796M and No Debt: POET Technologies Puts Its Cash to Work

Share this article
Spread the word on social media
The war chest isn't sitting idle. It's buying capacity, lasers, and partners.
POET Technologies $POET closed June with $796.3 million in cash and short-term investments and no debt. That's a pile of money for a company that booked $569,925 of revenue in the quarter. It's also the point. The $400 million raise in May wasn't a rainy-day fund. It's being spent on capacity, R&D, lasers, and partners across the AI optical interconnect stack. The cash is the strategy.
Cash is a weapon in silicon photonics
Silicon photonics and co-packaged optics eat capital. Everyone in the space is racing to qualify optical engines at 800G, 1.6T, and 3.2T. Most are doing it while sweating the next raise. POET isn't. CEO Suresh Venkatesan said the May financing "removes the uncertainty associated with our staying power." In plain English, nobody at a qualification meeting has to ask if POET will still be around in 2028. That matters when partners and hyperscalers pick who they'll build around.
Run the burn. Operating cash outflow was $12.2 million in Q2. Call it $50 million a year, more once the ramp builds. Against $796 million, that's not a runway. It's a highway. The cash also earns its keep. Interest and other income came in at $4.3 million for the quarter. That covers a third of the burn before a single engine ships.
Where the $400 million goes
POET's filings earmark the proceeds for manufacturing capacity, R&D, the light source business, acquisitions and corporate development, and general expansion. The first four are the story.
1. Manufacturing capacity. The Lumilens order is the driver. POET's September investor deck pencils about $60 million of 2026 capex, with Penang assembly scaling from 1 million optical engines a year toward 10 million.
2. R&D. Spending hit $5.8 million in Q2, up from $4.5 million in Q1 and $3.1 million a year ago. The company calls it the move from tech development to product development. We'd call it building parts a customer can qualify. That money feeds the 800G Infinity engines now shipping for qualification. It also funds the Teralight 1.6T and 3.2T chipsets POET is building with Mitsubishi Electric's EML lasers, and the Lessengers 1.6T module.
3. Light sources. Blazar, POET's hybrid laser, just got an exclusive supplier for a component that significantly boosts its output power. After quarter-end, POET signed with a Tier 1 laser company to co-develop an external light source engine.
4. Corporate development. POET took a 4.99% stake in Lessengers, about $3 million, on top of the joint module program. That turns a contract into a structural tie. Expect more of it.
The Lumilens order anchors it all
In May, Lumilens placed a $50 million initial purchase order for optical engines inside a framework that could top $500 million over five years. Engineering samples are targeted for late 2026. And Lumilens isn't a lightweight. It raised more than $700 million in a Series C at a $5.51 billion valuation while still in stealth.
The deal structure tells you how both sides see it. POET granted Lumilens warrants on up to 22,921,408 shares at $8.25. About 2.3 million are exercisable now. The rest vest only as Lumilens pays toward future orders, up to that $500 million. POET booked the warrants at a $30.1 million fair value as deferred customer consideration. Read that twice. POET's spending $30 million of equity value to tie a customer's upside to volume. You don't make that trade from a weak balance sheet. You make it when you can afford to spend ahead of revenue.
That's the sequence the cash funds. Capex now. Qualification next. Production revenue when the samples pass.
Capacity before the ramp, not after
POET Technologies runs its own assembly and test in Singapore and is building out Penang. Headcount went from 115 in May to 155 by late August. Mitsubishi Electric supplies the lasers inside the Teralight engines. Lessengers is co-developing a 1.6T transceiver module with POET. A cash-strapped fabless startup doesn't get those relationships. Partners pick suppliers that can finish what they start. Management says the optical engine production ramp is still on schedule for the second half of 2026. Substantial numbers of production units ship for qualification through year-end.
Light sources are the sleeper
Co-packaged optics needs integrated, high-power, multi-wavelength light sources at scale. The industry is short on them. POET's treating Blazar as a product line, not a side project. The Tier 1 laser partner validates the external light source approach. At the June AGM, POET put Blazar on track for at-scale deployment in 2028. Cash is what lets a company carry a 2028 program while it funds a 2026 ramp.
What could go delay an AI infrastructure ramp
The risk here isn't demand. It's whether the industry can build fast enough to meet it. We laid this out in our supply chain analysis: 800G and 1.6T optical modules are running lead times past 40 weeks, InP laser supply is the chokepoint, and capacity across the chain is sold out. An AI cluster is a convoy. A slip anywhere in the chain shows up everywhere in the chain. POET's ramp rides that same convoy: lasers, substrates, assembly tools, test capacity. In this market, a missed ship date is usually the supply chain talking, not a company-specific red flag. Cash buys priority with suppliers. It doesn't buy time. Six straight quarters of sequential revenue growth say the engine's turning over. They don't say it's at speed.
Bottom line
No debt. A $50 million order with a $500 million path. Capex going into the ground. R&D rising by design. A light source business with a Tier 1 partner. Can POET Technologies afford the jump from development-stage company to production-scale optical engine supplier? Yes, with room to spare. Whether it makes the jump is an execution question. The next two quarters of qualification shipments start answering it.
Want the next $POET read before the tape prices it? Subscribe to StockAlpha.ai.