Jakarta, INTI - Chinese artificial intelligence (AI) startup DeepSeek is reportedly designing its own chip to power its AI systems.
According to a Reuters report, the chip is designed specifically for the inference phase, the phase where an AI model generates responses to user queries, rather than for training new models from scratch. DeepSeek has not yet issued an official response about the matter.
DeepSeek's move aligns with the trend where many of the world's leading AI developers are reportedly producing their own chips. This strategy aims to gain a competitive advantage by tailoring hardware to work as efficiently as possible with their systems.
Global Competition in Chip Development
Last month, OpenAI launched its first custom chip in partnership with Broadcom. Meanwhile, Amazon.com is in negotiations to sell specialized data center chips to other companies, and Alphabet has also seen a significant surge in demand for its Tensor Processing Units (TPU) chips.
DeepSeek has become a key driver in China's ambitions to compete in the global AI industry. Last year, the company developed an AI model that they claimed to require minimal resources and chips to build.
This AI development demonstrated that Chinese companies can compete with those from the US or other countries, despite pressure from US export restrictions on advanced devices.
Conclusion
Chinese AI startup DeepSeek is reportedly developing a standalone chip for inference processing to improve its AI system efficiency. This aligns with the global trend of tech giants like OpenAI, Amazon, and Alphabet designing their own custom chips. This strategic move confirms DeepSeek's position as a key driver of AI computing in China.
Read more: BRIN Integrated Electrochemical Sensors and AI to Develop Smart Sensing Technology