Nvidia is scaling its co-packaged optics (CPO) capacity in response to the limitations of copper interconnects in AI data centers. CEO Jensen Huang announced at GTC Berlin that the company, in partnership with Siliconware Precision Industries (SPIL), aims to increase optical-engine capacity by 15 times by mid-2027, with a further increase to 30 times by the end of 2028.
The CPO architecture integrates optical engines directly alongside silicon, which improves power efficiency and signal integrity. Nvidia is launching its Spectrum-6 CPO switches for its upcoming Vera Rubin AI factories to address bandwidth and energy consumption challenges. These switches are designed to handle the demands of large-scale AI infrastructure.
Nvidia is also investing $1.1 billion to enhance semiconductor testing capacity in the US through a partnership with KYEC, with production scheduled to start in April 2027. Additionally, packaging partner Amkor intends to increase its interposer wafer production capacity by more than 300% by the end of 2028 to support the growing demand for high-speed chiplet connectivity.
