Confidence legend: ✓ verified-primary · ◐ partial / aggregator · ⚠ inferred / estimate.
This page is the canonical reference for any thesis claim about POET’s market size, growth rate, or share assumptions. Each segment carries explicit primary-source attribution; analyst syntheses are flagged ◐. POET’s own management has been disciplined about not publishing aggressive long-range revenue targets, so the SOM modeling on this page is built bottom-up from research-house TAM data + share-assumption sensitivities, not management framework.
1. Top-of-the-funnel: pluggable optical transceiver TAM
LightCounting baseline (the dominant industry source)
LightCounting tracks PAM4 + coherent DSP + transceiver markets quarterly. The most-recent published numbers (LightCounting newsletters, OFC 2025 presentations):
| Year | TAM ($B) | YoY % | Notes |
|---|---|---|---|
| 2024 actual | ~$11–12 | — | Pre-AI-buildout-acceleration ⚠ |
| 2025 actual | $16.5 | +40% | ✓ LightCounting February 2025 newsletter |
| 2026E | $26.0 | +60% | ✓ Same source |
| 2028E | ~$40 | +25% CAGR 2026-2028 | ◐ Mordor Intelligence trajectory extrapolation |
| 2030E | ~$60 | +30% CAGR 2025-2030 | ◐ Same Mordor source |
The LightCounting 2025 actual of $16.5B and the 2026 forecast of $26B are the two anchor data points anyone analyzing POET’s TAM should validate first. Neither is a forecast — the 2025 print is a measured market figure, and the 2026 figure is a near-term forecast that LightCounting updates quarterly.
Dell’Oro’s optical-transceiver forecast
- 2030 forecast: $4.4B for the IPoDWDM-systems sub-segment alone (datacenter-internal carrier-grade optics; not the broader pluggable market). ✓ Dell’Oro Group press release
- The broader datacenter-Ethernet-transceiver TAM that Dell’Oro tracks is significantly larger than the IPoDWDM slice and aligns directionally with the LightCounting trajectory above.
650 Group view
650 Group covers switch-ASIC + transceiver markets jointly. Their analyst commentary on the broader silicon-photonics-vs-discrete optics shift includes:
- ◐ 650 Group — Marvell Investor Day highlights surge in custom ASIC demand — referenced as a directional view of the related custom-silicon market that drives optical-port demand.
TAM segmentation by data rate
Because POET’s product roadmap targets specific data rates (currently 100G / 400G / 800G; with 2026 focus on 1.6T per the 20-F), the TAM segmentation matters:
| Generation | Per-lane | Per-port | Volume year(s) | LightCounting segment share 2026E |
|---|---|---|---|---|
| 400G | 50G PAM4 | 8×50G = 400G | 2022–2025 (volume) → tail | ~10% ⚠ |
| 800G | 100G PAM4 | 8×100G = 800G | 2024–2027 (volume) | ~50% ($13B) ⚠ |
| 1.6T | 200G PAM4 | 8×200G = 1.6T | 2026–2029 (volume ramp) | ~30% ($8B) ⚠ |
| 3.2T | 400G PAM4 | 8×400G = 3.2T | 2028–2030 (forecast volume) | <5% in 2026 ⚠ |
The 800G + 1.6T combined share of the 2026 TAM is 80% ($21B of $26B) — POET’s product strategy aligns to this dominant slice.
2. The silicon-photonics-integration SAM
What is “SAM” for POET specifically?
POET sells Optical Engines — passive silicon waveguide substrates with III-V active devices flip-chip mounted, packaged together — to transceiver-module makers. Not every dollar in the LightCounting pluggable-transceiver TAM is addressable to POET; the market splits roughly into:
| Module-internal architecture | 2026E share | Addressable to POET? |
|---|---|---|
| Discrete optics (separate III-V devices in TO-cans / standard packages, hand-aligned) | ~30% | NO — not the integration approach POET sells |
| Vertically-integrated module-maker silicon photonics (Coherent / Lumentum captive) | ~15% | NO — captive |
| Captive silicon-photonics from Cisco-Acacia / Broadcom Sipho | ~10% | NO — captive to specific OEMs |
| Merchant silicon-photonics platforms (Marvell+Inphi, OpenLight, POET, etc.) | ~25% | YES — addressable |
| Mixed / undisclosed integration | ~20% | Partially addressable ⚠ |
So POET’s SAM = roughly 25–35% of the pluggable-transceiver TAM, depending on how aggressively merchant silicon photonics displaces discrete and captive integration over time.
| Year | TAM ($B) | Merchant-SiPh SAM share | POET SAM ($B) |
|---|---|---|---|
| 2025 | $16.5 | 25% | ~$4.1 |
| 2026E | $26.0 | 28% | ~$7.3 |
| 2028E | ~$40 | 33% | ~$13.2 |
| 2030E | ~$60 | 38% | ~$22.8 |
These SAM-share figures are inferred ⚠ — research-house data does not provide a clean merchant-vs-captive split for the silicon-photonics-integration layer specifically. The directional pattern (merchant share growing from ~25% to ~38% over 2025-2030) reflects the broader industry consensus that vertical integration is becoming less economic as silicon-photonics platforms commoditize.
