
Coherent vs Lumentum Stock: Who Has More AI Upside?
Artificial intelligence (AI) data centers need far more than powerful GPUs. They also need fast optical connections to move huge volumes of data between servers without creating traffic jams.
That is why Coherent vs Lumentum stock has become a serious comparison for investors seeking AI infrastructure exposure. Coherent is a diversified photonics company with greater scale, while Lumentum remains more focused on AI optics. The better fit depends on whether you value focused growth or a more diversified technology company.
Key Takeaways
Lumentum offers greater near-term AI upside because its revenue growth and exposure to high-speed datacom transceivers, EML chips, and networking components are currently stronger.
Coherent is the larger and more diversified business, with exposure to optical communications, industrial lasers, engineered materials, silicon carbide, and other photonics markets.
Lumentum's focused growth comes with greater risks, including customer concentration, inventory swings, supply-chain bottlenecks, and a sharper slowdown if AI networking demand weakens.
Investors should compare revenue growth, margins, free cash flow, debt, and consistently defined forward earnings multiples rather than relying on forward PE alone.
The better stock depends on portfolio role and risk tolerance: Lumentum may suit investors seeking higher potential upside, while Coherent may offer a more balanced way to own AI optics.
Coherent vs Lumentum stock: The short answer
Lumentum has greater near-term upside potential if AI networking demand stays hot. Its revenue is rising much faster, and its business has direct exposure to high-speed transceivers and laser chips used by hyperscale computing customers.
However, that upside comes with more risk. Lumentum is smaller, more reliant on a concentrated group of customers, and more exposed to sudden shifts in networking demand.
Coherent is the steadier of the two businesses. It is much larger, sells into more markets, and spans lasers, networking components, optical materials, and industrial applications. Its broader materials exposure includes silicon carbide, though that is a diversification benefit rather than the main current revenue driver. This mix can soften the impact when one end market cools.
For a long-term investor, Coherent may offer a more balanced way to own AI optics. Lumentum may offer more explosive returns and greater market upside if the current demand cycle continues, but its stock can also move harder in the other direction.
A single optics stock should rarely carry an entire portfolio. A wider current investment portfolio can put high-growth technology shares in a more sensible context.
Why AI data centers need optical communications
Training and running AI models requires thousands of processors to exchange information at extreme speed. Copper cables work over short distances, but they lose efficiency and create heat as bandwidth requirements rise.
Optical links use lasers to turn electrical data into light, send it through fiber, then convert it back into electrical signals. That is why 800G datacom transceivers have become so important. These modules move up to 800 gigabits per second between processors, switches, and servers, helping GPU clusters avoid slower connections.
Data-center operators are moving toward 800G and eventually 1.6T connections because AI workloads create more traffic inside the data center, not only between users and the cloud. A widely cited review of optical data-center technologies explains why electrical interconnects become harder to scale as speed and distance increase.
VCSELs, EMLs, and silicon photonics are not interchangeable
Both companies manufacture laser technology, including vertical-cavity surface-emitting lasers, or VCSELs. These tiny lasers emit light straight upward from the chip surface. They are useful in short-range optical links and consumer 3D-sensing systems, including facial authentication on Apple devices. That Apple use case differs from hyperscale optical networking, so consumer 3D-sensing demand shouldn't automatically count as AI datacom demand.
Yet investors should avoid lumping every laser into one AI category. Many long-distance, high-speed datacom applications use electro-absorption modulated lasers, known as EMLs, rather than the short-range emitters described above. Lumentum's recent component growth has been tied to strong shipments of EML chips, pump lasers, and other networking parts.
Silicon photonics could change the mix again. It combines optical functions with silicon-based manufacturing techniques, allowing smaller and more densely integrated optical components. Research into silicon photonics for terabit communications points to its role in pushing network speeds higher while keeping power use in check.
Coherent has more scale, Lumentum has sharper exposure
Coherent took its current form after II-VI acquired the original Coherent and adopted its name. The combined company is now a large photonics company with significant scale.
Its product range is broad. Coherent sells optical transceivers and components for communications, lasers for industrial manufacturing, and engineered materials used in semiconductors and electronics. Its experience with materials such as indium phosphide, gallium arsenide, and silicon carbide gives it technical depth that Lumentum does not match. Consumer 3D-sensing applications associated with Apple form a separate market, not evidence of a current customer relationship.
Silicon carbide has mattered because it is a wide-bandgap semiconductor used in high-voltage power applications, especially electric vehicles and industrial equipment. Coherent's materials history gives it another route into semiconductor spending. Investors should separate this silicon carbide exposure from current AI-optics revenue.
Lumentum's more focused laser business sells the laser chips, tunable components, and transceiver-related products used in telecom networks and cloud data centers. That focus can be powerful when demand rises fast. It also leaves less room to hide when customers pause orders.
Photo by Brett Sayles
Co-packaged optics is another long-term issue to watch. Instead of placing pluggable optics at a switch edge, this approach puts optical engines closer to the networking chip. That could improve bandwidth density and reduce power use, but heat management and servicing remain difficult. Silicon photonics may enable higher-density optical engines, while packaging, yields, and customer qualification remain decisive.
The data-center silicon photonics market is already developing around these design choices. Coherent and its focused peer both have a place in that shift, though the eventual winners may depend on packaging, yields, and customer qualification.
Revenue growth favors Lumentum, but Coherent remains larger
Recent financial reports show why revenue growth and absolute scale tell different stories. Coherent's revenue base is much larger, while Lumentum's growth rate is far higher.
