Goldman Sachs has issued a research report maintaining a "Buy" rating on Innoscience (02577), with a 12-month target price of HK$114. This valuation is derived from a discounted EV/EBITDA methodology based on 2030 forecasts, applying a 30.1x target multiple that aligns with comparable companies' trading EV/EBITDA ratios, relative to Innoscience's projected one-year EBITDA growth rate and EBITDA margin. Based on the September 8 closing price of HK$43.20, this target implies a potential upside of 163.9%.
During the bank's China AI roadshow, Goldman Sachs met with Innoscience's management in Suzhou on September 7, with discussions centered on the adoption of gallium nitride (GaN) technology in data centers and product expansion, the company's business model spanning GaN wafers, devices and modules, as well as potential competitive dynamics in the sector. The bank maintains a positive outlook on Innoscience's comprehensive GaN product portfolio, which covers a voltage range from 15V to 1200V, and its diverse end-market applications across multiple segments.
Management indicated that the company has accelerated its GaN capacity expansion efforts, targeting a monthly output increase to 70,000 wafers in the coming years. This expansion is designed to capture the growing demand for GaN in automotive and data center applications, as well as anticipated demand from the robotics sector over the medium to long term. Management also noted that GaN adoption in data centers remains in its early stages, with the company still exploring various business models, including potential partnerships with power device suppliers or establishing direct relationships with cloud service provider (CSP) customers.
Elaborating on the product roadmap, management said GaN products are primarily deployed across three stages of the 800 VDC rack power architecture, specifically the conversion steps of 800V to 54V, 54V to 12V, and 12V/6V to 1V, with the low-voltage products in the final stage commanding higher per-unit value. For this final conversion stage, the company has developed 15V DrGaN products designed to power AI GPUs from 6V down to 1V, and these products have already been submitted for customer qualification. The technology enables higher switching frequencies, improved conversion efficiency, and a reduction in the number of BOM components required.
Goldman Sachs views Innoscience favorably as an early mover in this space, leveraging its accumulated technical expertise to capture growth opportunities as Nvidia's new architectures begin to ramp in volume production.
Comments