Unverified50% confidenceFactExact time
NVIDIA's officially supported maximum interconnect configuration for DGX Spark at the time was 3 units.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/4/2026
First Seen
Valid until: 1/2/2027
Sources
Related Entities
Related Claims
VerifiedBoth NVIDIA Triton Inference Server and TorchServe have built-in mature dynamic batching schedulers68% similarUnverifiedNVIDIA is pushing products like RTX Spark to capture a share of the on-device AI hardware market.68% similarVerifiedDGX Spark is a desktop-level AI supercomputer released by NVIDIA at the 2025 GTC conference, featuring the GB10 superchip based on the Grace Blackwell architecture.66% similarUnverifiedDGX Spark基于NVIDIA的Grace Blackwell架构，采用Grace CPU与Blackwell GPU的紧密耦合设计66% similarVerifiedThe CPU and GPU of DGX Spark share up to 128 GB of LPDDR5X memory, achieving memory-coherent access via NVLink-C2C high-speed interconnect.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/974693API
curl https://kongchang.com/api/v1/knowledge/claims/974693MCP
get_claim(id=974693)