GPU entrepreneurship "high fever does not retreat", Moore thread completed the A round of 2 billion financing

Reporter | Peng Xin Edit |

Investment in the GPU (graphics processing) field is "high fever does not go away".

On November 25, moore thread, a domestic GPU chip start-up, announced the completion of the A round of 2 billion financing, which was jointly led by Shanghai Guosheng Capital, Wuyuan Capital and Bohai Zhongsheng Fund, and jointly invested by nine institutions such as CCB International, Qianhai Fund of Funds, China Merchants Securities and Hubei High-quality Development Industry Fund. According to the Tianyancha APP, Tencent is also a participant in this round of investment.

This is the third round of financing after the establishment of Moore Thread in October 2020, with a total of billions of yuan, and the funds raised in this round will focus on the mass production and manufacturing of the first GPU chip, THE IP research and development associated with GPU SOCs, and the expansion of the domestic GPU ecosystem. Moore Thread's last Pre A+ round of funding was in February, with three rounds of funding in less than a year. At the same time as completing this round of financing, Moore Thread also announced that the first domestic full-featured GPU chip was successfully developed as scheduled and has begun to adapt to domestic mainstream CPUs and operating systems.

According to Moore Thread, it has a software and hardware design team that designs a modern full-featured GPU architecture, and the chip has built-in self-developed 3D graphics computing cores, AI training and inference computing cores, high-performance parallel computing cores, ultra-high-definition video codec calculations and other cores. In terms of the founding team, the core members of the company are mainly from NVIDIA, AMD, ARM, etc., the CEO is zhang Jianzhong, the former general manager of NVIDIA China, and CTO Zhang Yubo has worked at NVIDIA and has a deep understanding of graphics and GPU parallel computing.

Moore Thread said that it has established cooperative relations with hundreds of ecological partners to jointly promote the joint development, performance optimization and application innovation of domestic GPU application software, and plans to empower GPU technology to key areas of digital economy development such as digital twin, industrial simulation, digital cultural creation, smart energy, smart city, smart medical, autonomous driving, robots, digital people, and biocomputing.

From 2020 onwards, GPUs have become one of the most gold-absorbing tracks in the domestic chip field, and emerging entrepreneurial GPU companies include Biling Technology, Muxi, Moore Thread, Tianshu Zhixin, Denglin Technology, etc.

Founded in September 2019 and headquartered in Shanghai, Biling has raised 4.7 billion yuan from more than 40 investment institutions, with a valuation of more than 10 billion yuan, and Hillhouse Venture Capital, Walden International, Ping An of China, China Merchants Capital, BAI Capital, and Guosheng Group State Reform Fund are all on the list of investors. On August 25, GPU manufacturer Mu Xi announced that it has completed a round of financing of 1 billion yuan, jointly led by China State-owned Enterprise Structural Adjustment Fund Co., Ltd. and China Internet Investment Fund, and the old shareholders Jingwei China, Heli Capital, Sequoia China, Lightspeed China, Guochuang Zhongding, Smart Interconnection Industry Fund, Shanghai Science and Technology Innovation Fund, Lenovo Venture Capital, China Merchants Jintai, Fosun Ruizheng, Oriental Fuhai, and Chuangtu Investment. On November 4, Denglin Technology announced that it has received joint strategic investment from Qualcomm Ventures and several top companies in the industry, but did not disclose the amount of financing.

GPUs are accelerators that have a fundamental bearing effect on the entire artificial intelligence industry. The core of AI is algorithms, computing power and data, all three are indispensable, when the industry's demand for AI is increasing, the existing computing power is difficult to meet, manufacturing GPUs, can better serve users. At present, NVIDIA is the semiconductor manufacturer that has benefited the most in the AI era. Its main GPU is the graphics processor, originally designed for game and video rendering, and later became an indispensable infrastructure for AI computing, known as GPGPU (General Purpose Graphics Processor): GPGPU removes the graphics rendering part and is born to handle accelerated applications such as artificial intelligence and professional computing.

In the GPU technology route, each family has a choice, Moore thread, Biling Technology wants to take into account the GPU's accelerated computing and graphics functions, MuXi, Denglin first started from the GPGPU, focusing on accelerating computing, and entering the data center market with the strongest localization demand.

Once each company starts the mass production and landing of the first product, the gap will appear, some companies with poor products will fall behind, and start-up GPU companies need to race against time. In terms of research and development progress, companies are expected to achieve GPU tape-outs this year and next year, and provide tests to customers. At the same time, GPU start-up products also need large customer verification, such as Tencent, ByteDance, Ali and other companies with large data centers, if they participate in GPU companies, it will promote product verification and testing.