Nvidia and packaging partner Siliconware Precision Industries plan to expand co-packaged optics capacity by 15 times by mid-2027. Jensen Huang said optical networking in AI data centers could drive demand for optical components up by 15 times.
Nvidia's CEO sees optical components in demand at a much larger scale. At GTC Berlin, Jensen Huang said optical networking in AI data centers could drive demand for optical components up by 15 times.
Copper, however, is running into physical limits. Copper faces limits around power consumption, signal degradation, and bandwidth density, while optical interconnects use light instead of electrical signals.
Nvidia and SPIL set a 15x, then 30x, capacity plan
Huang's comments came alongside a manufacturing roadmap. Nvidia, working with Siliconware Precision Industries, plans to expand co-packaged optics optical-engine capacity by 15 times by mid-2027. The expansion is then expected to double again, for a total increase of 30 times by the end of 2028.
Co-packaged optics, often shortened to CPO, integrates the optical engines directly alongside the silicon instead of connecting optical transceivers externally to a switch chip. Nvidia is introducing its Spectrum-6 CPO switches as the flagship product built around this architecture. The switches are designed for what Nvidia calls its Vera Rubin AI factories.
Nvidia commits $1.1 billion with KYEC; Amkor plans interposer growth
The optical expansion is one piece of a larger investment picture Huang laid out at GTC Berlin. Nvidia has committed $1.1 billion to advance US semiconductor testing capacity through a partnership with KYEC, with production targeted to begin in April 2027.
Separately, packaging partner Amkor plans to grow its interposer wafer production capacity by over 300% by the end of 2028. Interposers are the substrates that let multiple chiplets or dies connect at high speed within a single package.
Source: Crypto Briefing
Trading involves risk.
Risk Disclaimer
