The AI Boom Has A Power Problem. Nvidia Is Investing Billions To Solve It

As artificial intelligence systems demand more power and faster data movement, Nvidia is betting heavily on photonics technology to overcome growing infrastructure limits.

The AI chipmaker has committed more than $6.5 billion in recent months to companies developing photonics technology systems that use light instead of electricity to transfer data across AI infrastructure.

Since March, Nvidia has committed $2 billion to photonics firms including Lumentum, Coherent and Marvell, CNBC reported. The company has also pledged $500 million to Corning for advanced optical connectivity technology. It joined the $500 million Series E funding round of optics startup Ayer Labs.

Photonics is emerging as an alternative to traditional copper-based electrical connections currently used in servers and networking systems.

“Photonics represents a way for Nvidia to scale their AI infrastructure without the energy costs that staying with electrical and copper will incur,” Alvin Nguyen, senior analyst at Forrester, told CNBC.

The technology allows data to move between GPUs, memory chips, servers and data centers using light instead of electrical signals. Nvidia is increasingly focusing on optical connectivity as it expands its AI networking business.

Brian Colello, senior equity analyst at Morningstar, said Nvidia's “next-generation AI rack solutions” would require much larger amounts of optical connectivity to handle rising bandwidth needs driven by advanced AI models and growing usage.

At Nvidia's GTC conference in March, CEO Jensen Huang said the company was scaling silicon photonics technology across its ethernet networking platform and GPU interconnect systems used in AI factories.

“Which means the amount of silicon photonics technology capacity that we need is substantially higher than the world has today,” Huang said.

He added that Nvidia was working closely with suppliers to expand manufacturing capacity ahead of future demand.

Investor interest in photonics-related firms has surged alongside the AI boom. Shares of Lumentum have climbed 134 per cent this year while Coherent is up 96 per cent. Marvell has gained 122 per cent and Corning has risen 111 per cent.

Nvidia is not the only company increasing investments in the sector. AMD joined the Ayer Labs funding round and previously acquired startup Enosemi while also investing in Teramount and Celestial AI. Venture arms linked to Alphabet and Microsoft also backed photonics startup nEye in an $80 million funding round in April.

But expanding the technology on a large scale remains difficult.

Nick Patience, AI lead at the Futurum Group, said manufacturing optical assemblies is difficult because even small alignment errors between optical and silicon components can disrupt the packaging process. Patience added that large-scale adoption is likely to begin from 2028 onwards.