Unverified50% confidenceFactExact time
Training used 3,800 NVIDIA Grace Blackwell GPUs in European data centers, and the public preview also runs on the same infrastructure.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
Unverified前沿模型训练需要极高的GPU间互联带宽，要求训练集群物理上集中在同一地点，不能简单分散到多个数据中心64% similarVerifiedLarge model training primarily consumes computing power from tens of thousands of high-end GPUs (such as NVIDIA A100/H100).64% similarUnverifiedThe EU's European High Performance Computing Joint Undertaking (EuroHPC JU) is deploying multiple supercomputers dedicated to AI training.63% similarVerifiedTechniques such as mixed-precision training and gradient accumulation enable training of larger models under limited GPU memory.62% similarUnverifiedTraining-time methods enhance explanation capability by modifying training objectives, which is costly and requires full access to model weights and large-scale GPU computing resources.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1000080API
curl https://kongchang.com/api/v1/knowledge/claims/1000080MCP
get_claim(id=1000080)