Rebutan Supremasi AI: Raksasa Teknologi Tiongkok Tantang Dominasi Nvidia
Sanctions imposed by the United States on China’s semiconductor industry over the past few years has spurred Beijing to pursue technological self-reliance. The surge in popularity of artificial intelligence (AI) further intensifies China’s drive to compete with the US for global AI leadership.
Currently, Nvidia, a US-based company, holds a dominant position in the market with its advanced graphics processing units (GPUs). This is because Nvidia designs the core hardware components crucial for training large AI models, such as ChatGPT from OpenAI.
This dominance has propelled Nvidia’s stock price to unprecedented heights. CEO Jensen Huang’s wealth has soared to US$101.1 billion (Rp 1.547 trillion), placing him at the 15th spot on Forbes Real-Time Billionaires list.
Despite facing restrictions on its access to China’s market (with the Biden administration blocking the shipment of high-end chips), Nvidia still manages to send older chip models to China. This hasn’t deterred the Chinese tech giants, who are now actively developing their own alternatives to challenge Nvidia’s hegemony and narrow the technological gap.
Below are some of the leading Chinese tech companies vying for a slice of the AI chip market:
Huawei
Based in Shenzhen, this multinational conglomerate designs the Ascend processors specifically for AI data centers. Huawei sells these chips as part of server systems used in data centers that train AI models.
Their AI server brand is Atlas, and Huawei is currently developing the next-generation chip, the Ascend 910C, which is expected to rival Nvidia’s H100 chip, according to The Wall Street Journal.
Alibaba and Baidu
Both Alibaba and Baidu, giants in their own right, procure chips from Nvidia while concurrently developing their own semiconductors for AI processing.
Baidu has developed its own chip brand, Kunlun, intended for use in both servers and autonomous vehicles. Meanwhile, Alibaba’s semiconductor design unit, T-Head, has introduced the Hanguang 800 chip.
Biren Technology
Like Nvidia, Biren Technology designs GPUs and offers a software development platform for building AI applications on its hardware. These chips, part of the Bili product line, are designed for data center AI training and processing.
Cambricon Technologies
This company specializes in designing a wide range of semiconductors, spanning from chips for AI model training to those for running AI applications on various devices.
Moore Threads
Founded in 2020, this company is developing GPUs specifically designed for training large AI models. They also possess a data center product unit, MTT KUAE, which houses their proprietary GPUs.
Moore Threads is backed by ByteDance, the owner of TikTok, alongside investors such as Sequoia Capital and GGV Capital.
Enflame Technology
This startup positions itself as a viable alternative to Nvidia in the Chinese market. Enflame’s focus is on designing chips for data centers used in both training and processing AI.
Tencent, a leading Chinese tech giant, is one of Enflame’s investors.
The Road Ahead: Fierce Competition in the AI Chip Arena
The race to dominate the AI chip market is heating up, with China demonstrating a clear commitment to narrowing the technological gap with the US. As Chinese companies continue to develop and refine their own AI chips, the global landscape of AI development and implementation is poised for significant transformation. The years to come will likely witness fierce competition between US and Chinese tech giants, shaping the future trajectory of artificial intelligence.
