Nvidia CEO Jensen Huang said the chip giant is working with Taiwan Semiconductor Manufacturing Company (TSMC) on silicon photonics technology, but does not expect it to be deployed soon.

“It’s still several years away,” Huang told reporters from Taiwan’s iNEWS in a clip posted to social media by SemiAnalysis. “We should stay with copper for as long as we can, and then after that, if we must, we’ll use silicon photonics.”

Photonic chips use light to transmit data, where traditional semiconductors use electrons. While electronic chips struggle to meet that need, photonics can manage massive data rates at far greater bandwidth and lower latency.

Nvidia is already working on photonic projects, with plans to launch photonic versions of its Spectrum-X networking switches featuring co-packaged optics (CPO) in 2026.

Created with TSMC using a new packaging architecture called Compact Universal Photonic Engine (COUPE), the switches comprise a 65-nanometer electronic integrated circuit (EIC) and a photonic integrated circuit (PIC).

The offering can support intense data transmissions across a variety of distances, with Nvidia touting the CPOs offering 3.5-times more power efficiency and 63-times greater signal integrity versus “traditional methods” upon their unveiling at GTC earlier this year – though the company failed to disclose exactly what it was comparing the hardware to.

While the technology promises massive data rates, Huang’s comments suggest Nvidia is content to extend the life of high-performance copper interconnects before making the jump.

Nvidia’s manufacturing partner has experience working on silicon photonics, having teamed up with Ansys, Synopsys, and Cadence in May 2024 to build out its silicon photonics integration system capabilities.

TSMC is also a founding member of the Silicon Photonics Industry Alliance, a group of some 30 Taiwanese companies working to develop silicon photonics and co-packaged optics (CPO) technologies.

Nvidia has also previously backed a startup developing silicon photonics, Xscape, as part of its $44 million Series A funding round last October. The chip giant joined Cisco’s investment arm in backing the Santa Clara-based startup, which is developing a photonics platform it claims can maximize GPU “escape bandwidth” and scale inference performance for AI workloads.