Chinese artificial intelligence chip maker ThinkForce has raised RMB450 million (US$68 million) in series A round from Sequoia Capital China, Hillhouse Capital, Jack Ma-backed Yunfeng Capital and Yitu Technology, an AI start-up specializing in machine vision and hardware, according to media reports.
This is Yitu Technology’s second investment in the AI space this month after joining IDG Capital in a US$15 million round in AI drug maker Accutar Biotech.
Yitu Technology itself raised a RMB380 million (US$55 million) series C round led by Hillhouse Capital in May.
“Yitu started the AI discussion with ThinkForce a long time ago. We think in the era of AI, China and the U.S. will be the globe’s dual engines. Chip technology and talents are China’s strategic priority and we want to have a world-class chip team,” said Jiao Huiru, director of strategic development at Yitu, “We believe there is industry asymmetry in cross-disciplinary developments, posing great barrier for the multidisciplinary developments in AI.”
On its website, ThinkForce claims that core members of its team have worked for top chip firms such as IBM, Intel, Broadcom, AMD and ZTE. It also says members of its team have participated in projects such as IBM’s high performance microprocessors IBM PowerPE, Microsoft XBOX game console chip, Sony PS3 Cell processing Device, and 56G Serdes.
ThinkForce focuses on building intelligent semiconductors with AI algorithms and AI hardware platform. ThinkForce plans to launch a chip with Manycore processors. Manycore processors are specialist multi-core processors designed for a high degree of parallel processing, containing a large number of simpler, independent processor cores.
ThinkForce says that its technology combined with its self-developed firmware will accelerate calculation of various types of neural network models, achieving more than five times the power consumption and cost saving compared with the mainstream AI chips.
Chip virtualization is another focus of ThinkForce. Chip virtualization enables a central processing unit to act as if you have several independent computers, in order to enable several operating systems to run at the same time on the same machine.
“Chip virtualization technology can double the AI cloud usage rate. We have worked with Firmware in chip architecture to ensure the chips’ virtualization capabilities in cloud applications,” said Zhenning Zhang, a vice president at ThinkForce.