Co-packaged optics adjacency
CPO is a separate market that overlaps the SAM but has different dynamics:
- 2028E CPO TAM: $2–3B (LightCounting CPO Book) ◐ SemiAnalysis CPO Book
- 2030E CPO TAM: $10B+ (same source) ◐
- POET’s 20-F discusses CPO neither as a current product nor as an explicit roadmap milestone. POET’s Optical Interposer is a wafer-level chip-scale-packaging (WLCSP) approach to discrete pluggable transceivers, not co-packaged optics. POET could technically build CPO modules using the interposer architecture, but the company has not announced a CPO product.
For the time being, CPO TAM is outside POET’s explicit SAM. ⚠ inferred / no management commitment
3. POET’s SOM (Serviceable Obtainable Market)
What is realistic for POET to capture?
POET is competing against larger merchant silicon-photonics platforms (Marvell+Inphi, OpenLight, Cisco-Acacia internal, Broadcom Sipho internal) and against module makers’ own captive integration. POET’s competitive advantages — merchant-only positioning (no competing module business unit), wafer-level integration, multi-data-rate platform — are real but unproven at scale.
A realistic SOM build:
| Scenario | 2028E SOM share of merchant-SiPh SAM | 2028E SOM ($M) | Implied 2028 POET revenue |
|---|---|---|---|
| Bear case | 1% | ~$130M | $130M |
| Base case | 2% | ~$265M | $265M |
| Bull case | 5% | ~$660M | $660M |
For comparison, POET’s fiscal 2025 revenue is nominal (the 20-F discloses operating loss of $63M and continuing R&D + SMA expenses of ~$43M combined; revenue line would have been low single-digit millions but is not separately quoted in the Item 5.A discussion provided here). ⚠ POET-revenue exact figure not pulled into this version
Sensitivity to TAM and SAM
Two-factor sensitivity (POET 2028 SOM share × merchant-SiPh SAM as % of TAM):
| 25% merchant SAM | 33% merchant SAM | 40% merchant SAM | |
|---|---|---|---|
| 1% POET SOM | $100M | $132M | $160M |
| 2% POET SOM | $200M | $264M | $320M |
| 3% POET SOM | $300M | $396M | $480M |
| 5% POET SOM | $500M | $660M | $800M |
The bull case (~$660M revenue at 5% × 33% × $40B TAM) requires three things simultaneously: (a) the 2028 TAM hits $40B per the LightCounting trajectory, (b) merchant silicon photonics displaces captive integration to 33% SAM share, (c) POET captures 5% of the merchant pool. Each of these is plausible individually; their joint probability is materially lower.
The base case (~$265M) requires only the LightCounting TAM trajectory plus modest merchant-SiPh share gains plus a 2% POET SOM — which is consistent with multiple design wins at 2-3 named tier-1 customers.
The bear case (~$130M) is what happens if TAM grows but POET fails to break out of the design-in-only stage at named tier-1 customers — i.e., POET stays a sample / engineering-revenue company through 2028.
What would push POET into the “tier-1 silicon-photonics merchant” category
Three signals would lift the SOM-share upper bound:
- Public confirmation of multiple 1.6T design wins at named hyperscaler-supplying module makers (Innolight, Eoptolink, Coherent, Lumentum) — this would shift base case toward 2.5%+ SOM share.
- Volume-revenue inflection — first quarter of revenue >$10M (vs. current low single-digit-million baseline) — confirms POET has achieved commercial scale.
- A second-source design at NVIDIA / Broadcom switch CPO program — would establish POET at the very top of the merchant SiPh stack and lift SOM-share toward 5%+.
None of these signals had been publicly disclosed as of the 2026-03-31 20-F. ⚠ POET 20-F disclosure cadence
4. The “POET deserves a higher multiple” thesis stress
A POET-bull thesis often invokes a comparison: “POET trades at $X market cap with a $260M+ SAM by 2028; this implies a 2x-3x revenue multiple at the SOM, which is conservative for a silicon-photonics merchant in an AI-tailwind industry.”
