AXTI因InP衬底短缺及中美价差套利，受益于AI供应链瓶颈，具非对称上涨潜力。

英文原文

The InP substrate shortage and "Spread" Arbitrage - $AXTI Market research is showing InP and optical components to become analogous to the HBM shortage and price craze. The initial bottleneck was caused by $NVDA monopolizing of the production capacity of EML from $COHR to $LITE (TrendForce). Here's the breakdown: $NVDA has reportedly engaged in strategic capacity reservations for EML laser chips, pushing lead times for these specific components beyond 2027. By booking out the finished goods capacity of IDMs like $LITE and $COHR (who are constrained by InP), Nvidia has effectively removed the capacity for optical engines. The "Second Order effect": As a result, many hyperscalers have likely delayed their roadmaps/ramp from $MSFT to $AMZN and are looking for capacity. Optical makers have also scrambled for alternative sources. Since the primary IDMs are booked, hyperscalers are moving upstream to secure wafers, substrate, and feedstock capacity directly, creating a effect that is amplifying demand at the materials level. Both 6N and 7N Indium are at all time highs on the Shanghai Metal Markets; prices sit higher on Western markets, representing a price bifurcation between China and Western markets. 6N for example sits ~$412.83-$425.46 per kg. 6N Western prices are ~$668.00 (Jan 1st, source: Indium Corporation at 5-9 kg quantities), over 50%+ premium. (Spot direct negotiations to hyperscalers would likely have different prices). Premiums for 7N would likely higher margin, with substrates commanding the most premium (not public data, but some estimates go into triple digit premiums) However, the bottleneck has now migrated upstream to substrate itself (and third order effect, feedstock) to secure capacity. $AXTI perhaps plays one of the most critical roles and unique roles as "40% of the InP supply chain" ( $AXTI CEO ER quote) in the InP bottleneck: Low costs: AXTI's factory is in Beijing. They source materials (own 10+ materials suppliers) and refine them themselves into finish substrates. Sell High: They sell finished InP substrates to the global market, where prices are inflated by the shortage and the Western Indium price. Arbitrage: As long as AXTI continues their export permits, they capture the massive spread between low Chinese manufacturing costs and high global selling prices. This is compared to Western companies that are largely affected by upstream material costs and Western premiums. As a result $AXTI, margins next quarter should expand rapidly on top of their production capacity funded by the $100m share sale injection by Northland. This is implicated from reports from late December reports from Asian Metal, indicating that Chinese suppliers explicitly "raised offering prices" for exports to the US: December 19, 2025: Asian Metal reported that "Indium ingot suppliers raise offering prices in United States," marking the beginning of the specific divergence in Western pricing. December 24, 2025: They subsequently reported that "Suppliers of indium ingot raise quotes in China," confirming that the pricing pressure had solidified across both domestic and export channels by the end of the week. This stems from capacity monopolization by hyperscalers, extended lead times, and a decoupling of "AI" material pricing from commodity baselines. The conditions for a "Massive Price Increase" for InP substrates are not just present, they are active. Throughout 2026, both the price of InP substrates, and by one-hop, the commodity price of Indium would likely see a rapid surge as hyperscalers go upstream to secure demand. The company that benefits both ways from the a possible parabolic rally in InP substrate + InP feedstock prices would be $AXTI, presenting an asymmetrical upside at $16, ~$850M MC. However, the risks are present as they're heavily reliant on export permits and US-China trade relations. TLDR, AXT: 1. Sits in the center of both InP substrate and InP feedstock bottlenecks 2. Benefits from margin arbitrage and substrate price surges between China production and US hyperscaler premiums during allocation battles. 3. Controls a major a major percentage of the world's merchant output for substrates required for the Western hyperscaler AI buildout. Hyperscalers saw Nvidia monopolized EML capacity and are likely looking at companies like $AXTI, Sumitomo or feedstock (Vital, $AXTI, $DOWA) to secure their roadmaps now. AXT is one of the most interesting companies at $850m given it controls the critical bottleneck to the hyperscaler AI buildout and captures both the InP substrate price surge + margin arbitrage opportunities.

在 X 上查看原推 ↗