IQE의 최고경영자(CEO)는 블룸버그를 통해 중국이 이미 AI 하드웨어 제조에 필수적인 소재의 70%를 장악하고 있다고 경고했다.
A Welsh chipmaker has reported a 40% jump in revenue, with its shares up 849% this year. In an interview with Bloomberg, Sofia Gerace reported that Jutta Meier, chief executive of Cardiff-based IQE, warned that “access to indium phosphide substrates is becoming a key risk for the whole semiconductor industry.” Chinese export controls have rendered the supply increasingly uncertain.
Indium phosphide underpins photonic chips, which generate and amplify the light signals that move data around AI data centres. Demand is rising faster than supply. According to the US Geological Survey, China holds 70% of the global supply—a single figure that encapsulates the core vulnerability.
Meier stated that IQE is managing the risk through “broad engagement with different suppliers,” while adding the caveat that “export control licences will still affect IQE if the policies behind them are revoked.” Both points warrant close attention. Diversifying suppliers is the logical response, yet it faces a hard ceiling: 70%. An executive publicly acknowledging this constraint, alongside an 849% stock rally, reflects a level of candor that underscores the severity of the bottleneck. This dynamic mirrors earlier reporting by The Next Web in June, when Beijing began slow-walking export licences and wafer prices rose around 250%. As noted at the time, “The market moved first and the disclosure followed, which is the usual order and the reason prices are worth watching more closely than statements.” Three months later, the chief executive of a major substrate buyer has put her name to the warning, confirming the sequence.
IQE’s financial trajectory highlights the tension between opportunity and constraint. After a 2025 in which revenue fell 18% to £97.3 million, the company initially forecast 20% sales growth for 2026 in July. It has since surpassed that projection and reaffirmed guidance for full-year growth exceeding 30%, a revision made within two months without prior indication during the original forecast. The surge is driven by AI data centre expansion, smartphone sensors, wireless communications, defence applications, and robotics. Wireless recovered in the first half following a weak 2025, though photonics had already overtaken it as the largest part of the business for the first time.
This creates a structural paradox: the same supply tightness constraining IQE’s inputs is simultaneously inflating its order book. Scarcity benefits a supplier only for as long as its own inputs keep arriving. Meier confirmed that the export licence risk is already built into the second-half outlook, signaling to investors that a geopolitical variable beyond the company’s control has been priced in. Her hedge against that exposure is breadth. “IQE’s exposure to AI runs beyond data centres, into robotics and sensing, which she said limits its reliance on the current infrastructure boom.” While strategically sound, this approach does not resolve the geographic concentration of substrate sourcing.
The situation underscores Europe’s structural vulnerability. The continent hosts very few compound semiconductor firms of any scale. IQE is one of them, it is in Cardiff, and the substrate it depends on comes overwhelmingly from China. That is dependency in miniature, and it is more specific than the sovereignty debate usually gets. The argument is normally about fabs and chip design. This is about a raw material, upstream of both, where one country holds most of the world’s supply. It has already shown it will use licensing as leverage. Compounding the concern, a Belgian researcher has been held since May over gallium nitride secrets allegedly passed to China, in another corner of the same compound semiconductor industry.
There is a symmetry here that neither side seems to enjoy. US controls on chip exports pushed Chinese firms towards custom silicon rather than stopping them. Chinese controls on indium phosphide are now pushing Western buyers to hunt for alternative suppliers. Eventually that means alternative sources being developed outside China. Both moves are rational on their own terms. Both accelerate the thing they were meant to prevent, which is a second supply chain owned by the other side. The difference is timing. Substrates take years to bring online, and the AI build-out is not waiting for anyone.
Separately, IQE intends to move from AIM to the Main Market of the London Stock Exchange, to reach institutional capital and improve liquidity in its shares. That is a small line in a story about supply chains, and it points somewhere else. A company on an 849% run deciding it wants institutional money is a company that thinks the run has further to go. It is also a European technology firm choosing to move up within a European exchange rather than out of one, at a moment when the traffic runs the other way. France’s Pasqal listed on Nasdaq in August rather than at home.
Three developments will determine whether this bottleneck translates into systemic disruption or manageable friction. First, whether the indium phosphide licensing tightens further, which would show up in wafer prices before it shows up in any announcement, exactly as it did in June. Second, whether IQE’s second-half results hold the guidance, given that the company has already told the market the licence risk is inside the forecast. Third, whether anyone outside China moves to build substrate capacity. That is the only structural answer, it takes years, and nobody has announced it.
Correction, 8 September 2026: An earlier version of this article said that IQE “cannot source its own substrate”, and described the supply position in more definitive terms than IQE’s chief executive used in her interview with Bloomberg. Jutta Meier said access to indium phosphide substrates is becoming a key risk for the semiconductor industry, and that IQE is managing that risk through broad engagement with different suppliers. She did not say the company is unable to obtain the material.