Share this article
Spread the word on social media
POET Investor Presentation 2026: by the numbers
POET Technologies POET Technologies (POET) has a new investor presentation out for September 2026. Nineteen slides, no press release attached. That's fine by us. A deck is where management shows its work, and this one shows more of it than most small caps bother with. It has a cap table. A burn rate. Even a capex number. Those three things alone put it ahead of half the decks we read.
We went through all nineteen pages so you don't have to. Here are the five numbers that matter, and what we think each one tells you.
1. Cash: about $800 million, and the deck tells you how fast it goes
Slide 18 is the money slide, literally. Total cash sits at about $800 million as of June 30, 2026. Operating working capital is the same figure. That squares with the $796.3 million in cash and short-term investments from the Q2 report in August.
The new part is the burn. POET cash burn is projected at about $15 million a quarter. Capex for 2026 is about $60 million. Do the math. Sixty million a year of operating burn plus sixty million of capex is $120 million. Against $800 million, that's more than six years of runway at today's pace. The pace won't hold once production ramps, and the slide says "projected" for a reason. Still, the company put its burn rate in front of investors. Most pre-revenue names bury it.
2. Sales: $1.1 million in the first half, against $830 million raised
Here's the tension the whole deck lives inside. POET sold $1.1 million of product from January through June. It raised $830 million through private placements between May 2025 and May 2026, on 66 million shares. Nobody buys $POET for $1.1 million of first-half sales. They're buying what that balance sheet is supposed to turn into.
That's what the $50 million purchase order with $500 million of potential is for. Slide 12 lists it under "Modules and Light Sources for Hyperscalers." We know the buyer as Lumilens, the customer that came out of stealth in August. The Q2 release said substantial numbers of production units for qualification would ship in the remaining quarters of 2026. The deck doesn't update that timeline. It doesn't walk it back either.
3. Fully diluted: 264 million shares, with the warrants deep in the money
Slide 18 does something else we wish every small cap would do. It prints the whole capitalization. There are 173,035,169 common shares, 78,734,968 warrants at a $4.39 average strike, 5,369,393 options at $2.10, and 7,035,953 RSUs. That's 264,175,483 fully diluted shares.
Read the warrant line twice. At $7.45, every one of those warrants is more than $3 in the money. Exercise would bring in roughly $346 million at the listed average strike. That's the trade the company printed on the slide: more shares, more cash to build with. Traders who quote the $1.3 billion market cap should know the fully diluted figure is closer to $2 billion at Monday's price. The deck lets you do that math yourself, which we respect.
4. Headcount: 153 people, up 60 since January
We wrote in August that POET is hiring like a company with customers. The deck confirms the count. Total headcount is 153 across six countries, 143 full-time and 10 contractors, with 60 added since January 2026.
The map is the interesting part. Singapore holds 93 full-time employees plus five contractors doing interposer and process development. Penang has 20 full-timers in optical engine manufacturing. Kulim, at the SilTerra interposer foundry, has 10. The Bay Area runs management, sales and marketing with 12. Toronto, the corporate headquarters, has one full-time employee. That isn't a company built around a head office. It's a company built around a fab and an assembly line.
5. Capacity: 1 million units a year, going to 10 million
Slide 11 says assembly and test capacity is growing from 1 million units a year to 10 million. Cleanroom space goes from more than 8,000 square feet to 20,000. More than 25 semiconductor assembly and test tools are already installed: optical and electrical die bonders, a reflow oven, wire-bonders, burn-in and testers. Contract manufacturers are described only as "established partners in key jurisdiction." No names.
The capacity number matters because of how POET says it builds. The pitch is wafer-level assembly instead of one-at-a-time job-shop work. Optical engine chips built hundreds at a time, with no active alignment. The deck's own comparison claims 10% to 30% lower module cost, 20% to 30% lower power, and 10x lower assembly and test capex than conventional approaches. Those are internal estimates, and the slide says so. But a 10x capacity target is the kind of claim you can check against shipments.
What else is in the POET investor presentation
The market slide leans on LightCounting. Ethernet transceiver sales went from $2.7 billion in 2016 to $10.7 billion in 2024. LightCounting projects $31.4 billion by 2030, with cloud at 94% of it. Optics for scale-out plus scale-up run from $9.6 billion in 2024 to $29.5 billion in 2030. Call it a $30 billion market growing about 20% a year. The speed jumps are compressing too. Ten gig to 100G took about eight years. Then 400G to 800G took two, and 800G to 1.6T about one.
The customer slides name names. Mitsubishi, for 1.6T and 3.2T pluggable modules that put Mitsubishi's 200G and 400G EMLs on the POET Optical Interposer. NTT, for 1.6T and 3.2T engines and for 100G BiDi and ER. Adtran $ADTN, for data center edge optics. One line on the Mitsubishi slide deserves a highlighter. Mitsubishi's 400G lasers can't go into conventional wire-bond assembly because of crosstalk. They need an integration platform. That's the whole POET thesis in one sentence, told through the customer's problem instead of the vendor's pitch.
The IP count is 49 patents and 37 applications, all tied to the Optical Interposer. The deck says POET is one of very few companies that can reach 1.6T right now, with schematics up to 6.4T. The management slide is four people: CEO Suresh Venkatesan, CFO Thomas Mika, Chief Revenue Officer Raju Kankipati, and COO Sandeep Kumar.
What the deck doesn't do
It doesn't guide. There's no revenue target for 2026 or 2027. The hyperscaler behind the $50 million order stays unnamed. The ramp timeline doesn't move past what the Q2 release already said. And it prices the stock at $7.46, the August 31 close. $POET traded near $7.45 on Monday morning, down about 4% on the day. A month of tape brought the stock right back to the deck's own snapshot. What changed in between sits in the slides, not the price.
Our read: the POET investor presentation is a balance sheet with a roadmap stapled to it. The things that will move POET stock from here aren't on any slide yet: production shipments, a dated ramp, the Q3 report in November. What the deck gives you is the base case in the company's own arithmetic. That's worth more than a press release.
Subscribe to StockAlpha.ai for daily coverage of $POET and the AI photonics trade.