Jane Lanhee Lee

(Reuters) – Computers that use light instead of electric currents to process data, which were only years ago considered research projects, are gaining traction. Startups who have successfully solved the engineering problem of photons being used in chips are receiving large funding.

Ayar Labs is a company that has been developing silicon photonics technology and announced Tuesday that it raised $130 Million from Nvidia Corp.

Although the silicon-based transistor chip has made computing more powerful over the past decade, as transistors reach the width of many atoms and are shrinking further, it is still challenging to make them smaller. You can’t make them so small. Signals also leak between them.

Moore’s law which predicted that the density of transistors on a single chip would double every two decades and reduce costs is slowly fading. The industry must find new methods to address its growing artificial intelligence computing demands.

Data firm PitchBook claims that silicon photonics startup raised more than $750 million last year, which is more than double the amount in 2020. This was $18 million in 2016.

“A.I. A.I. is expanding like mad and taking over large portions of the data centre,” Charles Wuischpard, CEO at Ayar Labs told Reuters. “The energy use in the data movement and the data movement challenge is an enormous problem.”

Problem is, large-scale machine learning algorithms may require hundreds of thousands or more of processors. There is also a problem with data transmission speeds between servers or chips using the current electrical method.

Since decades, light is used to transmit data over fiber optic cables. However, it wasn’t possible to get it down to the chip level because devices for controlling or creating it were not so easy to shrink.

PitchBook’s senior emerging technology analyst Brendan Burke expects silicon photonics to become common hardware in data centers by 2025 and estimates the market will reach $3 billion by then, similar to the market size of the A.I. In 2020, the market for graphic chips will be $3 billion.

Startups using silicon photonics to build supercomputers and quantum computers as well as chips for self-driving cars are raising large funds.

PsiQuantum has raised $665 Million so far. However, the promise that quantum computers will change the world in the future is still many years away.

Lightmatter is a company that builds light-based processors to accelerate AI workloads within the datacenter. It raised $113 million. The chips will be released later in this year, and customers will test them soon thereafter.

Luminous Computing is a company that builds an AI supercomputer with silicon photonics. The startup was backed Bill Gates.

PHOTONIC FOUNDRIES

These startups are not the only ones pushing forward this technology. Photonics is also a growing area of interest for semiconductor companies.

GlobalFoundries head of Computing and Wired Infrastructure Amir Faintuch explained that the collaboration with PsiQuantum and Lightmatter have helped create a platform to produce silicon photonics for other users. It was officially launched on March 1.

Peter Barrett is the founder and investor of Playground Global Venture Capital, a venture capital firm. He believes that silicon photonics will speed up computing in the long term, although it’s a difficult road ahead.

He said, “What Ayar Labs does so well is that they solved the data-interconnect problem for traditional high performance computing.” It will take a while until we can have digital photonic computing for non-quantum systems.