UBS Resumes Coverage on Innolight with ¥1,500 Target Price, Views Pullbacks as Buying Opportunity
Nashnova编辑部
UBS resumed coverage of Zhongjixuchuang (中际旭创) with a Buy rating and a ¥1,500 target price, implying roughly 63% upside from the Aug 12 close of ¥921. The bank argues the market is over-discounting cycle and policy risk while under-pricing three demand multipliers and the company's shift from module assembly toward optical engines and chip design.
Where does UBS find 63% upside?
Zhongjixuchuang has pulled back ~35% from its late-June peak; UBS sees the sell-off as over-pricing cycle and regulatory headwinds.
At UBS's forecasts the stock trades at roughly 15.2× 2027 and 10.8× 2028 earnings — a ~50% discount to peers.
UBS anchors valuation at 25× 2027 earnings, yielding the ¥1,500 base target; the bull case is ¥2,011, the bear case ¥370.
This means → UBS is not betting on a short-term bounce but on the market eventually re-rating the stock for its optical-engine and chip-design capabilities.
What keeps optical-module demand growing through 2030?
UBS breaks demand into three multipliers: GPU/ASIC shipments × optical modules per GPU × average selling price per module.
Multiplier one: global GPU/ASIC shipments rise from ~13 million in 2025 to ~28 million in 2027.
Multiplier two: as clusters scale, two-tier Fat-tree Clos architecture — a network topology that links large numbers of servers at high speed — spreads, lifting modules per GPU from 2–2.5 to 4–5.
Multiplier three: spec upgrades push the blended ASP above $500 by 2028–2030. Combined, the global datacom optical-module market could grow from $19 bn in 2025 to $120 bn–$185 bn by 2028–2030, a ~60% CAGR.
In 1.6T competition, who ships at scale first?
UBS argues the moat in high-speed modules lies not in sample launches but in who reaches stable mass production earliest.
Zhongjixuchuang led on 400G, 800G, and 1.6T, shipping at scale roughly six months ahead of its closest rival.
In silicon photonics — chip technology that carries data with light instead of electricity — the company began R&D in 2017; by Q1 2026 silicon-photonics products accounted for ~70% of its high-speed revenue.
UBS projects 2027 shipments of ~21 million 800G and ~31 million 1.6T units, translating to ~25% 800G share and ~40%–45% 1.6T share.
Why does UBS say 3.2T NPO is not a stopgap?
NPO — near-package optics, placing the optical engine next to the switch ASIC rather than on a front-panel plug — outperforms traditional pluggable modules while staying removable and serviceable.
The market reads the roadmap as "pluggable → NPO → CPO" in linear succession, but UBS sees all three coexisting for three to five years. NPO is not a transition product.
In plain terms = CPO delivers the best raw performance but cannot be removed and locks in a closed ecosystem; NPO hits the sweet spot between performance and flexibility.
UBS expects Zhongjixuchuang to begin shipping 3.2T NPO in 2027, contributing ~10% of revenue in 2028 and potentially ~30% by 2030. This means → the company is moving up the value chain from "selling modules" to "selling optical engines + chip design," lifting per-unit value along the way.
How fast can earnings grow — and will margins hold?
UBS forecasts revenue growth of 187% in 2026, 112% in 2027, and 42% in 2028; 2026–2029 revenue CAGR of ~52%.
Profit grows faster: 2026–2028 EPS CAGR of ~69%, roughly 16 pp above comparable peers.
Gross margin is expected to peak at ~45.4% in 2026, then ease to ~43.9% in 2027 and ~42.9% in 2028, remaining above 40% through 2030.
This reflects a key UBS assumption: rising new-product value + leader-scale production advantages + tight supply can offset mature-product price erosion, producing a gentle margin glide rather than a cliff.
How large is the US-policy risk?
In 2024–2025 the US accounted for ~60% of Zhongjixuchuang's revenue; Singapore, the Netherlands, and other markets added another ~30%, making the company highly exposed to US regulatory action.
UBS sees a full ban as low-probability, citing four frictions: Chinese suppliers account for roughly half of global optical-module supply and offshore substitution takes time; DSPs and ASICs are still mainly US-supplied; module assembly margins are thin, limiting US producers' incentive to enter; and China's role in the indium-phosphide supply chain means broad restrictions would disrupt global optical communications.
The company has expanded offshore manufacturing to ~85% of total capacity, but UBS cautions that policy could be enforced by supplier nationality, control, country of origin, or customer procurement rules — each yielding very different outcomes.
In plain terms = UBS judges "soft restrictions" most likely, a full hard ban least likely — but if a hard ban materialises, the entire revenue and profit model must be rebuilt. That tail risk is a core driver of the sharp sell-off over the past month.
Zhongjixuchuang at current levels — undervalued or priced for risk?
BULL
Share gains + value-chain升级
1.6T share holds, NPO ramps early, margins stay above 40% — the market re-rates on optical-engine capability.
Steep valuation discount
Trading at just ~15× 2027 earnings, a ~50% discount to peers.
BEAR
Triple-threat risk
AI capex slowdown, policy restrictions, and price competition could hit simultaneously, undermining both growth and margin assumptions.
Policy uncertainty
60% of revenue from the US; a hard ban would force a full model reset.
In plain terms = the pace of 3.2T NPO ramp will be the key window for resolving the bull-bear debate — fast ramp validates the bull case, slow ramp confirms the bear thesis.
Content is for reference only, not financial advice.