POET Technologies
This is the archived full research. Read the story
A small company making the light-carrying parts inside AI data centre cables, sitting on a pile of cash and still waiting for its first real production customer.
The scores
Growthno evidence →
Revenue is $1M a year of fees and samples; doubling from that base shows no engine yet, and the one production order was cancelled.
Business quality2/4 →
Patents protect a method nobody has proven at volume, while the largest makers build their own engines and grow fast.
Financial position4/4 ↑
$796M of cash and deposits against $5.8M of debt and a burn near $50M a year: no outside money is needed.
Execution record2/4 →
Lost its production customer over a confidentiality breach, dropped the US move, and its 2026 shipment goal has nothing behind it yet.
Per share capture2/4 ↓
Shares up 126% in eighteen months, 78.7M warrants, heavy advisory fees, an unnamed loan and a likely PFIC tax on US holders.
Potential2/4 →
A $26B market, but the only order is conditional on qualifying a module not yet built, for a private buyer.
Risk profile2/4 →
The case turns on one start-up's qualification, with a class action open and orders cancellable without penalty.
Valuation3/4 ↑
The price is about the cash; the business is valued near a year's burn, but the cash is earmarked for spending, not holders.
Evidence quality2/4 →
The cash is filed, but customers are unnamed, orders conditional, there is no earnings call, and the plan rests on management's word.
If it works 2.5× · if it fails you keep about 74%.
The Verdict
We rate POET Technologies Neutral, with low conviction. Its way of building optical engines could be cheaper, but it sells almost nothing yet, its one production customer walked away in April, and its next big order, from Lumilens, waits on samples that have not shipped, so the business is mixed. The shares are worth about the cash in the bank, which keeps it alive for years, so the market pays close to nothing for the business: fair. The bet is speculative and fairly priced: it pays well if Lumilens qualifies the engines and orders follow, and holders keep most of the cash if not.
The long caseIf AI needs millions more optical links a month, making the core of each one on a wafer without alignment could win orders from the hand-built way.
How we weighed it
The price against our value
Cuts both ways
No fair value is drawn: the forecasts show no lasting profit after stock pay, so the bet prices it.
The business
Against it
Mixed
A mixed business near the weak end: almost no sales and its production customer left.
Growth against what the price assumes
Cuts both ways
Growth cannot be measured until a customer qualifies the samples.
What could go wrong
Against it
Orders can be cancelled, a US tax trap may apply, a class action is open and more shares may come.
Against competitors
Cuts both ways
No filed rival compares directly, so the price is judged on its own.
How its group is behaving
Cuts both ways
Optical names move together on AI spending news.
The latest news and report
Cuts both ways
The latest report cleared revenue and cash and kept Lumilens on schedule, but the engine ramp slipped half a year on shortages and yields.
The evidence under this - 5 claims
The share count is June's, since the Q3 statements were not read; warrant exercises would raise it. The $45M line and the burn will lower cash before engines sell (S21).
- 2026-11-17editedre-struck at the 16 November close against the 30 September balance sheet
- Fulfilment depends on qualifying a module not yet built, so the order can lapse without a breach (S06).
- Purchase orders can be cancelled or delayed without significant penalty, as POET's own risk factors say (S04).
- 2026-11-17editedrolled to the 30 September balance sheet from the Q3 release: cash and investments fell $22.0M in the quarter
The company
POET designs optical engines and light sources. Its trick is the Optical Interposer: a wafer on which lasers, light detectors and electronics are placed and wired by machine, without the slow hand alignment most makers use.
Engineers working on an industry standard estimated this kind of passive assembly could cut a transceiver's cost by about 60%. POET holds 81 patents, 49 of them on the interposer, and says its 800G receive engine has been designed into several customers' modules.
The evidence under this - 4 claims
The estimate is for 100G modules in 2013 and from a supplier's own engineers; no recent outside measure of the saving at 800G was found.
No customer is named, and revenue from these engines stayed below $1.1M a year (S01).
How it makes money
Today it barely does. Revenue was about $1.07M in 2025 and in the first half of 2026, all engineering fees and sample sales. The plan is to sell engines to module makers by purchase order, and those orders can be cut or delayed without much penalty.
The evidence under this - 3 claims
Where the money comes from, stream by stream
| Stream | Share of sales (FY2025) | How it earns it |
|---|---|---|
| Engineering fees and samples | 100% | Fees for custom engineering work and sales of sample engines to customers testing them, booked in Asia. |
Who runs it
Suresh Venkatesan has run POET since 2016 and chairs the board. Sandeep Kumar became operating chief in May, and Giovanni Pacelli, from Silicon Labs, took over as finance chief on 2 November.
