Huawei is planning to begin mass production of a cutting-edge glass substrate by 2027 to elevate the performance of its Ascend AI chips, according to a new report from DigiTimes. The OEM will join hands with native suppliers, developing the new technology in its home ground, the latest move in Huawei’s broader strategy to reduce China’s reliance on US chip technology.
Glass substrates are flat, ultra-smooth sheets of glass that hold multiple chipsets together. Chinese chip component suppliers are already making TGV (Through-Glass Via) and glass interposer components, technical parts that serve as an important connector between different chipsets and help to improve chip speed and performance.
Bypassing Lithography Constraints Through Packaging Innovation
Since Huawei cannot access EUV lithography machines for chip production due to US export controls, these glass substrate materials will enable it to produce cutting-edge Ascend processors without relying entirely on advanced node scaling. Advanced packaging techniques allow manufacturers to combine multiple “chiplets” into a single, powerful package, effectively increasing performance even if the individual chiplets are manufactured on older, less advanced nodes.
Glass substrates like Interposer and TGV can resist higher temperatures and pressure compared to traditional organic substrates. As a result, these materials will elevate AI chips’ reliability. Huawei adds that these components will make semiconductors 25% thinner and 30% more power efficient, meaningful gains in a data center context where power density and thermal management are critical constraints.
As EastFrontier has previously reported, Huawei’s Ascend 910C demand has surged 20% as DeepSeek V4 pushes to break CUDA dependence, and the company’s 2025 annual report identified AI as its core growth driver. The glass substrate initiative fits squarely within this trajectory: as demand for Ascend chips grows, Huawei needs to continuously improve performance per watt to remain competitive with Nvidia’s offerings.
The Race for Advanced Packaging
The shift toward glass substrates is not unique to Huawei; global semiconductor leaders like Intel are also heavily investing in this technology. However, for Huawei, mastering advanced packaging is not just about performance optimization, it is a strategic necessity to remain competitive in the AI hardware market while cut off from the most advanced manufacturing tools.
China has also been investing in adjacent photonics capabilities. EastFrontier has previously reported on Huawei’s backing of North Ocean Photonics to build domestic optical chip autonomy and China’s launch of a photonics lab to power AI with light, bypassing US chip curbs. Together, these investments paint a picture of a systematic, multi-front effort to build a domestic AI hardware stack that is insulated from Western supply chain disruptions.
It is unclear whether the company will actually implement this plan next year or if the Chinese chip component suppliers will take time to develop glass substrate technologies on a wide spectrum. For the time being, however, the 2027 target represents a credible milestone in China’s push for semiconductor self-sufficiency and a signal that innovation in packaging can provide a viable alternative pathway to continued performance gains in AI hardware.
The broader implication is that China’s semiconductor strategy is increasingly diversified. Rather than betting everything on catching up in leading-edge logic fabrication, a domain where the gap with TSMC remains large and the equipment restrictions are most severe, Chinese firms are pursuing parallel tracks in packaging, photonics, and materials science. Each of these tracks offers a different way to improve system-level AI performance without requiring the most advanced lithography. Huawei’s glass substrate initiative is one node in this multi-front effort, and its 2027 timeline, if met, would represent a meaningful step toward the kind of hardware self-sufficiency that Beijing has made a national strategic priority.
For the global semiconductor industry, the implications extend beyond the US-China competition. If Chinese firms successfully commercialize glass substrate technology at scale, it could accelerate the broader industry’s transition away from organic substrates, a shift that would benefit all manufacturers of high-performance AI chips. Technology developed under the pressure of export controls has a history of finding its way into the mainstream supply chain, and Huawei’s packaging innovations may ultimately prove as consequential for the global industry as they are for China’s domestic AI hardware ambitions.
The 2027 timeline also puts Huawei’s glass substrate ambitions in direct competition with Intel’s glass substrate roadmap, which targets commercial production around the same period. The race between the two companies, one constrained by export controls, the other by the competitive pressure of a resurgent Chinese rival, may ultimately accelerate the global transition to glass substrates faster than either would have managed independently.
That is the paradox at the heart of the US-China technology competition: the restrictions designed to slow China down are also, in some domains, speeding up the pace of global innovation. For investors and policymakers tracking China’s semiconductor progress, the glass substrate timeline is a concrete milestone worth watching. A 2027 mass production date, if achieved, would mark a meaningful inflection in China’s ability to deliver competitive AI hardware without depending on the Western supply chain.