AI Optics Earnings Start Nov 4th: POET Runs Anchor Nov 16

The AI optics prints start next week. Coherent $COHR reports November 4. Lumentum $LITE and Applied Optoelectronics $AAOI go November 5. And the group walks into earnings season off a drawdown the three-month charts hide: Coherent sits 45% below its June record close, AAOI is 67% off its high, and Lumentum is still 24% under its October peak after a two-day bounce. The small caps got it worse. Aeva $AEVA and Lightwave Logic $LWLG are down about two-thirds from their highs.

POET Technologies POET Technologies (POET) reports last. Monday, November 16, 7:00 a.m. ET, before the open, and there's no conference call behind it. The first regular call are intended to come with the year-end numbers on March 1, 2027. So for one more print, the press release is the whole event. The stock walks in at $5.69, about 73% below its May high, with $796.3 million of cash and short-term investments at June 30. Here's what we want that document to say, and why the tape isn't in the mood to wait for March.

How we got here is not an isolated case

Rewind to spring. In late April the Company said a purchase order had been cancelled, and the stock took the hit. On May 14 the Lumilens deal landed: a $50 million initial order and a framework that could run past $500 million over five years. The stock printed $20.81 that day. The next morning came a $400 million raise, priced at a premium to market, and the air started leaking out. June brought the AGM. July brought nothing, and the stock ground from near $10 to the mid-$6s on no company news at all.

Then August 13. POET put its Q2 release out pre-market, led with six straight quarters of sequential growth and $796 million in cash, and got its first green opening reaction in four prints. That was the template working. Since then the Company has been heads down on the ramp, the September update arrived as a recorded CEO interview on the website, and the stock drifted again. It sits at $5.69 now, under every moving average on the chart.

The sector's problem is the opposite of POET's. Lumentum's CEO says the company is sold out through 2029, and the stocks sold off anyway on AI-funding nerves. Nvidia Nvidia (NVDA) boss Jensen Huang stood up in Berlin last week and told the room the next leg of the buildout runs on light. So the order books say unicorns & rainbows and the charts say horror story, for the whole group. $POET has the horror-story chart without the sold-out order book yet. Closing that gap is what November 16 is about.

The calendar: POET goes last

By the time POET's release crosses the wire on the 16th, Coherent, Lumentum and AAOI will have been through their calls, and the market will know whether the laser and optics shortage is still real. Lumentum already says it is. That leaves one question for POET: has it started shipping into that shortage? Nobody is claiming it has. The Company's own timeline says qualification units this year and volume next. The sector testifies first, POET takes the stand last, and there's no cross-examination.

What Jensen built, and the part POET makes

Strip the Berlin keynote down to hardware. Nvidia's Spectrum-X switch is in full production with co-packaged optics: silicon photonics bonded straight onto the switch chip, with lasers feeding light up into it. Huang's line on stage was that scale-up, the links inside the rack, is going to become optical. Add scale-up to the optical domain and the volume, as he put it to SDxCentral, "is going to be, well, 15x." Nvidia said it's scaling optics manufacturing capacity fifteen-fold to match.

Here's the part that matters for $POET. Silicon moves light; it can't make it. So in Nvidia's design the laser doesn't live inside the package. It sits in an external laser source, a high-power, field-replaceable module on the front panel that feeds several photonic engines at once. That makes the light source its own product category with its own supply chain. And it's the scarce one. Nvidia put $2 billion each into Lumentum and Coherent in March to lock up laser capacity, and Lumentum now says it's sold out through 2029.

POET's Blazar is an external light source. It's a hybrid laser built for co-packaged optics and chip-to-chip links, with high output power and multiple wavelengths, demoed at OFC in March and nominated for an ECOC award in September. The August release added an agreement with a Tier 1 laser company to develop an external light source engine, plus an exclusive supplier deal for a component that lifts Blazar's output power. The AGM put at-scale Blazar deployment in 2028. The Company called its partner a Tier 1 laser company and left it there. So will we.

The other half of Jensen's problem is a counting problem. Fifteen times the engines doesn't happen on lines that align every part by hand. POET's whole pitch, the Optical Interposer, is wafer-level assembly with no active alignment. That's why the Lumilens agreement was written around optical supply that scales with GPU fleets instead of against the limits of conventional assembly. Nobody at POET has claimed a CPO customer, and we're not claiming one for them. What's on the record is a product aimed at the exact slot Jensen just drew, with a 2028 date on it. That's why item six on the list below is the one we'd watch closest.

How the market is pricing it: revenue gets paid, roadmaps don't

Run the arithmetic, not the story. At the May high, POET's market cap was about $3.2 billion against roughly $430 million of cash at March 31. The market was paying close to $2.8 billion for the roadmap, the Lumilens framework and the engagements. Today the cap is about $985 million against $796.3 million of cash. That's under $200 million for the business. The cash nearly doubled. The price on the business got cut 93%.

