"Industrial MSG" Ignites Secondary Market: Public Funds Sweep Full Indium Phosphide Industry Chain
As the most fundamental and irreplaceable semiconductor cornerstone for AI computing optical interconnection, the Indium Phosphide (InP) sector has been strengthening since Q2, becoming a core track where institutions and hot money are simultaneously increasing positions, showing a trend-like rally with rising volume and prices.
On June 22, Xingye Technology announced a 55 million yuan acquisition of InP-related business assets, subsequently hitting five consecutive limit-up boards. In early June, Suqian Liansheng announced the establishment of a joint venture to develop, produce, and sell InP substrates, and its stock subsequently rose for 17 consecutive sessions. The overall secondary market heat rose simultaneously; the Wind InP index has surged 25.31% since June and hit a new stage high of 4211.71 points on June 25. Trading activity in the sector increased significantly, with average daily turnover exceeding 32 billion yuan in the past month.
"This rally is not a short-term thematic speculation; behind it is the industrial logic of the global AI computing construction wave continuously transmitting upstream," said several public fund managers in interviews with Securities Times reporters. They pointed out that the InP industry chain has extremely high entry barriers and faces severe supply-demand mismatch. Coupled with factors such as accelerated domestic substitution, the industry has a multi-year basis for performance realization, and medium- to long-term allocation value is prominent.
Public Funds Sweep Full Industry Chain
The InP industry chain can be clearly divided into four core links: upstream high-purity indium/phosphorus raw materials, midstream substrates and epitaxial wafers, downstream optical chips and optical modules. As the main line of AI computing optical interconnection transmits from top to bottom, public funds have been continuously mining rare targets upstream along the chain, laying out track dividends in advance.
Since June, the Wind InP index has risen 25.31%, hitting a stage high of 4211.71 points on June 25. The index has risen more than 1,000 points in the past month, and the sector's profit-making effect continues to spread. The sub-sector leader Yunnan Lincang Xinyuan Germanium Industry has risen 279.46% year-to-date, and 32.13% in June alone, with its stock price hitting a record high of 132.88 yuan/share on June 25.
From the first-quarter 2026 public fund holdings data, several leading public funds have already heavily weighted Yunnan Lincang Xinyuan Germanium Industry, including large and medium-sized institutions such as Penghua Fund, GF Fund, Huaan Fund, and China Asset Management, with multiple equity products.
In addition to the substrate leader, funds continue to flow into various segments of the chain. On the upstream indium resource side, companies with high-purity indium smelting capacity such as Xingye Yinxi, Tin Industry Co., and Zhuzhou Smelter Group performed strongly. Tin Industry Co. saw more than ten public funds newly entered or increased positions in the first quarter, among which multiple products managed by famous Jing Shun Chang Cheng fund manager Yu Guang increased significantly. Zhuzhou Smelter Group's stock price rose sharply in the past half month; Yu Guang's three star funds—Jing Shun Chang Cheng Core Competitiveness, Jing Shun Chang Cheng Core Zhao Jing, and Jing Shun Chang Cheng Core Youxuan One-year Holding—also increased positions simultaneously in the first quarter.
In the midstream supporting link, Bojie Technology's stock price doubled year-to-date, benefiting from the dual logic of participating in substrate companies and special cutting equipment. In the downstream optical chip field, Yuanjie Technology, Changguang Huaxin, and Shijia Photon continue to receive increased positions from technology-themed funds.
From a market evolution perspective, the path of this round of AI computing rally is very clear: funds started from GPU computing chips, then through optical modules, focused upward on CPO/NPO/OCS technology routes, traced all the way to the most upstream, and finally locked on Indium Phosphide—its essence is capital betting on the scarcest and irreplaceable underlying material in the computing transmission link.
Irreplaceable "Industrial MSG" in the AI Optical Interconnection Era
"Indium Phosphide is the 'industrial MSG' of the AI optical interconnection era—small in quantity but indispensable," described a Shenzhen-based fund manager heavily positioned in technology tracks.
As a semiconductor material with strong optoelectronic properties, InP has extremely high electron mobility and superior high-frequency performance. Its substrate is a key basic material for EML optical chips and CW light sources, and an important upstream material for high-speed optical modules and CPO. Global tech giants have long made clear its strategic position: Intel CEO Chen Liwu clearly stated that InP is a rigid demand material for high-speed optical interconnection (such as 800G, 1.6T optical modules and CPO) in AI data centers, and is the core to solving the bottleneck of high-speed data transmission between chips.
Industry data shows that currently over 80% of global InP demand is driven by AI data centers. NVIDIA predicts that demand for InP wafers will surge about 20 times between 2026 and 2030; Lumentum expects the compound annual growth rate of AI data center demand for InP to reach 85% by 2030.
Golden Eagle Fund manager Liu Zhongteng told reporters that as optical module rates increase, InP consumption shows a nonlinear explosion. For example, an 800G optical module requires 4 to 8 InP-based EML lasers, while a 1.6T optical module doubles the number of chips to 12 to 16, and InP substrate consumption is 2.7 to 3 times that of 800G; future 3.2T optical module demand will further expand.
However, while demand surges, the supply side faces hard constraints, making it difficult to release sufficient new capacity in the short term, leading to a persistent rigid supply-demand gap. According to Yole, global InP demand will soar to 2.6 million to 3 million pieces in 2026, while effective capacity can only increase to about 750,000 pieces, leaving a gap of over 70%.
