NVIDIA currently occupies an absolute dominant position in the global AI chip market. It has long dominated the computing power supply chain with its technology and ecological monopoly. The industry has been suffering from NVIDIA for a long time. Recently, there is news that ByteDance has officially entered the field of self-developed chips. According to the latest revelations from people familiar with the matter, ByteDance’s chip R&D team will begin large-scale recruitment.
At present, the core focus is on the chip design process, carrying out dedicated hardware customization and optimization around the company's own business, and developing a variety of complex chips using advanced semiconductor processes for cloud scenarios to improve performance and reduce computing power costs.
"At present, the Byte Chip team has achieved multiple successful first-version tape-outs, and many early projects have entered the mass production deployment stage, covering multiple mainstream advanced process nodes. The overall R&D and implementation rhythm is steadily advancing." said the person familiar with the matter. In response, Byte did not respond.
Previously, it was reported that ByteDance is promoting self-developed artificial intelligence chips, project code-named SeedChip, and is currently negotiating related cooperation with Samsung Electronics.
Another source revealed that ByteDance aims to obtain chip samples before the end of March and plans to mass-produce at least 100,000 AI inference chips this year, with subsequent production capacity gradually increasing to 350,000.
As a technology company with a relatively comprehensive AI layout, ByteDance's core business is highly dependent on AI technology. In recent years, it has continued to focus on cutting-edge areas such as large models, and launched AI products such as bean bags.
The company has already increased investment in the underlying computing infrastructure and built a large-scale AI computing cluster.
According to previous public reports, ByteDance has launched self-research on cloud training and inference chips since 2022, with the intention of building a complete and independent AI technology stack and reducing dependence on external chip supply chains.
The current global AI competition is becoming increasingly fierce, and AI chips have become a battleground for technology giants.Overseas Google TPU, Amazon Inferentia, domestic Alibaba Pingtouge, etc. are all self-developed AI chips by manufacturers. The core logic is to reduce costs, improve efficiency and reduce over-reliance on external suppliers through self-research.
