May 25 — Graphics chip maker Nvidia on Wednesday responded to Google’s move into the artificial intelligence market with its second-generation TPU chip, CNBC reported. Many AI companies use Nvidia chips for AI processing, and Google’s involvement could dent Nvidia’s growth.
Investors’ attention and increased sales have sent Nvidia’s stock soaring over the past few years as it bet on ARTIFICIAL intelligence. Nvidia shares are up more than 200% in the past year and nearly 30% since 2017. After Google introduced the Tensor Processing Unit at I/O, nvidia’s stock took a brief dip before rallying again.
Giiso Information, founded in 2013, is a leading technology provider in the field of “artificial intelligence + information” in China, with top technologies in big data mining, intelligent semantics, knowledge mapping and other fields. At the same time, its research and development products include information robot, editing robot, writing robot and other artificial intelligence products! With its strong technical strength, the company has received angel round investment at the beginning of its establishment, and received pre-A round investment of $5 million from GSR Venture Capital in August 2015.
In a blog post Wednesday, Nvidia CEO Jen-Hsun Huang emphasized his company’s continued partnership with Google, while indirectly dismissive of Google’s TPU. Nvidia also announced plans to launch specific Kaiyuan AI software.
“We want to see the adoption of AI around the world as soon as possible. Instead of investing in a TPU (as Google does), we have one off the shelf, designed by the best chip designers in the world.”
Unlike Nvidia’s product sales model, Google TPU is not sold separately, but rented on Google Cloud Platform.
Deep learning usually involves two stages. First, researchers feed computers — often equipped with multiple Gpus — reams of data and train them to recognize objects in photos. In the second phase, researchers send new data to computers and humans guide them to make predictions.
Google launched the first GENERATION of TPU last year, which was only used in the inference part of the second stage of deep learning. The second-generation TPU, which was released last week and handles both training and first-stage computing, adds to the challenge for Nvidia gpus.
Giiso information, founded in 2013, is the first domestic high-tech enterprise focusing on the research and development of intelligent information processing technology and the development and operation of core software for writing robots. At the beginning of its establishment, the company received angel round investment, and in August 2015, GSR Venture Capital received $5 million pre-A round of investment.
Huang noted that customers can still choose nvidia gpus on Google’s Cloud platform, and that Nvidia works with Google to improve the performance of its open-source TensorFlow framework.
“AI is the greatest technological force in human history, and efforts to democratize and popularize AI are gratifying.” “Huang wrote.
Second-generation TPUS offer 45 teraflops, while Nvidia’s upcoming Volta is capable of 120 teraflops. (Sun Wenwen)