INFN: photonic chips being developed…



What does this development mean for INFN?

Even though this was an academic discovery, the article implies it could be put into production quickly using the same manufacturing processes as today’s electrical chips.

I don’t understand this subject well enough to infer possible outcomes of this new technology. Any help is appreciated.


1 Like

What does this development mean for INFN?

Not sure how it will affect INFN. The article mentions 2 start-up companies that are developing the technology. They are SiFive and Ayar Labs:


This is the CEO of SiFive. Interesting background.

I was wondering whether INFN would simply acquire those little companies while they are small, if they perceived a real threat?



I was wondering whether INFN would simply acquire those little companies while they are small, if they perceived a real threat?

Saul, it would take years and hundreds of millions of dollars of research, development, and manufacturing work to make these technologies a threat. Ayar Labs seems to be targeting the data transport market. I think that the bigger threat is that these companies can enable one of the larger competitors (Cisco, Ciena, Alcatel-Lucent, or Juniper) to more effective compete with INFN. I think this would still take several years for an acquiring company to fully develop, scale, and demonstrate the technology so that customers would be comfortable purchasing. However, the big boys will have the resources to achieve this.


There was also some speculation on the boards about Infinera already being research mode for such a photonic processor.

I imagine Infinera has been working on their own version of this process for some time now.

There is some evidence that they have at least thought about it. Whether or not they have a strategy on it… who knows for sure. We’d likely see more strategic hires from the “processor” side of the industry. If we were to speculate, perhaps it was one of the reasons the ex-CEO of Altera was brought onto the board. Either that or for his experience in mergers & acquisitions with his prior company being bought out by Intel.

Here’s the article:…

Integration of Optical and Electronic functions

Here is my take …since silicon could be used for both the electronic and optical functions, this means that we could have computer chips that use light instead of electrons. The idea is great in theory, except that the feature size (i.e. the size of the silicon features used to implement optical functions on the chip) is in the range of 100 nm. Today’s Broadwell CPUs from Intel use a 14nm process that is currently many times too small for optical applications. The good news is that an older 100nm silicon CPU fab could be given a new lease on life to build silicon photonics, but the electronics it could support would be extremely dated, and simply not powerful enough to keep up with the optical functions integrated onto the same chip. To achieve modern electronic performance it would be better to connect the silicon optical chip to a separate silicon electronic chip, which is the architecture you will see being used by silicon photonics proponents.