Lumentum’s Quarter And The AI Plumbing Trade
What an optical networking winner says about the real bottlenecks in AI
The AI Plumbing Trade
If you’ve followed the AI trade over the last year, you’ve probably seen plenty of charts of GPUs and cloud stocks. What gets less attention is the plumbing that has to be built to move all of that data around—inside data centers, between data centers, and out to end users.
That’s where companies like Lumentum (LITE 0.00%↑) come in.
On Tuesday, Lumentum reported a quarter that looked, from an AI-infrastructure perspective, about as clean as you could ask for:
Fiscal Q2 revenue came in around $665.5 million, up roughly 65% year-over-year and ahead of Wall Street’s estimates.
Non-GAAP EPS was about $1.67, versus consensus closer to $1.40.
Management guided for Q3 revenue and earnings above prior expectations, explicitly highlighting demand tied to cloud and AI build-outs.
The market’s initial reaction after hours was exactly what you’d expect when a key piece of the AI supply chain overdelivers: the stock spiked sharply higher into the high-$470s to $480 area before settling back somewhat as traders digested the guidance and the recent run.
That’s the “what.” The more interesting question, at least for me, is how you end up in front of a name like this before the obvious beat, and how you structure the trade so you’re not just buying common stock and hoping.
That’s where the rest of this story comes in.
Step 1: Let the system find the hunting ground
If you’ve been reading me for a while, you know I don’t start with a story and go hunting for a stock to match it. I do the opposite:
Start with the numbers.
Every trading day, our system ranks thousands of names by estimated 6-month return after factoring in the cost of hedging them.Focus on the very top.
I look at the top ten names (“top names”) and then decide which ones to actually trade, and how.
Since I launched this Substack at the end of 2022, I’ve been tracking each weekly top-ten cohort for six months. Across 136 rolling 6-month periods, those top-ten baskets have averaged 20.27% over the next six months, versus about 10.22% for the market-tracking SPDR S&P 500 Trust (SPY 0.00%↑).
So when a theme keeps popping up in the top ten, I pay attention. Optical networking / photonics started doing that well before it was fashionable to talk about “AI plumbing” on financial TV. Lumentum was one of the names that kept bubbling up there.
In other words, the model led us to the optical space first, and then I went to work on structure.
Step 2: Why this structure on LITE?
On January 14th, I sent out this Top Names trade in Lumentum:
The stock is Lumentum Holdings (LITE), and our trade is a combo consisting of these four legs:
Buying the June 18th, 2026 $460 calls,
Selling the June 18th, 2026 $480 calls,
Selling the February 6th, 2026 $345 puts, and
Buying the February 6th, 2026 $340 puts,
For a max net debit of $2.50. The max loss per 1-contract combo is $750, and the max gain is $1,721. This trade filled at $2.10.
A few key design choices there:
1️⃣ Defined upside: call spread, not a naked call
On the day we entered, the June $460 call by itself was trading around $55.83 per contract. That’s a big upfront outlay for a single line of upside.
Turning that into a $460–$480 call spread cut the gross cost of the upside leg down to about $4.82 per contract, while still giving us $20 of potential upside between the two strikes.
That still gives us plenty of upside if the AI optical story keeps playing out, but with a defined maximum payout per combo.
For readers with larger accounts, that “max gain of $1,721” is per 1-contract package. If you want more exposure, you scale the number of packages, not the risk per package.
2️⃣ Financing the upside with near-term fear
We paired the June call spread with a short February put spread:
Short $345 put, long $340 put, same February expiration.
The premium from that put spread helped finance the long-dated call spread, bringing the net entry cost for the whole four-leg package down to our actual fill of $2.10 per combo.
In other words, instead of paying roughly $55.83 for a single long call, we used structure to get a defined-risk package that participates in the same theme for just over $2 per combo. That’s the kind of leverage you can get from intelligent trade construction.
Structurally, this is our standard hybrid 4-leg combo:
Long-dated upside exposure (the June call spread),
Short-dated put spread “floor” that pays us while we wait,
All risk defined at entry.
3️⃣ Aligning with our 6-month horizon
June 2026 isn’t a random date. Our whole system is aimed at the next six months, so when we use longer-dated calls or call spreads, it’s usually:
Out around six months, sometimes a bit longer for slower-developing theses.
Far enough to give the story time to play out,
Close enough that the options still meaningfully respond to catalysts along the way—like this week’s earnings.
Step 3: What this week’s earnings did to the payoff profile
Going into the print, our position looked like this in simple terms:
If LITE disappointed badly and traded back toward the mid-300s in the near term, the Feb put spread would come into play, but the loss would be capped, and we’d still own a June call spread that could recover if the longer-term AI optical story was intact.
If LITE delivered, the first thing to move would be the short-dated options—that February put spread would deflate quickly, while the June call spread would gain value as the market priced in higher odds of LITE spending time above our 460 strike over the next few months.
What actually happened:
Lumentum crushed expectations on both revenue and earnings, with revenue up about 65% versus a year ago and non-GAAP EPS nearly 4x what it earned in the same quarter a year earlier.
Management didn’t just beat the quarter; they raised guidance on revenue and earnings for the next quarter in a way that clearly tied to AI infrastructure demand.
The stock spiked after hours, trading up into the high-$470s to $480 area before settling back somewhat as liquidity thinned out in the post-market.
From the trade’s perspective:
That after-hours spike put us in a world where our $460–$480 call spread was effectively “in play”—a path to near-max payout came into view much faster than usual.
Even if there’s some give-back when regular trading resumes, LITE is now much further above our short put strikes (345/340) and closer to our June call strikes (460/480) than it was when we entered the trade in mid-January.
We haven’t “won” the trade yet—there’s still time and volatility between now and June—but:
The downside risk remains capped at the original $750 per combo,
The earnings reaction validated the thesis that optical networking is a real beneficiary of the AI build-out,
And the whole package is doing what it was designed to do: give us a levered, defined-risk way to express a system-driven idea in a sector the model was already pointing to.
Why this matters beyond Lumentum
I’m not writing this to get you excited about one ticker. The point is the process:
Let the model do the heavy lifting.
The “top names” engine sifts through thousands of securities and identifies where the best 6-month risk-reward has been, empirically. Since late 2022, that’s meant roughly double the S&P 500’s average 6-month return for those top-ten baskets.Use that as a radar for themes.
When similar kinds of names—optical networking, shipping, small-cap biotech, whatever—start showing up repeatedly at the top of that list, that’s usually a hint that something is changing under the surface before the narrative catches up.Express the idea with structures, not just shares.
Four-leg combos like the LITE trade let us:Cap downside in dollar terms per contract,
Use short-dated options to finance longer-dated exposure,
And scale position size up or down depending on account size without changing the per-contract economics.
Lumentum’s quarter is a nice case study where all three layers lined up:
System said: optical / photonics names belong near the top of the list.
Theme said: AI build-out isn’t just about GPUs; somebody has to build the optical highways.
Structure said: we can take that bet in a way that survives being early, benefits from being right, and doesn’t blow up the account if we’re wrong.
That’s what I’m trying to do with every trade alert: not just pick a name, but encode a repeatable way of taking risk when the numbers and the story rhyme.