter revenue of $1.58 billion and GAAP gross margin of 36.6%.
Earlier, in August 2025, Coherent announced a definitive agreement to sell its Aerospace and Defense business to Advent for $400 million. The company said it expected to use the proceeds to reduce debt, and that the transaction would be immediately accretive to EPS.
In September 2025, Coherent pushed further into the electronics that power modern optical links, announcing a new family of quad-channel integrated circuits aimed at faster, more efficient optical transceivers for cloud, AI, and telecom networks. Coherent said it has been developing high-speed fiber-optic communication ASICs for more than 25 years, and is leveraging that experience to expand its selected open-market portfolio. The products include a four-channel driver for silicon photonics Mach-Zehnder modulators designed for 800G and 1.6T pluggable modules, along with a chipset for 400G ZR/ZR+ coherent optical links. Coherent positioned the lineup around two key benefits: industry-leading performance and lower power consumption—helping operators scale capacity while reducing energy use.
That same month, Coherent said it would showcase its next-generation optical communications innovations at ECOC 2025, held September 29 to October 1 at the Bella Center in Copenhagen, Denmark, with company leaders participating in the event’s Market Focus panels, workshops, and roundtable sessions.
Coherent also picked up external validation for its pace of product development at OFC 2025, where it received six Product Innovation Awards recognizing work across components, modules, and systems for datacenter and communications markets.
And in March 2025, Coherent announced general availability of its 800G ZR/ZR+ transceiver in a QSFP-DD form factor—aimed at growing demand for high-speed, efficient, scalable metro, regional, and data center interconnect applications.
On the Street, Morgan Stanley maintained coverage with an Equal-Weight recommendation in December 2025, while JP Morgan maintained an Overweight recommendation that same month.
Spotify
Apple Podcasts
Amazon Music
Audible
YouTube
Castbox
Pocket Casts