Picks Behind the Picks: Credo, Ciena, AXT & VIAVI — Four High-Conviction Upstream Photonics Names and Their Risks

TL;DR These four are the picks behind the picks — the companies that supply the suppliers. Credo tripled revenue in six quarters ($135M→$437M), and Ciena's backlog jumped $2B in 90 days to about $7B. AXT is one of only a handful growing the indium phosphide wafers every optical laser is built on, and VIAVI is the arms dealer whose test gear every laser and transceiver has to pass through. The growth is explosive — but each carries a very specific risk.

This is my favorite part — the picks behind the picks, the companies that supply the suppliers, plus a name like Credo that straddles the connection layer. The growth is explosive across the board, but the risks are sharply different name to name. Let's take them one at a time.

1. Credo Technology: the bridge between copper and light

Credo is the bridge between the old world and the new.

Its proprietary low-power chip technology squeezes every last inch out of the copper inside the rack, while it also builds the optical chips and cables that take over once the signal has to travel further. It just bought a silicon photonics company to round out an end-to-end stack running all the way to 1.6 terabit, and it now ships to all five of the largest US hyperscalers.

The growth is a little bonkers: revenue went from $135 million a quarter to $437 million in just six quarters — more than three times over. And here's the tell that there's a real chip company hiding inside this cable maker: gross margin runs about 68%. That's the kind of number you mostly see from software, not hardware. Operating margin nearly doubled to 37% as it scaled, and it's guided for over 80% more revenue growth next year.

But the risk is real and specific. Even though it ships to all five, just three of those customers make up 88% of revenue. If a single hyperscaler pauses its orders, the stock gets cut hard. Insiders have also been selling heavily into the run, and at about 35 times sales, it's priced for everything to keep going right. With a PEG near 1, the growth justifies a lot — but this is a high-conviction, sharp-pullback kind of name.

2. Ciena: the cheat code that carries more over existing fiber

Ciena is the Western leader in coherent optics.

Its proprietary WaveLogic technology was the first in the world to push 1.6 terabit down a single wavelength of light — the kind of cheat code that lets existing fiber carry far more without digging up new cable. That one product has already won 49 customers in just two quarters.

The relationships matter just as much. Its systems have already been adopted by three of the four largest hyperscalers, and cloud providers now make up almost half of everything it sells. Here's the number that matters most to me: last quarter its backlog jumped about $2 billion in 90 days to roughly $7 billion. Nearly all of it is scheduled to ship next year — more than a full year of revenue already booked. And as revenue hit a record, up 40%, operating margin doubled from under 8% to over 15%.

The item to respect here is valuation, because the market already knows all of this. The stock trades at around 120 times forward earnings. It's priced for absolute perfection.

3. AXT: growing the crystal every laser sits on

AXT is the most upstream name in this whole story.

It's one of only a handful of companies on Earth that grows indium phosphide wafers — the special crystal every single optical laser has to be built on. You can't just spin up a crystal-growing operation overnight, which is about as close to the definition of a moat as it gets. Its backlog for this material just hit a record above $100 million, as every laser maker scrambles for supply.

On risk, though, this company sits near the top of the list, and the risk is very specific. Almost all of AXT's manufacturing is in China, and China now requires a government export permit for every single order — which already caused it to come up short on revenue last quarter, even with that record backlog sitting right there. On top of that, it raised about $550 million that diluted shareholders, and insiders have net sold over $70 million into the run. The company still loses money, though the losses are shrinking and gross margins have climbed from 17% to about 30%. The story and the bottleneck are very real, but at roughly 66 times sales, this is a small, speculative lottery-ticket position.

4. VIAVI Solutions: the arms dealer of the optical world

VIAVI is the arms dealer of the optical world.

Every fiber link, transceiver, and system has to be tested before it goes live and monitored once it's running — and VIAVI makes the gear all of them have to pass through. The beauty of that position is that it's pick-agnostic. It doesn't matter which laser or transceiver wins; they all get tested on VIAVI's equipment. Honestly, this is exactly the kind of setup I love to find.

For years, VIAVI's revenue basically went nowhere, stuck around $285 million a quarter. Then the AI build-out hit and its network testing business grew over 54% as data centers raced to validate all that new optical gear. Revenue jumped to over $400 million a quarter, and operating margin climbed back into the double digits. The business also gets a high-margin kicker from a second, quieter division that makes the anti-counterfeiting coatings printed on the world's currency.

The caution here is that insiders, including the CEO, have been selling steadily into the move. At a PEG around 1.4, it isn't outrageously priced, but like the rest of the list it has already run hard — so I'd wait for it to cool off a little before stepping in.

The four at a glance

Name	Role	Growth	Key risk	Valuation
Credo	Copper↔light bridge, low-power chips & cables	Revenue 3x in six quarters, 68% gross margin	3 customers = 88% of revenue, insider selling	P/S ~35x, PEG ~1
Ciena	Coherent optics systems, 1.6T WaveLogic	Backlog +$2B in 90 days to ~$7B	Priced for perfection	Forward P/E ~120x
AXT	Indium phosphide wafers (most upstream)	Backlog record above $100M	China production & export permits, dilution, losses	P/S ~66x
VIAVI	Optical test gear (arms dealer)	Network testing +54%, revenue $400M+	CEO and insiders selling steadily	PEG ~1.4

My take

I see these four not as one basket but as a risk spectrum.

VIAVI is the most defensive as a pick-agnostic arms dealer; next is Ciena, priced for perfection; then Credo, with clear customer-concentration risk; and AXT is the most speculative, stacking China production, export permits, losses, and dilution. The bottleneck and the story are real for all four. But all four have already run hard, so I plan to wait for pullbacks and size each one to its own level of risk. (I'm not a financial advisor — this is for educational purposes only.)

FAQ

Q: Which of the four is safest? A: I'd say VIAVI. Because every laser and transceiver gets tested on its gear regardless of who wins, it's pick-agnostic. Just factor in that insiders, including the CEO, have been selling steadily.

Q: What's AXT's biggest risk? A: That almost all of its manufacturing is in China, and China now requires a government export permit for every order — which already caused a revenue shortfall last quarter despite a record backlog. Add dilution from a ~$550M raise and ongoing losses, and it's best treated as a small, speculative position.

Q: How serious is Credo's customer-concentration risk? A: It ships to all five major hyperscalers, but three of them make up 88% of revenue. A single one pausing orders could cut the stock hard. At about 35 times sales, any break in the 'everything keeps going right' assumption stings.