The On-Coming of AI Chips

Arguably the fastest developing technologies at the moment are IoT, AI and blockchain, through constant investment all three are demonstrating the capability to reshape the modern industrial landscape.

Each of them are often interconnected and used across the same industries, but sometimes compete against each other. Currently, we are starting to see the momentum shift to AI and AI chips, which have enjoyed several successes over the past few months.

In April, Facebook announced they were hiring a team of AI chip designers to work with “next-generation technologies within Facebook”. The posting also specifically mentioned AI and ML (machine learning), which many analysts have interpreted to mean a heavy concentration on pushing new hardware. Part of the reason for designing its own chips while very efficient alternatives are on the market already, is fine-tuning its chips to Facebook’s specific use-cases. These include social media and advertising algorithms, as well as data centre information collection.

One of the most popular AI chip systems on the market currently is Nvidia, whose graphics processing units have become the industry standard for ML systems, leading their share price to multiply ten times within just the last three years. Recently the company also announced a partnership with ARM, a British semiconductor and software design company. This huge collaborative deal should enable chip makers to implement deep-learning capabilities into various devices, including mobile phones and IoT gadgets.

There has also been much investment into AI chips: one notable case is SambaNova, who secured $58 million in Series A funding just this year. The company, who are concentrating on developing ML algorithms and software-defined hardware, have support from Google’s venture capital arm, GV, and Walden International, who are notable for funding innovative ideas within technology, internet and IT. It is also worthy of noting that one of SambaNova’s co-founders, Rodrigo Liang, is the former vice-president of processor development at Oracle.

Lastly, AI chip-equipped cameras are garnering much attention in the industry. Cameras that are able to understand what they’re seeing, e.g. accidents at work, or children arriving home safely, can be linked to applications enabling them to alert management or parents of these events. Many start-ups, including Chinese Horizon Robotics, are working to bring this concept to real life. Horizon also has support from executives who previously worked for Baidu, Facebook and Huawei.

Clearly, AI chips are on the rise, and the shift to explore solutions within the field is noticeable. Dashboard believes this surge will provide excellent supporting foundations for future IoT technologies as industry prepares to start mapping and developing a smart future.