Directors and officers together own under 0.1% of the company. Key managers were paid $7.3M in 2025, about half in shares, and the annual report says no set rule fixes bonuses.
The evidence under this - 4 claims
Why it wins, and who could beat it
Why customers come back
If it works, the edge is cost and power: an engine assembled by machine on a wafer should be cheaper to build at volume than one aligned by hand, and the patents cover the method.
That edge is not yet proven by repeat buyers. The customer that had committed to production, Celestial AI, was bought by Marvell, which cancelled every order in April citing a breach of confidentiality. Lumilens is now the case's main test.
The evidence under this - 5 claims
The estimate is for 100G modules in 2013 and from a supplier's own engineers; no recent outside measure of the saving at 800G was found.
No customer is named, and revenue from these engines stayed below $1.1M a year (S01).
- Fulfilment depends on qualifying a module not yet built, so the order can lapse without a breach (S06).
- Purchase orders can be cancelled or delayed without significant penalty, as POET's own risk factors say (S04).
The threat
The biggest transceiver makers, InnoLight above all, are growing fast with their own designs, and Lumentum and Coherent sell lasers, packaging and silicon photonics into the same links. The market also runs in booms followed by flat years, so a newcomer may arrive at a slowdown.
The evidence under this - 4 claims
The rivals: COHR, AVGO, LITE, and 1 more that is private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| InnoLight and Eoptolink, the largest transceiver makersPriced against it: Listed abroad | Optical engines and light sources on the Optical Interposer | On sale | Listed abroad | |
| COHR · AVGO | Coherent and Broadcom silicon photonics and co-packaged opticsPriced against it: Not checked | Optical engines and light sources on the Optical Interposer | On sale | Not checked |
| LITE · COHR | Lumentum and Coherent high-power laser sourcesPriced against it: Not checked | Optical engines and light sources on the Optical Interposer | On sale | Not checked |
How big it could get
How big this could get
The market is big. LightCounting puts sales of AI cluster transceivers and co-packaged optics at $16.5B in 2025 and $26B in 2026, with up to six transceivers per GPU.
POET's revenue was about 0.007% of that. How much it could win depends on qualifications it has not passed yet, so nobody can size its share honestly today.
The evidence under this - 4 claims
LightCounting's later forecast, reported by LightWave, put 2026 growth at 120% rather than 60% (S15).
POET sells engines, not whole transceivers, so its share of the engine market alone would be larger than this ratio, though no one publishes that market.
Who has signed
Lumilens placed a $50M first order, subject to development and qualification, and POET says it could grow to $500M over five years. Management also cites a $5M order from a systems integrator and a $2.4M order since June. None of these is binding beyond firm orders.
The evidence under this - 4 claims
- Fulfilment depends on qualifying a module not yet built, so the order can lapse without a breach (S06).
- Purchase orders can be cancelled or delayed without significant penalty, as POET's own risk factors say (S04).
- Nothing above the $50M order is committed, and Lumilens is a private start-up whose ability to pay is not disclosed (S06).
- The warrant pays Lumilens up to 22.9M shares at $8.25 for buying, a cost to holders if the volume arrives (S03).
Half-year revenue of $1.07M implies few production engines shipped by June (S03).
The results
What growth costs
So far growth costs far more than it brings in. Operating expenses doubled to $43.2M over three years while revenue stayed near $1M, staff rose past 115 with about 50 more planned, and management plans about $50M of equipment this half.
The evidence under this - 4 claims
- Revenue in the half was $1.07M, and POET's 2026 goal of 30,000 engines has no shipment figure behind it yet (S03, S02).
- The Q3 release moved the 30,000 engines targeted for 2026 to the second quarter of 2027, citing wafer and component shortages and a yield ramp (C42).
- 2026-11-17editedtested by Q3: the volume ramp slipped to the first half of 2027
The last report (16 Nov)
Quarter ended June 2026, filed figures.
In the third quarter of 2026 POET booked $1.03M of revenue, up 244% on a year earlier and its first quarter above $1M. It lost $7.6M, helped by a $6.4M paper gain on warrants, and used $12.7M of cash in operations.
Cash and investments were $774.3M. The 30,000 engines it had aimed to ship in 2026 now ship by mid-2027, blamed on wafer and part shortages and yields, while a new production line worth over $45M is built. Lumilens's first products are on schedule.
What it did to the thesisIt cleared the bars we set, but moved nothing that changes the view: sales are still a fifth of the level that would, Lumilens has not qualified, and the volume ramp slipped half a year.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $-0.04 against $-0.06 expected.