"Due to the supply-demand gap of up to 70%, InP prices continue to rise. The price of 2-inch substrates has risen from US$800-850/piece at the beginning of 2025 to recently US$2,300-2,500/piece," Liu pointed out. "Delivery cycles have also extended to 24-40 weeks, and downstream manufacturers need to prepay 30%-50% of the deposit to lock capacity."
Broad Growth Space for High-End Product Independent Development
On June 16 this year, Coherent (COHR) officially broke ground on its 6-inch InP wafer expansion project in Texas, USA. The project relies on the world's first mass-produced 6-inch InP production line, and after completion, the company's own capacity will increase fourfold. NVIDIA invested US$2 billion in March to lock upstream wafer capacity from Coherent, essentially a defensive move to ensure the AI computing supply chain, also indirectly confirming the strategic position of InP in global computing competition.
From a global perspective, the InP market has long been dominated by Japanese and American leaders, with overseas manufacturers holding over 90% market share, and supply chain security risks continue to highlight. In this context, promoting the full independent controllability of the domestic InP industry chain has become a common direction for policy and industry, and domestic manufacturers are facing a rare window to break through.
Currently, several domestic new material companies have achieved stable mass production of 4-inch InP substrates, but 6-inch large-size products are still in the pilot or small-batch stage, with mass production yield and process stability still significantly behind overseas leaders. Core equipment such as crystal growth and epitaxy still depend on imports, and the independent development of high-end InP has broad growth space.
CITIC Prudential Fund researcher Sun Guomeng believes that the current domestic substitution presents a pattern of "small size is stable, large size just starting." "From the latest research, small sizes (2-4 inches) have gained a foothold, and domestic companies have basically achieved large-scale mass production substitution in this area. This is the report card of the first stage of domestic substitution," Sun said. "Breakthroughs in large size (6 inches) are imminent, but volume production remains a challenge. The technical bottlenecks that domestic leading companies can break through in the short term still focus on the single crystal growth stage. How to control the probability of process defects in the invisible crystal growth process is the current core pain point."
Liu Zhongteng further analyzed that to assess the real progress of InP domestic substitution, three core sticking points need attention: First, the substrate link: the yield of 6-inch InP substrates from domestic manufacturers still lags behind overseas. Second, upstream raw materials: InP is smelted from high-purity indium and high-purity phosphorus, and 95% of domestic high-purity phosphorus depends on imports from Japan, constituting an important capacity constraint. Third, equipment: InP crystal growth furnaces, MOCVD epitaxial equipment, and high-end electron beam lithography (EBL) all need to be imported from abroad. For example, in EBL equipment, Vistec (Germany), JEOL, and Nikon (Japan) monopolize 90% of the global market, while domestic manufacturers can currently only produce scientific research-grade low-precision EBL equipment, and the choke point problem remains significant.
Dongwu Securities analyst Wang Zijing pointed out that domestic companies are achieving in-depth synergy through "upstream resource guarantee — midstream core process breakthrough — downstream application verification," not only effectively responding to international controls and raw material price risks, but also providing a solid guarantee for China to achieve key material independent controllability in the market opportunities brought by the computing explosion.
In addition, some interviewed industry insiders expect that relying on the huge domestic AI computing demand market, achieving independent controllability of key semiconductor materials has broad long-term industrial dividend space.
Fund Manager Decision Differentiation
From a hierarchical perspective of industry chain investment value, fund managers' judgments show some divergence.
Sun Guomeng is more optimistic about the midstream substrate wafer link, where current supply-demand contradictions are more prominent, believing this is a highly attractive sector with high earnings elasticity, currently in a good phase of rising volume and prices. On one hand, substrate prices continue to rise; on the other hand, domestic companies are facing a historic window for domestic substitution. If capacity is successfully released, performance elasticity could be very substantial.
However, F&R Fund manager Guo Liangliang pointed out from a performance realization perspective that InP substrates still need to go through capacity construction, yield ramp-up, and customer introduction, with a relatively long realization cycle. Therefore, his focus is on the optical module and optical chip links that have entered a continuous performance realization cycle.
Liu Zhongteng believes that the midstream substrate link may have the greatest elasticity—its market space could grow 20 times from 2026 to 2030, making it the most demand-intensive and highest investment value link. The downstream optical chip link has high barriers and large profit margins, and its future pattern is also relatively good, thus also having high investment value. As for upstream raw materials, they need to be viewed separately: China is the world's major producer of high-purity indium, while high-purity phosphorus has long been constrained overseas, which could also become a potential investment direction.
Some funds also warn of the risk of too fast InP capacity investment. Sun Guomeng pointed out that the current high valuation of the InP link is based on the core logic of substrate shortage and continuous price increases, but it should be noted that this is not a hard shortage but a phased mismatch of supply and demand. As capacity is continuously deployed, InP prices may fall.
Regarding the widely discussed risk of substitution of InP by silicon photonics and gallium arsenide materials, fund managers' views are consistent: there is no "substitution for InP" logic; it is more scenario complementarity, with no substantial substitution risk.
Sun Guomeng believes that in the long run, InP is still the optimal solution in long-distance transmission and has certain irreplaceability. Previously discussed silicon photonics technology routes only replace InP in the modulator and receiver links with thin-film lithium niobate and silicon-germanium chips, while the light source link still requires InP.
