Nvidia Pours $4 Billion Into Photonics to Power Next-Gen AI

Nvidia is investing $4 billion in photonics companies Coherent and Lumentum — $2 billion each — to scale up manufacturing of optical interconnects that will power the next generation of AI data centers. Both deals include multi-billion dollar purchase commitments and future capacity rights for advanced laser and optical networking products.

The move signals that the AI infrastructure bottleneck is shifting from chips to the connections between them. Traditional copper wiring cannot keep up with the bandwidth demands of training frontier AI models across thousands of GPUs, and photonics — using light instead of electricity to move data — is emerging as the critical enabling technology.

A Week That Reshaped AI Hardware Funding

The day after Nvidia's announcement, Ayar Labs closed a $500 million Series E led by Neuberger Berman, with Nvidia and AMD among the strategic investors. The MIT spinout, which develops TeraPHY optical engine chiplets capable of 8+ Tb/sec bandwidth per engine, is now valued at $3.75 billion with $870 million in total funding.

Ayar Labs CEO Mark Wade called it the company's "final round before positioning ourselves for a potential IPO," with customer volume ramps targeted for 2028.

Why This Matters for SBIR and Research Applicants

This $4.5 billion injection into photonics in a single week creates downstream opportunities. DOE, DARPA, and NSF all fund photonics research, and the commercial validation from Nvidia strengthens the case for academic and small business proposals in optical interconnect, silicon photonics, and co-packaged optics. Coherent and Lumentum shares jumped 15% and 12% respectively, reflecting market confidence in the sector. Grant seekers in photonics and AI infrastructure can track relevant opportunities on Granted.