The evidence under this - 3 claims
In June management said volume shipments would begin in the second half of 2026 (C12); the release blames wafer and component shortages and a yield ramp, so the 2026 goal has already slipped once.
- 2026-11-17editedrolled to the 30 September balance sheet from the Q3 release: cash and investments fell $22.0M in the quarter
What management did with the money
Most of the money raised is still in the bank. Beyond that, POET paid a financial adviser $9.5M in the first half, lent $30M to an unnamed company and bought a small stake in a module partner. None of these has been explained in detail.
The evidence under this - 3 claims
The price, and where we disagree
In short. Fair: the shares trade at about the cash in the bank, so the business is priced near nothing, which is roughly right until orders turn into sales.
The shares fell 77% from their high after Marvell cancelled the orders in April, a short seller's report and a class action followed, and they fell 9% more the day the third quarter was reported, to $4.66, well below the $21 an investor paid in May.
The next tests are close: Lumilens samples by the end of 2026, fourth-quarter shipments promised to beat the first nine months', and the first earnings call with the year's results on 1 March 2027.
The evidence under this - 5 claims
- Fulfilment depends on qualifying a module not yet built, so the order can lapse without a breach (S06).
- Purchase orders can be cancelled or delayed without significant penalty, as POET's own risk factors say (S04).
How we value it
No fair value: the analysts' forecasts show no lasting operating profit after stock pay, so a value built on earnings would rest on a guess; the bet prices it instead.
How we read the multiple
At $4.66 the 173M shares are worth about $806M, and the company holds about $766M of cash and investments net of debt. The business itself is priced at about $40M, close to nine months of spending.
That is a fair price for a company with no product revenue yet: the downside is mostly covered by cash, but some of that cash will be spent before engines sell.
Against its own history. It trades far below what it used to fetch against its net worth, because the cash raised this year is now almost all of what the company owns.
The evidence under this - 3 claims
The share count is June's, since the Q3 statements were not read; warrant exercises would raise it. The $45M line and the burn will lower cash before engines sell (S21).
- 2026-11-17editedre-struck at the 16 November close against the 30 September balance sheet
- 2026-11-17editedrolled to the 30 September balance sheet from the Q3 release: cash and investments fell $22.0M in the quarter
Priced against its rivals
None of the listed rivals above sells the same thing to the same buyers as a large part of what it does. A price comparison needs at least two, so the price is judged against its own history instead.
The bet: what it pays if it works
This is a bet on POET becoming a real supplier. If Lumilens qualifies the engines and its first order ships, sales of a modest size would value the business at more than twice today's price.
If the engines never sell, holders keep most of their money, because the cash left after a few years of spending is still most of today's market value.
| Revenue in two to three years | Dollars of yearly product revenue needed | At the stated multiple |
|---|---|---|
| $10.00M | about 10.0M at 100% kept | 1.0× |
| $25.00M (the big win) | about 25.0M at 100% kept | 2.5× |
| $100.00M | about 100.0M at 100% kept | 9.9× |
To pay 2× (a Fair bet) it would need about $20.23M of revenue (about 20.2M dollars of yearly product revenue); to pay 5× (a Cheap bet), about $50.56M of revenue (about 50.6M dollars of yearly product revenue), at the same multiple.
The steps it has to take
- Done Lumilens places a $50M first order
- Done $400M raised, funding the capacity build
- In progress Lumilens engineering samples delivered due 2026-12-31
- Slipped Volume shipments of production engines begin due 2027-06-30
- Not started Lumilens production ramp due 2027-12-31
- Not started Quarterly product revenue of $5M due 2027-12-31
The detail behind the label
The evidence under this - 3 claims
- Fulfilment depends on qualifying a module not yet built, so the order can lapse without a breach (S06).
- Purchase orders can be cancelled or delayed without significant penalty, as POET's own risk factors say (S04).
- 2026-11-17editedrolled to the 30 September balance sheet from the Q3 release: cash and investments fell $22.0M in the quarter
Where we disagree with the market
The cash will leak away on advisers, loans and spending before the engines ever sell in volume.
The cash covers the downside for years; the open question is only whether Lumilens qualifies and keeps buying.
Short sellers hold about a seventh of the shares, and the price says the market expects much of the cash to be spent without a business to show for it. The two analysts covering it rate it a buy.
We sit between them. The cash puts a floor under the value for some years, but the engines have not yet won a production customer that stayed.
What analysts say, and the evidence
The evidence under this - 4 claims
The share count is June's, since the Q3 statements were not read; warrant exercises would raise it. The $45M line and the burn will lower cash before engines sell (S21).
