JPMorgan Optical Communications Closed-Door Conference: Key Takeaways from LITE and AAOI Management Discussions
Nashnova编辑部
JPMorgan hosted a hardware and semiconductor closed-door forum in Palo Alto, where rising NPO demand and tight laser supply emerged as common threads across multiple companies; both LITE and AAOI are betting their margin improvement on high-end product ramps, making 2027 the critical validation window.
What exactly is Lumentum's NPO play, and why isn't it just a bridge to CPO?
NPO — near-packaged optics, placing optical components next to the chip package rather than inside it — is positioned as a standalone incremental market, not a temporary bridge to CPO (co-packaged optics).
This means → Lumentum sees NPO as a business in its own right, not a stopgap that vanishes once CPO matures.
CPO currently centers on one customer; NPO already covers roughly six meaningful customers/programs, offers higher optical lane density, and delivers about 85% of CPO's benefit at 25% lower complexity.
In plain terms = NPO captures most of CPO's upside with a simpler design and a much broader customer base — the company is treating it as the main event, not a workaround.
What do LITE's revenue trajectory and order visibility look like?
Management guided NPO revenue at $50–100 million in Q4 2026 and over $100 million in Q1 2027; recent demand revisions skew toward scale-up, but that contribution won't start until H2 2027.
EML — electro-absorption modulated laser, the core light-emitting component inside data-center optical modules — backlog extends through 2027, with negotiations underway for 2028–2029 coverage; pricing is described as trending up.
This means → LITE's order visibility now stretches two to three years out, and prices are still rising — an unusual combination in the optical-component industry.
Management expects EML unit shipments to grow over 50% by year-end; pump lasers hold ~80% market share — one of the highest-margin product lines — and shipment volume is set to quadruple by early 2028.
How do CW lasers and EML relate, and where does the technology cycle go?
Commercial CW lasers — continuous-wave lasers that emit a steady beam while an external modulator encodes the signal — are viewed as a new growth pillar, potentially reaching 10%–20% of the existing EML business; this is additive, not substitutional.
CW attach rates are expected to rise to 50%–60% in the 1.6T generation (versus 20%–30% at 800G), but at 3.2T the mix is expected to swing back toward EML.
This means → CW and EML are not rivals — they trade dominance across speed generations. CW peaks at 1.6T; EML reclaims the lead at 3.2T.
Transceiver (TRx) gross margins still need several quarters to climb; the inflection to mid-to-high-30s% gross margin is expected around early 2027.
What is driving AAOI's near-term growth, and where is the bottleneck?
Near-term growth is led by strong 800G demand, with some contribution from 400G at one major hyperscaler; a second large hyperscaler is ramping 800G, with growth directly tied to manufacturing capacity expansion.
In plain terms = AAOI's problem is not selling — it's making enough product. Through at least mid-2027, production capacity is the binding constraint.
Management is not overly concerned about large customers pursuing custom transceiver paths; the truly constrained component is the laser, and AAOI's vertically integrated in-house manufacturing provides a moat.
Since laser supply tightened late last year, external procurement has been essentially unavailable, pushing the company to accelerate internal expansion and a rapid transition to 4-inch wafers.
How far can AAOI's gross margin improve?
100G is a margin drag; 400G is a slight drag; 800G and 1.6T are accretive.
1.6T is expected to start contributing in Q4 and ramp meaningfully in early 2027; management targets mid-30s% gross margin by mid-2027.
This means → AAOI's margin improvement hinges on how fast the product mix shifts toward 800G and 1.6T — the faster low-end share falls, the faster margins step up.
If CPO opportunities are layered on top, gross margin could reach ~40%; CPO is expected to reach meaningful scale in H1 2028.
What new signals did private companies Ranovus and Avicena bring?
Ranovus (founded 2012, headquartered in Ottawa) spans the full optical stack and expects 2026 revenue to double, with over half of 2027 revenue from NPO; its integrated 3.2T chip bundles drivers, TIA (trans-impedance amplifier), photonic devices, and more into a single package — only the DSP stays external — enabling field-serviceable, swappable CPO.
Avicena takes a more radical path: replacing lasers entirely with microLEDs (microscopic light-emitting diodes). It has shipped the industry's first 1 Tbps microLED kit, fitting 1,500–2,000 fiber cores in under 1 mm².
This reflects two diverging bets in optical communications — Ranovus pursues integrated CPO/NPO on both tracks, while Avicena attempts to disrupt the laser paradigm with a new light source altogether.
Avicena management describes recent customer engagement growth as "beyond exponential" — consistent with the forum's overarching themes of rising NPO demand and tightening laser supply.
Content is for reference only, not financial advice.