AI - Is the Brain, Lumentum Is the Laser: How Micron, Coherent, and Friends Quietly Eat

The AI Diet: High in fiber, heavy on memory, and surprisingly low in copper. Finally, a plan where the only thing losing weight is the latency in your portfolio

Everyone’s staring at the flashy AI chips like they’re rock‑star brain surgeons. But those chips only shine if the behind‑the‑scenes plumbing — lasers, memory, cables, and connectors — can keep up.

That’s where companies like Lumentum, Coherent, Micron, Seagate, and others come in: they build the stuff that lets AI actually move data around at insane speed.

Why Fiber-Optic “Plumbing” Matters

Old‑school copper connections are starting to wheeze under AI‑level traffic.

New fiber‑optic systems move data with light instead of just electricity, using lasers zipping through glass fibers. That’s perfect for giant AI data canters that need to move oceans of data quickly and efficiently.

So instead of only chasing the obvious AI chip names, this approach says: also look at the companies selling the glass, lasers, memory, and connectors that those chips depend on.

1. Lumentum (LITE): Lasers With a Waiting List

What they do: Lumentum sells lasers and optical gear that help AI data canters move data at high speed over fiber. Think of them as upgrading the internet’s plumbing from garden hoses to fire hoses.

How their numbers exploded recently:

• In a recent fiscal quarter, Lumentum reported record revenue of around $665–800 million, up on the order of 60–90% year‑over‑year, driven largely by AI data center demand for optical gear.
• Earnings per share also beat Wall Street expectations, showing that this isn’t just “top‑line hype” — they’re actually making good money on this surge.

Demand and backlog:

• Management has talked about demand outpacing supply by roughly a third, even after they expanded capacity
• Their backlog (orders already in hand) for things like optical circuit switches and high‑power lasers has been described as hundreds of millions of dollars, including fresh, large, multi‑hundred‑million orders tied to AI networking gear

Translation: Customers are lining up; Lumentum is racing to build enough lasers to keep everyone happy

2. Coherent (COHR): Photonics Workhorse for AI

What they do: Coherent is a big player in photonics — lasers, optical components, and related materials that show up inside high‑speed networks and AI infrastructure.

Recent earnings pop:

• In its latest reported quarter, Coherent delivered revenue of roughly $1.8 billion, up about 20% year‑over‑year, and came in above analyst estimates
• Earnings per share also beat expectations, signaling that AI‑related demand is not just noisy headlines; it’s flowing into the income statement.

Demand and backlog:

• Coherent has highlighted strong bookings in AI‑related optical products like 800G and 1.6T transceivers used in high‑speed data center links
• They’ve discussed unmet backlog carrying forward from one quarter into the next, meaning orders were strong enough that not everything could be shipped on time and rolled into upcoming quarters.

In other words, their problem isn’t “where are the customers?” — it’s “how fast can we build the stuff they’ve already ordered?”

3. Micron (MU): Memory King of the AI Buffet

What they do: Micron makes memory chips, including high‑bandwidth memory (HBM) that sits very close to the AI chips and feeds them data at ridiculous speed.

Think of Micron as the short‑term memory and high‑speed snack bar for AI systems.

Recent earnings explosion:

• In a recent fiscal quarter, Micron reported around $13.6 billion in revenue, with strong margins and earnings per share that impressed the Street.
• Analysts highlighted that revenue, margin, and free cash flow all jumped meaningfully as AI‑driven memory demand accelerated.

Demand and backlog:

• Micron has said its HBM production for 2026 is essentially sold out under binding contracts, which is like having your entire restaurant fully booked for the year in advance.
• They’ve also pointed to an overall AI memory supply crunch, with data center and cloud customers driving very strong demand and giving Micron unusually clear visibility into future revenue from these segments.

Translation: AI customers are basically saying, “Take my money, just give me the memory.”

4. Seagate (STX): Long-Term Memory for the AI World

What they do: Seagate makes storage drives, especially for enterprise and cloud data centers. If Micron is the short‑term memory, Seagate is part of the long‑term archive where all the training data, logs, and backups live.

Earnings and demand flavor:

• Recent results and commentary around Seagate have emphasized strengthening demand from cloud and AI data center customers as storage needs grow with AI workloads.
• As AI models get bigger and datasets keep multiplying, the “file cabinet” side of the business (high‑capacity drives) continues to matter more and more for these infrastructure players.

Seagate’s story is less about explosive spikes and more about steadily rising storage needs in a world drowning in data.

5. The Connector Nerd: Amphenol (APH)

What they do: Amphenol makes connectors and cables — the plugs, sockets, and little hardware bits that physically hook all this gear together in data centers.

It’s the incredibly unsexy but absolutely essential Lego clips of the AI ecosystem.

Earnings and demand backdrop:

• As hyperscale and AI data centers expand, Amphenol benefits from higher volumes of high‑speed connectors and interconnect systems, which has supported steady revenue and earnings growth over time.
• Their business model tends to produce more gradual, stair‑step gains rather than wild AI‑style spikes, but the underlying driver — more data, more hardware, more connections — remains strong.

This is the tortoise in the race: not flashy, but it keeps plodding upward as long as the world keeps plugging more stuff into racks.

6. The Moonshot: Poet Technologies (POET)

What they do: Poet Technologies is the little guy with a big idea: an “optical interposer” that helps integrate optical components (like lasers) with chips more tightly and efficiently.

Plain English: They’re trying to put the laser closer to the brain, shrinking the gap between optics and electronics so data can move faster and cheaper.

Recent momentum:

• Poet doesn’t have the massive revenue numbers of Lumentum, Coherent, or Micron yet — it’s more about technology progress, partnerships, and design wins than huge earnings beats.
• The upside thesis is that if their optical interposer becomes standard in certain AI or networking applications, they could go from “tiny science project” to “very attractive takeover target” for a bigger player.

It’s a classic binary bet: boat or lesson.

Putting It All Together

Here’s how to think about this lineup in ultra‑simple terms:

• Lumentum & Coherent: Their earnings are popping because AI data centers are scrambling for fast optical links, and both companies are dealing with strong orders and meaningful backlogs in their photonics businesses.
• Micron: Earnings and margins jumped as AI memory demand exploded; its high‑bandwidth memory is sold out for 2026, giving it a very strong demand pipeline.
• Seagate: Benefits from the slow‑and‑steady need for more storage as AI eats more and more data.
• Amphenol: Quietly grows as every new rack of AI hardware needs more connectors and cables.
• Poet Technologies: A small, speculative play that could pay off big if its optical integration tech becomes widely used.

The theme: Instead of only betting on the AI “brains,” you’re also looking at the companies selling the plumbing that those brains absolutely can’t live without.

Disclaimer

This article is for informational purposes only and does not constitute financial, investment, legal, or tax advice. It is not a recommendation or solicitation to buy or sell any securities, strategies, or financial instruments. All investing involves risk, including the possible loss of principal. Company names and examples are provided solely for educational illustration. You are solely responsible for evaluating your own financial situation and investment objectives, and you should conduct your own due diligence and/or consult with a qualified financial professional before making any investment decisions.