- 2026-11-17editedre-struck at the 16 November close against the 30 September balance sheet
- Fulfilment depends on qualifying a module not yet built, so the order can lapse without a breach (S06).
- Purchase orders can be cancelled or delayed without significant penalty, as POET's own risk factors say (S04).
What could go wrong, and how we would know
The risks
- Orders can vanish Customers buy by purchase order and can cancel or delay without much penalty, as the April cancellation showed.
- A US tax trap Interest income is most of its gross income, so POET may be a passive foreign investment company, which taxes US holders harshly.
- A lawsuit A class action alleges misstatements about that tax risk and its customer relationships in April 2026.
- More shares The share count more than doubled in eighteen months, and 78.7M warrants could add more.
- Unexplained spending Large adviser fees and a $30M loan to an unnamed company draw on cash that belongs to holders.
The evidence under this - 5 claims
The test is applied by tax counsel on the whole year and on asset values, and cost of sales may reduce gross income from revenue further; the company has not said.
What we still do not know
- Lumilens publishes no accounts, so whether it can pay for $500M of engines, and who its customers are, cannot be checked.
- Who borrowed the $30M, and why, has not been disclosed.
- Whether the plan for up to 1 million units a month by end-2027 has buyers behind it.
The evidence under this - 3 claims
Lumilens committed to a $50M first order and took a warrant tied to paying more, which a buyer without means would not do (S06).
- Revenue in the half was $1.07M, and POET's 2026 goal of 30,000 engines has no shipment figure behind it yet (S03, S02).
- The Q3 release moved the 30,000 engines targeted for 2026 to the second quarter of 2027, citing wafer and component shortages and a yield ramp (C42).
- 2026-11-17editedtested by Q3: the volume ramp slipped to the first half of 2027
What would change our view
- Neutral→AttractiveProduction revenue arrives
It needs all of:
- ○Quarterly revenue reaches $5M from product sales, not engineering fees.checked at the next results, due by 1 Mar
- ○Lumilens's engines pass qualification and production orders follow.watched as it happens
- Neutral→UnattractiveThe cash runs out of the business
It needs:
- ○Cash and deposits fall below $500M with quarterly revenue still under $2M.checked at the next results, due by 1 Mar
What is coming
- 16 November 2026 Q3 2026 results: Whether production engines started shipping and how fast the cash is going.
Every dated event ahead
- 16 November 2026 Q3 2026 results: Whether production engines started shipping and how fast the cash is going.
- 31 December 2026 Lumilens engineering samples: Whether the joint engine works well enough to move toward production.
- 1 March 2027 Q4 2026 results and the first earnings call: The year's shipments against the 30,000-engine goal, and analysts' first questions on the record.
Since we wrote this
- 17 NovEarningsRevenue passed $1M; the engine ramp slipped to mid-2027Q3 revenue was $1.03M, cash held at $774.3M and Lumilens stayed on schedule, but the 30,000 engines meant for 2026 now ship by mid-2027. Neutral, low conviction, holds.
- 17 NovBet moveBet re-struck: Priced for success to AsymmetricRe-priced at the 16 November close; the rating moved from Expensive bet to Cheap bet.
- 17 NovBet moveBet re-struck: Asymmetric to SpeculationRe-priced at the 16 November close; the rating moved from Cheap bet to Fair bet.
- 9 NovBet moveBet re-struck: Not established to Expensive betRe-priced at the 6 November close.
- 6 NovBet moveBet re-struck: Expensive bet to Not establishedRe-priced at the 5 November close.
- 4 NovBet moveBet re-struck: Speculation to Priced for successRe-priced at the 3 November close; the rating moved from Fair bet to Expensive bet.
- 3 NovThesis writtenNeutral, low convictionWritten after the shares fell to $4.80, close to the cash per share, two weeks before the Q3 results.
Also checked 1 time with nothing new to act on.
Sources 22
Sources - 22
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-03-31business description, products, patents, manufacturing partners, risk factors, PFIC disclosures, five years of statements, ownership table, management pay, headcountbacksC01C02C05C24C25C28C29C35C38C39
- S02primarycompany release2026-04-01the 2026 shipment goal, the $5 million production order, the 2025 financings
In total, we expect to ship more than 30,000 optical engines this year across both high-speed and high-power segments, underscoring the scale of our ramp.
