UK start-up enters the AI hardware arms race


Software and hardware system is based on its IPU.

The artificial intelligence hardware market is getting a new UK-based competitor.

Graphcore, a Bristol-based start-up that creates hardware designed for machine learning applications, raised $30 million in a Series A funding found.

The company says that its mission is to make machine learning faster, easier and more intelligent.

Nigel Toon, CEO, Graphcore said: “Our technology will reduce the cost of accelerating AI applications in the cloud; the same technology will bring AI to low power consumer devices. It will enable recent deep learning applications to evolve more rapidly towards useful, general artificial intelligence.”

The company has created a software and hardware system that it has based on its Intelligent Processing Unit, which it says will accelerate the full range of current and future machine intelligence methods for things like training, inference and prediction.

The company plans to use the money invested to increase its scale and to continue to innovate. The investors in Series A included; Bosch, Samsung, Amadeus Capital, C4 Ventures, Draper Esprit, Foundation Capital & Pitango Capital.

Graphcore is basically trying to break into a market that is currently getting a lot of attention thanks to the advances being made in areas such as driverless cars, robotics, and artificial intelligence.

The company will have a long way to go before it can call itself a challenger to the likes of Nvidia, Intel, IBM, or Arm, all of which have made significant investments into the area. Intel for example recently acquired a chip start-up called Nervana Systems for $400m.

The aim is to tap into the AI market, which IDC research suggests will be worth more than $70bn by 2020. Hardware spending on deep-learning projects is predicted to account for $41.5bn by 2024, up from $436m in 2-15, according to researchers Tractica.

The opportunity is clear and the Bristol-based start-up will be hoping that its approach will win it market share.

The company said that machine intelligence has a very different compute workload and therefore needs a new approach.

“One of the fundamental questions we have addressed at Graphcore, is how to build highly parallel processors, complete with software tools and libraries, that are much faster, more flexible and easy to use, so that developers can explore machine intelligence approaches across a much broader front,” said Toon.

Graphcore was founded in 2014 and has spent the past two years in stealth mode.

Leave a Reply