Skip to content
primarysourced Photonics sector POET
POET

Disagreement log

Where primary sources, aggregators, or prior framings disagree. Each entry shows the competing claims, the resolution, and what it implies for the thesis. Keeps the KB honest about what we actually know vs. what was assumed.

  1. Identity of the 'leading systems integrator' behind the disclosed $5M production order

    POET IR (2025-10-22 press release)
    $5M production order from a 'leading systems integrator' for 2xFR4 + 2xDR4 400G transmit + 800G 2xFR4 + DR8 receive engines
    as of 2025-10-22
    Customer name not disclosed
    Open-questions inference (POET KB)
    Profile fits a Tier-1 Chinese module OEM (Innolight or Eoptolink) or a US module OEM (Coherent, Lumentum, Source Photonics); 'systems integrator' framing is unusual and could indicate a hyperscaler-direct or networking-equipment-vendor (Cisco, Arista) engagement
    as of 2026-04-29
    Highly speculative inference
    Resolution
    Cannot resolve from primary sources; classify as undisclosed customer until subsequent press release or 20-F revenue-concentration footnote names them Counter view: Some bull-case framings assume the customer is a hyperscaler-tier name and use that to justify higher-multiple valuation
    Implication
    bull_case Pillar 4; bear_case Pillar 3; risks C1, C2

  2. POET's actual fully-diluted share count post-January 2026 raise

    POET STOCK_PRICE_DATA.json (POET data folder)
    ~67.5M shares outstanding
    as of 2026-04-28
    Conflicts with primary 20-F + 424B5 filings
    424B5 Jan 2026 (accession 0001493152-26-003330)
    Pre-offering 132,021,526 / Post-offering 152,711,182 common shares O/S
    as of 2026-01-23
    Primary-source authoritative
    Bear-case dilution math (POET KB bear_case.md)
    Fully diluted ~195.86M shares including 5.79M options + 37.36M warrants (Q4 2025); ~245M+ if forward 2026-2027 raise lands
    as of 2026-04-29
    Late-2025 secondary sources; warrant/option count to be re-verified directly from FY25 20-F
    Resolution
    152.7M basic post-Jan 2026 raise is the load-bearing figure (per 424B5). Fully-diluted ~195.86M including warrants and options is the bear-case overhang. Counter view: POET's STOCK_PRICE_DATA.json shows 67.5M; relying on stale data understates dilution by ~50%
    Implication
    bear_case Pillar 1 dilution math; risks K1, K2

  3. Whether the Marvell-Celestial AI PO cancellation reflects substantive technical/commercial disagreement or strictly a confidentiality breach

    Marvell Semiconductor cancellation notice (cited in POET 2026-04-27 press release)
    Cancellation cited 'alleged disclosures of information related to the Purchase Order and shipping information in contravention of confidentiality obligations'
    as of 2026-04-23
    Marvell-stated official reason
    Inferred bear-case reading
    Marvell's Celestial AI integration may have led the team to consolidate around captive Photonic Fabric chiplets rather than third-party engines like POET's
    as of 2026-04-29
    POET KB customers.md flag — not primary-sourced
    Pre-cancellation Stocktwits CFO interview
    Mika referenced '$5M production order for 800G optical engines and outstanding invoices from Celestial AI' and shipment timing 'as early as next quarter'
    as of 2026-04 (pre-cancellation)
    Likely the trigger for the confidentiality cite
    Resolution
    Both true at different scopes — confidentiality cite is the formal cancellation reason; whether Marvell's strategic preference for captive Photonic Fabric chiplets is the deeper rationale is unverifiable from primary sources Counter view: Bull-case framings argue this is purely a technicality and POET's IP/customer pipeline is unimpaired
    Implication
    bull_case Pillar 4; bear_case Pillars 2-3; risks C1, X5

  4. Whether the 2024-2025 Sanan IC SPX buyout terms ($6.5M minority + $6.5M majority = $13M total) signal a distressed unwind or favorable structural exit

    Bull case Pillar 2 reading (POET KB)
    POET removed the China-JV overhang at favorable terms; $13M total payable-over-5-years for 100% of SPX is structurally underpriced for production capacity
    as of 2026-04-29
    Bull framing
    Bear case Pillar 4 reading (POET KB)
    $6.5M for 24.8% implies Sanan's accounting valued the SPX joint venture at ~$26M of total enterprise value at exit — a fraction of the cumulative capital and management attention POET deployed 2020-2024
    as of 2026-04-29
    Bear framing
    Resolution
    Both readings have merit; the load-bearing fact is that the JV underperformed revenue expectations relative to original 2020 framing. The $13M aggregate price is small relative to capital deployed, consistent with a distressed-unwind interpretation Counter view: Bull reading is correct that POET retains operational flexibility and full SPX control without paying a premium
    Implication
    bull_case Pillar 2; bear_case Pillar 4; risks F3, C4