backsC13 - S03primarycompany filing2026-08-13the June 2026 balance sheet, cash flow, share capital, warrants, the loan receivable, the Lessengers investment, the Lumilens warrant, the legal action notebacksC03C04C05C06C07C09C27C29C30C33
- S04primarycompany filing2026-08-13products, qualification history, 2026 milestones and budgets, finance advisory fees, liquidity, purchase-order risk, partnershipsbacksC01C06C18C19C23C26C27C31C37C40
- S05primarycompany release2026-08-13Q2 revenue and loss, the five-quarter table, the post-quarter orders, the outlook, the decision not to redomicile
The Company expects to begin shipping substantial numbers of production units for qualification in the remaining quarters of 2026
backsC03C05C08C14C19C25C26C28 - S06primarycompany release2026-05-14the $50 million initial order, the $500 million framework, the warrant, sample and ramp timing
Fulfilment of the purchase orders and associated revenues is subject to the successful development and ultimate qualification of the modules, as well as the successful scaling of manufacturing capability.
backsC08C09C11 - S07primarycompany release2026-05-18the $400 million raise from one investor, use of proceeds, headcount of 115
We are in the process of expanding our capacity by roughly ten-fold in both wafer production and optical engine assembly in order to support a transition to higher volume manufacturing into 2027
backsC07C32C35 - S08primarycompany release2026-04-27the cancellation of every Celestial AI purchase order
Marvell indicated that the Company had made disclosures of information related to the Purchase Order and shipping information in contravention of its confidentiality obligations.
backsC10 - S09primarycompany release2026-06-30the production ramp, the capacity goal, the capital plan, Blazar timing, the engagement count
The Company has more than 10 active customer engagements that combined are expected to exceed $100 million in future annual revenue.
backsC12C35C40 - S10primarycompany release2026-11-02the new finance chief, his pay in shares, the first earnings calls
Beginning with its fourth-quarter and full-year 2026 results, the Company intends to host an earnings conference call each quarter
backsC34C37 - S11primarycompany release2026-10-19the Q3 date and the first call date of 1 March 2027
it will report its financial results for the third quarter ended September 30, 2026, on Monday, November 16, 2026
backsC34C36 - S12primarycompany release2026-08-03two new directors and a resignation
Most recently, he served as Founder, Chief Executive Officer and Chief Technology Officer of Avicena Tech
backsC37 - S13primaryindustry data2026-01-31the size of the AI cluster transceiver and CPO market
We estimate that the market reached $16.5 billion in 2025 and it will reach $26 billion in 2026.
backsC15C39 - S14primaryindustry data2026-03-31growth in 2026, the cycle, transceivers per GPU, the $100 billion case
this should limit the growth in sales of optical transceivers to “only” 60% in 2026, unless customers create deep inventory reserves for the future
backsC16 - S15secondarynews2026-09-30LightCounting's later forecast and InnoLight's growth
Innolight reported close to 200% y-o-y growth in sales of optical transceivers for the first half of 2026
backsC20 - S16primaryindustry data2013-01-21Anderson and Hiramoto (Oclaro), PSM4 Technology and Relative Cost Analysis, IEEE P802.3bm task forcewhat passive alignment and surface mounting save on a transceiver, as a standards submission
Our analysis indicates ~60% cost reduction in optical transceiver is achievable with the proposed PSM4 approach
backsC17 - S17primarycompany filing2026-08-17Lumentum Holdings Inc. Form 10-K for the year ended June 27, 2026 Lumentum Holdings Inc.a listed rival selling high-power laser sources
we deliver customized solutions tailored to specific customer needs, such as high-power laser sources, compact laser arrays, and advanced packaging schemes
backsC21 - S18primarycompany filing2026-08-14Coherent Corp. Form 10-K for the year ended June 30, 2026 Coherent Corp.a listed rival selling transceivers, silicon photonics and co-packaged optics
Our portfolio also includes silicon photonics, CPO, OCS, and other advanced optical technologies that support evolving AI datacenter architectures.
backsC22 - S19secondaryother2026-06-01the case caption, class period and allegations, and the Wolfpack report date
Captioned Christopher Jones v. POET Technologies Inc., No. 26-cv-04717 (D.N.J.), the POET class action lawsuit alleges that POET and/or certain of its officers and/or directors violated federal securities laws
backsC30 - S20primaryvendor data2026-11-03the close, the 52-week range, returns, the analyst split, short interestbacksC31C32
- S21primarycompany release2026-11-16Q3 revenue, net loss, the warrant gain, cash used in operations, cash and short-term investments, the capacity plan, the 30,000-engine schedule, customer counts and the Lumilens status
We expect to ship our current backlog of optical engines, including the 30,000 engines we had targeted for 2026, by the second quarter of 2027
backsC04C33C41C42 - S22secondaryvendor data2026-11-17cash of $276.0M, short-term investments of $498.4M, total debt of $8.3M and equity of $848.3M at 30 September 2026; the 16 November close of $4.66backsC04C33