2025 Optical Transceiver Trend: 1.6T Has Entered Mass Production, When Should Enterprises Deploy?

2025 has become the "breakthrough year" for large-scale commercialization of 1.6T optical transceivers. Among leading manufacturers, Zhongji Innolight has achieved a shipment target of 3 million units, accounting for over 50% of the global market share. Its self-developed silicon photonics chips boast a yield rate of 95%, and its CPO mass production line is 6 months ahead of competitors. Relying on the differentiated LPO technology route, New Easy Semi launched small-batch delivery in August and entered the large-scale mass production phase in October. Its Thai factory has a monthly production capacity of 200,000 units, with North American market revenue accounting for over 78%.

Industry data shows that the price of 1.6T modules has surged from 1,200 to 2,000 at the beginning of the year, and order backlogs have been extended to 2027. The global shipment volume of 1.6T modules is expected to exceed 1 million units in 2025, and demand will soar to over 5 million units in 2026. Notably, the International Photonics and Electronics Council (IPEC) has initiated the 1.6T SR8 standard project, which will be officially released in 2026 to provide unified technical specifications for short-reach interconnection scenarios.

1.6T optical transceivers have become the core support for the "artery" of AI computing power, with their demand surge driven by three rigid factors:

• Computing Density Leap: A single cabinet of NVIDIA GB300 AI servers needs to be equipped with 162 1.6T modules, which is 3 times higher than the computing density of the previous generation. North American leading cloud vendors (such as NVIDIA and Google) account for over 70% of global purchases.
• Low Power Consumption Requirement: The power of AI data center cabinets exceeds 350kW. LPO/LRO technology can reduce power consumption by 30%-40%. New Easy Semi's 1.6T LPO modules have become best-sellers due to their compatibility with Meta and AWS clusters.
• Policy Infrastructure Catalysis: Driven by China's "Eastern Data and Western Computing" project and the U.S. CHIPS and Science Act, the global intelligent computing center scale will reach 10,670MW by 2027, spawning massive upgrade demands.

1. Leading AI Data Centers (NVIDIA/Google Ecosystem): Launch immediately! The demand gap will reach 10-15 million units in 2026, and locking in production capacity in advance can avoid price increase risks.
2. Medium-Sized Data Centers: Enter during Q4 2025-Q1 2026. Reduce initial investment by leveraging the transition solution from 800G to 1.6T. Trixon's cost-effective 800G modules can serve as a bridging product.
3. Global Enterprises: Prioritize suppliers with multi-regional service capabilities to avoid tariff and delivery delay risks.

As a leading enterprise with 17 years of experience in the optical transceiver industry, Trixon has made adequate technical and service preparations:

• Qualifications and Compliance: Fully certified with ISO14001, ISO9001, and ISO45001. All products comply with international standards such as UL, TUV, FCC, CE, and RoHS, meeting global data center specifications.
• Global Services: With its headquarters in Chengdu and branches in Beijing, Hangzhou, Shenzhen, Hong Kong, and the US, Trixon provides 24/7 localized technical support.
• Smooth Upgrade Solutions: The existing full range of 155M~800G modules is compatible with the 1.6T technology route. Having served over 50 major telecom customers in 56 countries and regions, Trixon can quickly respond to customized needs.

"When 1.6T becomes the standard for computing power, enterprises that deploy in advance will gain a competitive edge." With extreme cost-effectiveness and global services, Trixon helps enterprises overcome the threshold of technological iteration.