NVIDIA adds official eight-node DGX Spark clusters with 1 TB unified memory to run frontier models

KEY POINTS

1. At GTC Berlin NVIDIA announced official support for eight-node DGX Spark clusters, scaling unified memory to 1 TB and enabling local execution of frontier models including the full GLM 5.3 (753B) from Z.ai.
2. NVIDIA also introduced a Spark-optimized collective backend called RoCEnante (merged into FlashInfer) that reduces all-reduce latency—reporting a 4.43x decode speedup at eight nodes (vs 3.05x without it)—and published benchmarks (e.g., Qwen3.8-27B NVFP4 scaling from ~12 to 52 tokens/sec, GLM 5.3 ~80 TPS, and 7.79x pretraining speedup at 97.4% efficiency for a nanochat model); NeMo Switchyard and vLLM integration are highlighted for multi-model routing and distributed inference, and NVIDIA Sync eight-node support is due later this year.
3. This expands local infrastructure capability for running and serving frontier-class models (up to 753B) and shows practical optimizations and scaling trade-offs for multi-node inference and training.