The stress test:
- Market cap (2026-04-28 close): ~$8.03 × ~152.7M shares outstanding (post-Jan-2026 raise) = ~$1.23B fully diluted ⚠ shares outstanding inferred from 424B5 disclosure
- Implied 2028 P/Sales (base-case $265M revenue): ~4.6× — within reason for a development-stage micro-cap in AI hardware.
- Implied 2028 P/Sales (bull-case $660M revenue): ~1.9× — would be a re-rate trigger.
- Implied 2028 P/Sales (bear-case $130M revenue): ~9.5× — too high; would require multiple compression.
The current valuation is consistent with the base case revenue path, not pricing in the bull-case explicit. A bull-case re-rate requires breakout volume-revenue evidence in 2026 or 2027 prints.
Capital-raise dilution impact
POET raised approximately $293M net in 2025 (per 20-F Item 5.B) plus an additional ~$143.6M gross in January 2026 (per 424B5 take-down disclosed in timeline). Total recent dilution: ~$437M of fresh capital across 2025+early-2026.
Given a $1.2B fully-diluted market cap, this capital represents 36% of current value — a meaningful runway extension. The runway should fund through 2028 R&D + operational scaling at the current spend level ($31M operating cash burn 2025).
The dilution overhang for the next 12–24 months: any further capital raises would compress equity per share. The two F-3ASR shelves (Nov 2025 + Jan 2026) signal management’s expectation of additional follow-on capacity, but actual issuances depend on share-price levels and customer-engagement progress. ⚠
5. The “what if POET is acquired” frame
Acquisitions of merchant silicon-photonics platforms have been frequent over the past five years:
- Acacia Communications acquired by Cisco for $4.5B (March 2021) — primarily for coherent DSP + silicon-photonics integration. ◐ Cisco/Acacia press release
- Inphi acquired by Marvell for ~$10B (April 2021) — primarily for PAM4 DSP + coherent.
- Polariton Technologies acquired by Marvell (April 2026) — terms not disclosed; plasmonic-organic-hybrid modulators.
- Celestial AI acquired by Marvell for $3.25B base + earnouts (closed Feb 2026) — photonic-fabric for scale-up.
These data points provide a comparable-transactions multiple range:
- Cisco/Acacia at ~$4.5B for ~$500M run-rate ≈ 9× revenue multiple at acquisition.
- Marvell/Inphi at ~$10B for ~$683M LTM revenue ≈ ~14.6× revenue multiple. ◐ Inphi LTM context (cross-section reference)
If POET were to reach the base-case $265M revenue by 2028, an Acacia-comp 9× multiple would imply a ~$2.4B acquisition value. An Inphi-comp 14.6× multiple would imply ~$3.9B. Both are materially above the current ~$1.2B market cap, suggesting an acquirer-arbitrage opportunity if POET successfully delivers the base-case revenue trajectory.
Caveat ◐: The acquisition-comp framework requires POET to (a) reach the base-case revenue trajectory, (b) be acquired rather than continue independently, and (c) achieve revenue multiples consistent with the most-recent acquisitions. None of these is guaranteed.
6. Reading the TAM/SAM signals
For a refresh-cycle analyst tracking POET’s TAM/SAM evolution, the highest-priority data sources to monitor:
- LightCounting quarterly newsletter — the canonical pluggable-transceiver TAM update.
- Dell’Oro and 650 Group — broader market data; Dell’Oro publishes quarterly press releases on optical / data center networking.
- Hyperscaler 10-K / 10-Q capex prints — feed the demand-side of the LightCounting forecast.
- POET 6-K filings — first announcements of named-customer design wins or volume revenue inflections.
- Module-maker quarterly commentary (Innolight private; Eoptolink Shanghai-listed disclosures; Coherent NYSE; Lumentum NASDAQ) — channel-side reads on transceiver demand and silicon-photonics sourcing.
The single most-important POET-specific signal is the first 6-K disclosure of >$10M quarterly revenue from a named tier-1 customer — this is the inflection that would re-rate the SOM share assumption from base case toward bull case.
Cross-section pointers
- industry dynamics — Transceiver-maker customer landscape that defines the SAM addressability.
- ai capex cycle — Hyperscaler capex direction that feeds the TAM forecast.
- regulatory landscape — China-customer regulatory risk to the pluggable-transceiver TAM accessibility.
- overview — Capital runway that underwrites the multi-year R&D scaling.
- bull case — SOM share-gain assumptions feed the bull case directly.
- bear case — TAM compression and SAM-share-failure scenarios feed the bear case.