Provide Chinese cloud providers integrated AI supercomputer clusters that can train and serve billion-parameter and larger foundation models domestically.
Launching 2026. Tianchi 512 trains trillion-parameter models at cloud scale.
Checked 11 Oct0 updates this week
Tianchi bundles Kunlunxin processors with high-speed interconnect, power and cooling, ready to deploy as a unit.
Multiple Tianchi nodes link with optimized networking to distribute training, supporting models too large for one node.
Kunlunxin provides software frameworks tuned to large model training, based on Baidu's 10+ years of internal development.
| Spec | Tianchi Supernodes |
|---|---|
| Tianchi 512 training capacity | Trillion-parameter modelsR (reported) |
| Tianchi 256 training capacity | Hundred-billion+ parameter modelsR (reported) |
R reported by the company
Baidu Kunlunxin
AI chips, M100, M300, Tianchi, Chinese GPU
Kunlunxin is Baidu's semiconductor subsidiary, developing a full-stack AI processor portfolio from training to inference. In 2026, Kunlunxin filed for Hong Kong and Shanghai stock exchange listings, with valuation toward HK$30-100B. The company ships over 50% of its revenue to external customers.