Now look at who kept their multiples. Lumentum trades near 26 times trailing sales because its CEO can say sold out through 2029. Applied Optoelectronics is around 10 times sales with a $290 million quarter guided. Coherent sits near 6.5 times on $7 billion of revenue. All three are well off their highs, and all three still get paid for revenue the market can see. The story names got the other treatment: Lightwave Logic and Aeva down two-thirds, POET priced at cash. That's the flip. In May the tape paid for what POET was going to ship. Now it pays only for what's already shipped. Our sentiment screens have called it a "show me" tape for weeks, and this is what show me looks like in a market cap.

We don't publish price targets, so build your own number from that arithmetic. What we'll say is this: when the market pays only for revenue, the quarter that shows product revenue scaling is the one it re-prices, and the document that shows it has no call to lean on. That's why the first ask on the list below is the headline number.

Retail's top seven asks, by the sentiment data

We don't have to guess what the market wants from this release. Our sentiment screens pull StockTwits, X, the boards, news and call transcripts, and the asks rank themselves. Here are retail's top seven, loudest first, and what the document has to contain to answer each one.

1. Is revenue still growing? The single loudest question on the screens. Q2 was $569,925, up 13% from Q1 and 112% from a year ago, the sixth straight sequential gain. Retail wants a seventh, in the headline, the way the August release did it. Any sequential gain keeps the streak alive. The sharper accounts are asking for the mix line: product shipments versus engineering services.

2. Are the 800G engines shipping? The Q2 release said POET expects to ship substantial numbers of production units for qualification in the remaining quarters of 2026. The boards took that as a Q3 promise. A count, or a dated line that says the second-half ramp is on schedule, answers it. One sentence does it.

3. Are engagements turning into orders? At the June AGM the Company said it had 10-plus active customer engagements expected to exceed $100 million in future annual revenue. Nobody on StockTwits is asking for all of them. One moving from engagement to purchase order, named or not, would be the biggest line in the release.

4. What's the burn? Cash-burn questions dominate the news and transcript side of the screens, which is the side that's been bearish. $796.3 million at June 30 is $4.60 a share on 173 million shares. At $5.69 the market pays about a buck a share for the patents, the Lumilens framework and 150-plus people. The September deck penciled in roughly $15 million a quarter of burn and about $60 million of 2026 capex. The answer is the September 30 cash number in the subhead, the capex spent, and a reminder that this is equity money with no bonds or converts ahead of the common.

5. Where's Lumilens? The May 14 filing set engineering samples for late 2026, with a production ramp aligned to hyperscaler deployments in 2027. POET's engines aren't in anything Lumilens ships today; that's the timeline, not a problem. Retail wants one line that says late 2026 still holds, or that it moved and why. And the $2.4 million order disclosed in August: shipped, shipping, or scheduled?

6. Give us something besides "working hard." That phrase, and "making progress," is what the screens flag as the thing retail won't accept again. The specific ask is a dated milestone, and after Berlin the one that lands is the light source: the Tier 1 laser agreement for an external light source engine, or the higher-power Blazar. A sample date or a qualification step puts POET's light source on a calendar that overlaps Nvidia's capacity ramp. That's the slice of the AI-optics tailwind the X crowd wants POET attached to.

7. When do we hear from you next? The no-call complaint is loud, and it's fair. The release that announced the date also announced the calls. An outlook paragraph that dates the next milestones and restates the March 1 call turns November into chapter one instead of a cliffhanger.

What the tape is saying

Sentiment is split down the middle. Retail and the StockTwits crowd lean cautiously bullish on a technical bottom near cash value. News flow and transcripts are stuck on the 70%-plus drawdown and the burn question. On our sentiment screens, "stock near cash value, awaiting the Q3 catalyst" is the single most-mentioned theme right now.

The options board is louder. Calls outnumber puts by a wide margin on the November contracts. Open interest is stacked at $5.50 to $7 and again at $10 to $15, and implied vol sits around 85. That's upside positioning into the print. It's also fuel for the other direction: a document that says nothing turns those calls into supply by the November 20 expiration, four sessions after the release. You can likely guess which outcome we're rooting for.

One more thing on the chart. 5.74 is the level we've been tracking all month, and Tuesday's close landed a nickel under it. The chart isn't going to carry this print. The document has to. And the first tick won't be the verdict either way; in August the pre-market slot opened green and it still took the full session to mean anything.

November is the document. March is the conversation.

The Company has been heads down ramping since the AGM. That's where it should be. It's also why the stock drifted: no call, no narrative, no air cover. Our read is that POET has a lot to say and the first call in March is where it gets said. November 16 can't be that. It can be a well-built document, the way the execution story was the story in August: the number in the headline, the cash in the subhead, dated milestones in the middle, and an outlook that points at March 1. With the stock sitting on its cash, a document that does that work again is worth more than it was in August. We'll have the read before the open.

Subscribe to StockAlpha.ai for the November 16 read before the market opens.