Facebook To Develop Its Own Silicon to Power Devices and Servers

• Facebook is in the early stages of developing its own version of silicon for devices and servers.
• Facebook’s job listings indicate the company will put efforts into artificial intelligence servers.
• The move could allow the company to make improvements to their development process.

Facebook is looking to hire a manager for developing its self-branded chips for company devices and servers. According to their job listings, the social media giant is looking for a manager for assembling an end-to-end SoC/ASIC, firmware, and driver development organization.

Facebook’s interests are in making the company more self-reliant instead of partnering with outside companies like Qualcomm or Nvidia. Apple is one of the largest companies to adopt self-branded chips for its iOS devices and is set to replace Intel chips on Mac computers with Apple-branded silicon. Google also produces custom artificial intelligence chips, including the image processor found in the Google Pixel. Amazon has also been developing chips to make their proprietary smart assistant Alexa function better.

Image Courtesy of Embedded Computing Design

Facebook’s job listings fall under the infrastructure category, which makes it very likely that the team will be working on artificial intelligence servers. The social media platform has been working towards artificial servers with Nvidia over the years, and it seems like they want to move on to self-reliant development.

Mark Zuckerberg mentioned in the past that artificial intelligence is something Facebook is focused on to help the social media company flag inappropriate content across the social media network. With 2.2 billion monthly active users in the last quarter of 2017, artificial intelligence can help Facebook be more effective at dealing with inappropriate content.

Tight software and hardware integration with self-branded chips could also help Facebook develop smart speakers or virtual reality headsets. Effective improvements and optimizations to hardware are only possible with a company being part of the entire development process. The social media company claims that their deep learning AI is already 99% efficient at taking down terrorist content. However, there are flaws in the system, and one of the biggest problems is that it doesn’t detect fake news.

With the recent Cambridge Analytica scandal causing people to accuse Facebook of spying on them, there might be trust issues if the company starts putting its chips in consumer-grade appliances.