  5. Whether the 30,000+ engine 2026 shipment commitment is achievable

    POET Q4 2025 release (2026-03-31)
    'We have decisive transition from development to execution'; guidance of 30,000+ optical engines shipped in 2026
    as of 2026-03-31
    Management commitment
    Bear case Pillar 2 reading (POET KB)
    Pattern of 4 prior cycles where management 2-yr commitments slipped to 4-6 yr horizons (2018 SilTerra, 2020-21 SPX, 2022-23 100G/400G, 2024-25 partner-stack); reasonable Bayesian prior is 30,000 commitment is partially delivered with volume/ASP shading downward
    as of 2026-04-29
    Pattern-based skepticism
    Resolution
    30,000 is a stretch commitment; conditional on (a) production yield at SilTerra scale (currently unproven at volume) (b) replacement orders post Marvell-Celestial cancellation (c) on-time Sivers/QCi partner deliveries Counter view: Bull case Pillar 5: even at the most optimistic ASP assumptions ($300-500/engine), 30,000 engines = $9-15M product revenue — a plausible directional inflection if delivered
    Implication
    bull_case Pillar 5; bear_case Pillars 1-2; risks T1, F1, C1

  6. Exact date of TSXV delisting

    FY25 20-F Item 3.D risk factor
    '2024 (calendar year, exact date undisclosed in 20-F)'
    as of 2026-03-31
    Date not disclosed in 20-F narrative
    Project orchestrator brief
    2025-08-27 cited as TSXV delist date
    as of 2026-04-29
    Date appears in orchestrator instructions; not yet primary-sourced
    Resolution
    TSXV delisting occurred during 2024-2025 transition period; precise date requires SEDAR / TSXV bulletin lookup. The orchestrator-cited 2025-08-27 is recorded with low confidence pending primary-source confirmation Counter view: 20-F flags it as a 2024 event — could be timing inconsistency
    Implication
    FPI risk-factor disclosure (Item 3.D); risks IP5, K4

  7. Whether monolithic SiPh + CPO will compress merchant-pluggable-engine TAM (POET's market) by 2027-2028

    Tower Semiconductor announcement (2025-11-12)
    $300M SiPh capex expansion; new CPO foundry technology available on Tower's leading SiPho and EIC optical platforms
    as of 2025-11-12
    Industry capex signal
    Marvell + Celestial AI + Polariton stack (closed Feb 2/Apr 22 2026)
    End-to-end optical stack creates structural CPO scale-up bet
    as of 2026-04-22
    Major end-to-end competitor
    POET bull-case Pillar 1 counter
    Hybrid-integration architecture is cost-superior to monolithic SiPh for transceiver-component economics; addresses module-OEM customer set, not hyperscaler-direct CPO
    as of 2026-04-29
    POET architectural counter
    Resolution
    CPO architectural shift is a credible 3-5 year risk to merchant-pluggable layer; POET's response (Sivers ELS for CPO; QCi 3.2T) is structurally consistent with adapting rather than competing head-on Counter view: Bull case argues hyperscalers run multi-architecture stacks indefinitely; merchant-pluggable layer holds ASP
    Implication
    bull_case Pillar 1; bear_case Pillar 5; risks T5, X1, X3, X5

  8. Whether/when POET completes US redomicile (resolves PFIC + opens US institutional investor base)

    POET 6-K PFIC disclosures
    POET has signaled US-redomicile evaluation
    as of various 2024-2025
    No public timeline disclosed
    FPI redetermination 2026-06-30 (FY25 20-F Item 3.D)
    20-F warns 2024 TSXV delisting + US-investor-heavy capital raises may push POET below FPI threshold on next determination
    as of 2026-03-31
    Forced-deconversion path
    Resolution
    Redomicile is a probability-weighted forward catalyst; most likely path is AGM 2026 vote with completion 2026-2027. If POET loses FPI status involuntarily, the conversion is forced rather than strategic Counter view: Voluntary redomicile would be capital-allocation-positive; involuntary FPI loss would be incrementally costly (10-K, 10-Q, 8-K, DEF 14A, Section 16, Reg FD compliance)
    Implication
    risks K4, IP5

New disagreements added as primary sources land. If you spot one missing, add it in src/data/disagreements.json.