@rohanpaul_ai: Per The Information, Zhipu AI is also (after DeepSeek) exploring a custom ASIC after GLM-5.2 usage reportedly jumped 27…
Summary
Zhipu AI is exploring a custom ASIC after DeepSeek, as GLM-5.2 usage jumped 27x in one week, highlighting the shift towards specialized hardware for inference at scale in China's AI industry.
View Cached Full Text
Cached at: 07/07/26, 07:36 PM
Per The Information, Zhipu AI is also (after DeepSeek) exploring a custom ASIC after GLM-5.2 usage reportedly jumped 27x in one week.
A custom ASIC removes flexibility, but it can cut power draw and per-token cost.
Nvidia GPUs are strong general-purpose machines, but inference at scale has different economics. A fixed model can run better on silicon designed around its own repeated operations.
Zhipu has not chosen a partner, and the project may take more than 2 years.
The pattern is now bigger than one Chinese lab or one model launch. Chinese AI companies are trying to make software, hardware, and deployment less separable.
Rohan Paul (@rohanpaul_ai): DeepSeek is building an inference chip to cut dependence on Nvidia and Huawei in China’s $50B AI-chip market.
DeepSeek’s chip work is still early, with outside partners and private hiring of chip-design engineers.
The hard part is not drawing a chip, but making it at scale.
Similar Articles
Barron's reports on the spending strategies of Chinese AI companies Zhipu AI and DeepSeek, highlighting their financial moves in the competitive AI landscape.
Barron's reports on the spending strategies of Chinese AI companies Zhipu AI and DeepSeek, highlighting their financial moves in the competitive AI landscape.
Zhipu surges 33% as Wall Street raises bets on China AI after Anthropic curbs
Shares of Chinese AI developer Zhipu surged up to 48% after Wall Street banks raised price targets and as US restrictions on Anthropic's models boosted demand for Chinese alternatives. Zhipu announced its open-source GLM-5.2 model.
Z.ai Served GLM-5.3-Flash Entirely on Chinese AI Chips
Z.ai announced that it served the GLM-5.3-Flash model entirely on Chinese AI chips with per-token costs comparable to Nvidia GPUs, using a custom inference engine optimized for memory-constrained hardware.
@rohanpaul_ai: Brilliant piece by Zhipu Founder Tang Jie. AI scaling is moving past parameter growth. “How many parameters?” is becomi…
Zhipu Founder Tang Jie discusses how AI scaling is evolving beyond parameter count to include factors like training data, compute per forward pass, and post-training, with GLM-5.3 as an example.
Z.ai Built a Gigawatt-Scale AI Data Center (3 minute read)
Z.ai completed a 1-gigawatt data center powered entirely by Chinese-made chips, expanding computing infrastructure for training its advanced GLM models.