Baidu Kunlunxin P800
Baidu's third-generation Kunlun chip: 96 GB of HBM and 345 FP16 TFLOPS. Baidu powered up a 30,000-chip P800 cluster in April 2025 to train DeepSeek-class models, and it packs 256 of them into one Tianchi supernode. Sources and verification dates.
Every published figure
Source per value- Owner The company that specified the chip and runs it in its own cloud, data centers or products.
- Baidu 1
- Role What the owner says the chip is for: training, inference, both, or a general-purpose server CPU.
- Training and inference 2
- Design partner The ASIC house that co-designed or physically implemented the chip (Broadcom, Marvell, MediaTek, Alchip, GUC and others), as stated by the owner, the partner, a filing or named trade-press reporting. 'In-house' when the owner states it designed the chip itself, including through an acquired team such as Annapurna Labs.
- In-house (Kunlunxin) 1
- Memory On-package memory type (HBM2e, HBM3, HBM3E, LPDDR5) or, for CPUs, the memory the socket drives (DDR5).
- HBM 1
- Memory capacity GB On-package memory per chip in GB. Blank for CPUs, whose memory sits off-package.
- 96 1
- FP16 TFLOPS Published FP16 throughput per chip or accelerator module in TFLOPS. See the precision note for shared-format figures, sparsity and module scope.
- 345 1
- Precision note What the throughput figures mean: dense or sparse, which precision, per chip or per package, and any figure the owner states only in another unit.
- 345 TFLOPS is Kunlunxin's FP16 figure; the chip also supports FP8 and FP32. 1
- Largest pod chips The largest scale-up domain the owner states: chips per pod, UltraServer, supernode or rack joined by the chip-to-chip interconnect.
- 256 3
- Status Announced: public roadmap without verified deployment. Sampling: first silicon or limited internal use. Preview: customers can request early access, without a GA claim. Deployed: stated production use or general availability. Discontinued: program stopped.
- Deployed 1
- Deployment scale The largest deployment the owner or a named report states: chip counts, clusters, gigawatts or regions.
- 30,000-chip P800 cluster switched on in April 2025 (Baidu CEO Robin Li, via Reuters). 2
- How to get it How an outside buyer can use the chip: a named cloud instance or service, a product, or 'Internal only'.
- Baidu AI Cloud (Baige platform) 3
Not verified in the sources reviewed: partner basis, foundry, node, memory bandwidth, bf16, mxfp8, mxfp4, mx8, mx4, fp8, int8, fp4, power, announced.
Sources
- UDN: Kunlunxin shows M100 and P800 at WAIC 2026 accessed 2026-11-11
- Reuters: China's Baidu says its Kunlun chip cluster can train DeepSeek-like models accessed 2026-11-11
- TrendForce: Baidu rolls out Kunlun roadmap, Tianchi supernodes accessed 2026-11-11