Bristol, UK based chipmaker Graphcore has received $200m in Series D funding from the likes of Microsoft and BMW (iVentures arm, based in Mountain View, San Francisco and Munich).
"Patient capital" provider, Sofina was also involved (whose current Chairman is Sir David Verey, CBE, ex Lazard). This investment values the company at $1.7bn. (Note that Sequoia invested $50m at Series C).
Founders are Nigel Toon and Simon Knowles.
Nigel has been CEO of two VC-backed companies before founding Graphcore, one of which was XMOS in which Graphcore was incubated prior to spin-out. Simon is co-founder and CTO, and has founded and exited two fabless semiconductor companies, Element14 and Icera for a combined value of $1bn. Prior to that he headed microprocessor design at ST Micro.
CNNs (convolutional) and RNNs (recurrent neural networks) are among the skills being hired for in their new round of expansion as well as knowledge of PyTorch and Horovod. Horovod uses MPI (Message Passing Interface) to distribute machine learning load across multiple GPUs.
RNNs have been used to analyze temporal dynamical behaviour such as handwriting recognition and speech recognition.
"Patient capital" provider, Sofina was also involved (whose current Chairman is Sir David Verey, CBE, ex Lazard). This investment values the company at $1.7bn. (Note that Sequoia invested $50m at Series C).
Founders are Nigel Toon and Simon Knowles.
Nigel has been CEO of two VC-backed companies before founding Graphcore, one of which was XMOS in which Graphcore was incubated prior to spin-out. Simon is co-founder and CTO, and has founded and exited two fabless semiconductor companies, Element14 and Icera for a combined value of $1bn. Prior to that he headed microprocessor design at ST Micro.
CNNs (convolutional) and RNNs (recurrent neural networks) are among the skills being hired for in their new round of expansion as well as knowledge of PyTorch and Horovod. Horovod uses MPI (Message Passing Interface) to distribute machine learning load across multiple GPUs.
RNNs have been used to analyze temporal dynamical behaviour such as handwriting recognition and speech recognition.