Moore threads

Blacklisted Chinese Chipmaker unveils GPU-based computing centre, AI graphics card

Despite being added to the US Entity List in October, the company has announce a new graphics card along with China's first GPU-based computing hub for AI training-with CUDA compatibility.
Share this STORY

Introduction:

In a dynamic landscape where the fusion of technology and geopolitics shapes the future, Beijing-based GPU start-up Moore Threads Intelligent Technology has taken center stage. In October, despite being added to the US Entity List, the company announced a groundbreaking development and Blacklisted Chinese Chipmaker company. The introduction of a new graphics card, coupled with the unveiling of China’s first GPU-based computing hub dedicated to artificial intelligence (AI) training.

Join TechoVedas Community here

Background:

As of December last year, Moore Threads had successfully secured a total funding of US$525.7 million, including a significant investment of 1.5 billion yuan (US$205.4 million) from a Series B funding round. This funding round, led by Hexie Health Insurance and Hechuang Digital Private Equity Fund Management, propelled the company’s pre-market valuation to 28.95 billion yuan, as reported by private equity market tracker PitchBook.

Despite the notable achievements, Moore Threads did not disclose information regarding the foundry responsible for producing its latest graphics cards.

However, Moore Threads did share that they completed the construction of the KUAE computing facility in just 30 days. Notably, the facility is designed to integrate with Nvidia’s CUDA, a parallel computing platform and programming model that the US company developed for general computing using its GPUs.

Additionally, Moore Threads stated that its computing center has the capability to complete the training of a 130-billion-parameter model within 56 days, showcasing its efficiency in large-scale model training.

Additionally, following the US Commerce Department’s announcement of updated tech export controls in October, Nvidia now faces a prohibition on selling its A800 and H800 chips to Chinese companies. These chips, originally designed to comply with Washington’s earlier rules, reflect the impact of evolving geopolitical dynamics on technology trade.

The recent addition of Moore Threads and Biren Technology to the US Entity List has posed challenges for the Chinese AI chip start-ups.

The export controls have made it difficult for these companies to find wafer foundries willing to manufacture their chips, similar to the challenges faced by Huawei Technologies since late 2020.

Additionally, access to US electronic design automation software has been restricted, complicating their operations.

Read More: Soitec Inaugurates Groundbreaking EV Technology Plant in France – techovedas

Product Launch:

Moore Threads unveiled its latest GPU, integrating it into the KUAE Intelligent Computing Centre.To meet China’s escalating demand for large language model training, the company designed this computing hub.

The company claims that the KUAE center, equipped with Moore Threads’ GPUs, can complete the training of a 130-billion-parameter model within 56 days.

Notably, the facility is compatible with Nvidia’s CUDA, a parallel computing platform widely used for general computing with GPUs.

Read More: TSMC Abandons Plans for Most Advanced Chip Fab Amidst Local Protests – techovedas

CEO Statement:

Jams Zhang Jianzhong, CEO of Moore Threads, emphasized the company’s commitment to building a smart computing product line, encompassing chips, graphics cards, and clusters.

“Moore Threads has built a smart computing product line from chips to graphics cards to clusters, relying on the multifunctional computing advantages of GPUs,” 

~CEO, Jams Zhang

Despite the challenges imposed by the US sanctions, Zhang highlighted the multifunctional computing advantages of GPUs and expressed optimism about the company’s future.

Impact of US Sanctions:

The US sanctions have not only affected Moore Threads’ access to wafer foundries but also limited its access to Nvidia’s A800 and H800 chips.

This has led to a surge in demand for Nvidia GPUs among prominent Chinese tech companies, including Baidu, ByteDance, Tencent Holdings, and Alibaba Group Holding.

The heightened interest in developing large language models has fueled a substantial increase in purchases of Nvidia chips.

Strategic Alliances:

In a move to strengthen its position in the AI ecosystem, Moore Threads announced collaborations with 15 partners, including 360 Security Technology, Baidu’s PaddlePaddle, JD.com’s Yanxi, NetEase, Tsinghua University, and Fudan University.

This initiative, known as the Large Model Alliance, aims to foster collaboration and innovation in the realm of large language models.

Conclusion:

Moore Threads Intelligent Technology’s resilience and innovation in the face of US trade restrictions underscore the dynamic nature of the technology industry. The introduction of a new GPU and the establishment of the KUAE Intelligent Computing Centre demonstrate the company’s commitment to advancing AI capabilities in China. As geopolitical tensions continue to shape the tech landscape, the success and challenges faced by companies like Moore Threads provide valuable insights into the evolving dynamics of the global technology ecosystem.

Share this STORY