Goldman Sachs raised the target price for Zhongji Innolight to 762:800G, expecting revenue to double next year, and the 1.6T technology iteration is expected to bring a new growth cycle

Goldman Sachs shows strong confidence in the growth prospects of Zhongji Innolight.

On November 20th, according to the news from Chasing Wind Trading Desk, Goldman Sachs stated in its latest research report that it has significantly raised the target price for Zhongji Innolight by 62% to RMB 762, maintaining a "Buy" rating. This aggressive adjustment is based on the company's strong growth potential in the 800G/1.6T optical module and silicon photonics technology fields.

Goldman Sachs predicts that benefiting from the diversification trend of ASIC chips and the surge in demand for high-speed connections, Zhongji Innolight's net profit compound annual growth rate will reach 59% from 2025 to 2028. It is expected that the company's 800G optical module revenue will surge by 104% year-on-year in 2026, and the 1.6T optical module revenue will skyrocket by 110% year-on-year in 2027, supporting an overall revenue compound annual growth rate of 52% from 2025 to 2028.

The firm has raised its net profit forecasts for Zhongji Innolight for 2026 and 2027 by 23% and 28%, respectively, reaching RMB 21.645 billion and RMB 29.944 billion, mainly based on higher revenue and gross margin expectations. The operating profit margin is expected to increase from 25% in 2024 to 39% in 2028, reflecting sustained growth supported by long-term demand trends in AI networks.

The report states that Goldman Sachs' sensitivity analysis shows that a 20% increase in revenue from 800G optical modules could bring a 16% upside to net profit in 2026, highlighting the performance elasticity. Goldman Sachs clearly points out multiple catalysts in the first and second halves of 2026, including the growth of new rack-level AI servers and the deployment of new ASIC AI servers by cloud service providers.

Diversification of ASIC Chips Reshaping Demand Patterns

The report states that the diversification of chip platforms from GPU to ASIC is reshaping the demand for optical modules.

According to Goldman Sachs' global server market research, ASIC chips are expected to account for 38%, 40%, and 45% of the AI chip market in 2025, 2026, and 2027, respectively.

The key technical difference lies in the configuration ratio of optical modules. Leading GPUs typically require 2-3 optical modules per chip, while ASICs, due to their lower single-chip computing capability, may require 3-5 or even more than 10 optical modules per chip.

This architectural difference significantly increases the dependence of ASIC servers on multi-chip, multi-server, and multi-rack connections, thereby generating stronger demand for high-speed connections.

Specification Upgrade Accelerating to 800G/1.6T

Generative AI is accelerating the specification upgrade cycle of optical modules.

The transition from 400G to 800G has become a major driver for 2025, and Goldman Sachs expects this trend to continue in the coming years. The deployment of 1.6T optical modules has also begun in the second half of 2025, which is expected to support long-term demand growth.

Goldman Sachs emphasizes that as optical modules upgrade to higher speeds, market competition will decrease, as this raises the technical barriers for design, testing, and production efficiency, benefiting technology-leading companies to maintain or expand their market share From a design perspective, issues of signal loss, heat dissipation, and power efficiency need to be addressed, and testing requires more complex procedures.

According to Goldman Sachs' forecast, the revenue contribution of 800G/1.6T products will significantly increase from 64%/8% in 2025 to 71%/16% in 2026, and this product structure upgrade will become the core engine of the company's revenue growth.

Silicon Photonics Technology Supports Gross Margin Expansion

Goldman Sachs predicts that the company's gross margin will rise from 41.6% in 2025 to 46.4% in 2027, mainly due to faster speed migration and the transition from EML to silicon photonics technology.

The research report points out that, on a technical level, EML optical modules use expensive EML light sources, while silicon photonics has a more compact and simpler structure, which can reduce component costs and improve production efficiency.

Compared to EML, silicon photonics technology has better gross margin performance, and the impact of export controls will ease after the fourth quarter of 2025, further supporting the company's profitability.

Earnings Forecast Significantly Upgraded

Goldman Sachs has significantly upgraded its earnings forecast for Zhongji Innolight.

Goldman Sachs raised its revenue forecasts for 2026-2027 by 24% and 35% respectively, reflecting accelerated growth in mainstream 400G/800G products.

Although Goldman Sachs holds a relatively cautious attitude towards the first half of 2026, considering that major new ASIC AI server models will begin shipping in the second half of 2026, it still expects revenue in 2026 to grow strongly by 84% year-on-year, further accelerating from a 50% year-on-year growth in 2025.

While slightly adjusting the net profit forecast for 2025, it raised the net profit forecasts for 2026-2027 by 23% and 28% respectively. Compared to the market's general expectations, Goldman Sachs' net profit forecasts for 2026/2027 are 3%/12% higher.

At the same time, the firm raised its gross margin forecasts for 2026-2027 by 1.2 and 1.7 percentage points respectively. Goldman Sachs also introduced its first forecast for 2028, expecting the operating profit margin to reach 39%.

In terms of valuation, Goldman Sachs adopted a target price-to-earnings ratio of 31 times (previously 30 times), and combined with the revision of earnings forecasts and valuation adjustments, raised the 12-month target price from the previous 470 yuan to 762 yuan, a significant increase of 62%.

This target price-to-earnings ratio is consistent with the company's recent trading average, falling between the 9-year average price-to-earnings ratio of 34 times and one standard deviation below the average of 18 times