Chinese optical transceiver makers dodge US ban for now as FCC updates rules

The US Federal Communications Commission (FCC) has published its final decision on updates to its equipment authorisation rules, easing market anxiety over potential bans on Chinese optical transceivers. The rule, adopted in July and finalised on Friday, tightens authorisation requirements for devices containing certain hardware components made by entities on the Covered List – a US blacklist of foreign technology equipment and service providers deemed to pose an unacceptable risk to national security. Optical transceivers, devices located within data centres that facilitate the transmission of data, were not targeted by the ruling, which calmed market speculation after a Reuters report in August said that the Trump administration was drafting measures to block new Chinese models from entering the US market this year. Zhongji Innolight, a leading producer of optical transceivers used in data centres, ended the day up about 4 per cent in Shenzhen and 4.37 per cent in Hong Kong, as peer Eoptolink Technology gained nearly 3 per cent. Suzhou TFC Optical Communication fell 2.62 per cent in Shenzhen. Why are optical transceivers at the centre of the AI race? The regulatory scrutiny comes as optical communications is becoming increasingly important in global artificial intelligence infrastructure, a domain where Chinese suppliers hold a commanding market position. As clusters expand, electrical connections become harder to scale efficiently because of limits in bandwidth, distance and power consumption, making optical links necessary. “The competition is no longer simply about the computing power of a single card,” said Man Jiangwei, director of Huawei Technologies’ advanced optoelectronics laboratory, in an interview with the South China Morning Post on Thursday. “It is increasingly about the scale of the whole cluster.” This shift is accelerating both market growth and technical adoption. The global optical interconnect market was projected to soar from US$24.8 billion in 2025 to US$111 billion by 2030, according to industry research firm LightCounting. While traditional telecoms optical transceiver generations evolve every three to five years, data centre cycles can now churn in less than two years, InnoLight chairman Liu Sheng said during a Goldman Sachs event last week. Goldman Sachs expected global shipments of 800-gigabit-per-second (Gbps) optical transceivers – currently a mainstream specification for large AI data centres – to reach about 45 million units this year and 49 million in 2027. Shipments of the faster 1.6-terabit-per-second (Tbps) generation, meanwhile, are forecast to jump from roughly 33 million units this year to 71 million in 2027, overtaking 800Gbps transceivers within a year. The next 3.2Tbps generation is expected to begin ramping up in 2027. Why is China favouring NPO architecture while US leaders push CPO? To keep pace with rising data rates, the industry is moving optics closer to silicon to tackle electrical bottlenecks between switch chips and optical modules. However, a regional split in technical focus has emerged. Preferred by Chinese players like Huawei, Alibaba Group Holding and Tencent Holdings, near-packaged optics (NPO) keep optical engines in separate packages, placed close to the switch chip, shortening electrical connections while leveraging existing manufacturing ecosystems. Alibaba owns the South China Morning Post. Pursued by US leaders like Nvidia and Broadcom, co-packaged optics (CPO) integrates optical engines directly onto the same substrate as the switching silicon. Nvidia’s Spectrum-X Ethernet Photonics CPO switches are already in production for its Vera Rubin platform, alongside several generations of CPO Ethernet switches from Broadcom. Engineers at Alibaba and Tencent acknowledged that CPO offered the best performance, but argued that NPO was more practical and cost-effective today because it tapped into the mature, pluggable module supply chain, according to LightCounting. Huawei’s Man noted that this divergence stemmed from the companies’ current strengths, rather than geography. Chinese firms excelled at optical module manufacturing, making NPO a natural evolutionary step, whereas CPO relied more heavily on advanced semiconductor packaging and silicon integration dominated by US vendors. Does China really dominate the global optical transceiver market? By volume and headline revenue, China appears dominant. Chinese firms supplied nearly two-thirds of global transceivers by volume and accounted for 60 per cent of optical datacom transceiver revenue in 2025, according to Counterpoint Research. However, that dominance is strongest at the module level and does not extend across every part of the upstream supply chain. High-end optoelectronic parts remain heavily reliant on overseas suppliers. Investment bank Nomura noted that foreign companies continue to dominate critical components like 200Gbps EML laser chips used in 1.6Tbps transceivers, keeping global supply tight. Optoelectronic and electronic chips, including lasers and DSPs, typically accounted for over half of the material cost of a mainstream high-speed transceiver, according to China Insights Consultancy data from Innolight’s Hong Kong initial public offering prospectus published in July.

Chinese optical transceiver makers dodge US ban for now as FCC updates rules

.css-1c6uqr6{color:inherit;font-weight:inherit;font-size:inherit;font-family:inherit;line-height:inherit;overflow-wrap:break-word;}The devices are emerging as central to artificial intelligence race, and China’s suppliers hold key market position

The devices are emerging as central to artificial intelligence race, and China’s suppliers hold key market position