Metric | Coherent | Lumentum |
|---|---|---|
Fiscal 2025 Revenue | $5.81 billion | $1.645 billion |
Fiscal 2025 Revenue Growth | About 23% | About 21% |
Fiscal Q3 2026 Revenue | $1.81 billion | $808.4 million |
Fiscal Q3 2026 Revenue Growth | 21% year over year | 90% year over year |
Coherent's size means it needs far more new sales to post the same growth percentage as Lumentum. The 90% quarterly increase remains hard to ignore for Lumentum. It reflects aggressive hyperscale demand for datacom transceivers after a weak networking cycle.
Coherent reported $1.69 billion of revenue in fiscal Q2 2026, up 17% year over year. On a pro forma basis, after adjusting for its aerospace and defense sale, growth was 22%. Its official fiscal second-quarter release also showed AI-related communications demand as a major driver. The diversified mix means this isn't a pure AI-optics or silicon carbide story.
The company reported $665.5 million in fiscal Q2 2026 revenue, up 65.5% year over year. Its second-quarter results also showed a major improvement in non-GAAP operating margin.
Fast percentage growth does not automatically mean lower risk, especially when it follows a depressed networking cycle.
Revenue CAGR, or compound annual growth rate, matters more than one strong quarter. A two- or three-year view can show whether current AI spending supports a durable silicon photonics infrastructure cycle or merely filled a backlog.
Customer concentration and supply chain friction remain real
The AI optics story has risks that can disappear in a bullish earnings call.
Lumentum's fiscal 2023 filings showed that one customer accounted for 30% of revenue. Coherent also disclosed customers that crossed its 10% revenue threshold. A large buyer can quickly change supplier results by altering a product design, cutting inventory, or delaying a new infrastructure build. This risk spans end markets, with Apple serving only as a generic example, not a disclosed customer of either company.
Older articles often mention the company's earlier decline. That fall came during a severe telecom inventory correction, when customers stopped ordering parts they had already stockpiled. Recent AI-driven results have improved the picture, so that old drawdown should not be treated as the company's current trend.
There are also bottlenecks behind the finished transceiver, including qualified indium phosphide wafers, laser fabrication, and reliable foundry capacity. Optical communications depend on specialized semiconductor equipment for wafer processing, packaging, testing, and metrology. Tower Semiconductor has silicon photonics foundry capabilities; Axcelis Technologies supplies ion-implantation equipment, and Onto Innovation provides inspection and metrology tools.
These businesses may benefit from higher semiconductor spending, but no direct relationship with either company should be assumed. Axcelis Technologies and Onto Innovation illustrate the broader supplier ecosystem, where semiconductor equipment serves several markets. Silicon carbide has separate demand drivers, so it should not automatically enter optical-transceiver forecasts; the supply chain is connected, but not a simple one-to-one read-through.
Valuation needs more than a forward PE
A forward PE compares today's stock price with expected earnings per share over the next year. If a $100 share is expected to earn $5 per share, the multiple is 20.
That simple calculation can mislead when earnings are recovering from a downturn. Lumentum's earnings expectations have risen sharply alongside its sales. Coherent also reports a meaningful gap between GAAP and non-GAAP earnings, so a forward PE is meaningful only when both companies use the same earnings definition and valuation date. Otherwise, comparing adjusted earnings with GAAP figures can make a forward PE look like a false bargain and widen the valuation gap.
Use the same-date stock price and the same earnings definition for both companies. Then compare:
Sales trends, gross margin, free cash flow, and debt alongside forward PE.
Revenue CAGR to see whether growth has lasted beyond one AI spending burst.
A PEG ratio only when earnings estimates are credible and positive.
Management guidance, because forecast changes can move either stock within hours.
This is a technology comparison, not a quantum computing bet. That distinction matters when the stock market starts putting every fast-growing tech company into the same AI basket.
If you are learning how to invest, position size matters more than finding a perfect entry point. A compound interest calculator can show how consistent contributions often matter more than chasing a volatile move.
For those deciding how to buy stocks, brokers such as Trading 212 and eToro make access easy, but access is not research. If eToro is your chosen platform and you have made your own decision, Buy now on etoro.
Frequently Asked Questions
Which stock has more AI upside, Coherent or Lumentum?
Lumentum currently has more near-term upside potential because its revenue growth and direct exposure to AI networking components are stronger. That potential comes with higher volatility and greater sensitivity to customer orders and networking demand.
Is Coherent a safer investment than Lumentum?
Coherent is generally the steadier business because it is larger and operates across more photonics, materials, industrial, and communications markets. Diversification can reduce the impact of a slowdown in one area, although it does not eliminate stock-market or execution risk.
How are Coherent and Lumentum exposed to AI data centers?
Both companies supply optical technologies used to move data between servers, switches, and processors in data centers. Lumentum has sharper exposure to transceivers, EML chips, and related laser components, while Coherent combines AI optics with a broader product portfolio.
Should investors compare the forward PE ratios of these stocks?
They can be useful, but only when the stock prices, valuation dates, and earnings definitions are consistent. Investors should also examine revenue growth, margins, free cash flow, debt, guidance, and whether current earnings reflect a durable cycle or a recovery from weak demand.
What are the main risks of investing in AI optics stocks?
The key risks include customer concentration, inventory corrections, supply-chain constraints, technology changes, and a slowdown in hyperscale data-center spending. Co-packaged optics and silicon photonics may create opportunities, but adoption depends on cost, power use, packaging, yields, and customer qualification.
Final view on AI optics upside
Lumentum has the stronger current growth profile and greater upside if AI networking demand remains intense. That focus also increases its exposure to customer concentration, inventory swings, and any slowdown in transceiver demand.
Coherent offers a broader business, larger revenue base, and more ways to benefit from photonics demand. The better choice depends on volatility tolerance, portfolio role, and perceived investment potential. Monitor quarterly growth, margins, and forward PE, while tracking silicon photonics and co-packaged optics adoption as technology variables.












































Comments