PALO ALTO, CA, Aug 20, 2026 – PaleBlueDot AI announced that its NVIDIA HGX B300 cluster achieved NVIDIA Exemplar Cloud status for large-model training workloads. The company met NVIDIA‘s performance requirements across every recipe, exceeding the 95% performance threshold on all tests.
NVIDIA established the Exemplar Cloud program in 2025 as a standard benchmark for providers to validate their infrastructure against, letting buyers compare against a standard rather than a vendor claim.
PaleBlueDot AI’s campaign covered six training workloads: DeepSeek-V3, GPT-OSS, Nemotron-H, Qwen3, and two Llama 3.1 configurations. Every test run exceeded 98% of NVIDIA reference performance. Results held across divergent model architectures, parameter scales from moderate to frontier-class, and multiple numerical precision formats.
Cluster Architecture
The Blackwell Ultra cluster is built on NVIDIA HGX B300 systems. Each compute node has eight NVIDIA Blackwell Ultra GPUs connected through NVIDIA NVLink and NVIDIA Switch, creating an interconnected compute domain within each node.
The cluster uses an 800Gb/s non-blocking NVIDIA Quantum-X800 InfiniBand network. Each GPU has a 800Gb/s connection, for aggregate compute-network bandwidth of up to 6.4 Tb/s per node. The cluster also includes topology-aware scheduling, NVIDIA GPUDirect RDMA, collective communication optimization, and automatic isolation of unhealthy nodes.
“Achieving NVIDIA Exemplar Cloud status on NVIDIA HGX B300 is an important validation of the engineering discipline behind our AI infrastructure,” said Stephen Watts, CEO of PaleBlueDot AI. “Customers need more than access to leading GPUs. They need predictable performance and sustained reliability at scale. We focus on optimizing the full stack, from compute, networking and storage to scheduling and operations, so customers can run their most demanding training workloads with confidence.”
PaleBlueDot AI optimized the infrastructure as an integrated system covering:
- Accelerated computing and system tuning: Hardware configurations aligned with NVIDIA Blackwell Ultra GPUs and the data center’s high-density power and air-cooling design.
- High-performance network architecture: The 800Gb/s non-blocking NVIDIA Quantum-X800 InfiniBand fabric for large-scale distributed training.
- High-throughput storage: A parallel storage system and 63.36TB of local NVMe cache per compute node for training-data loading and frequent checkpoint operations.
- Workload scheduling and resource orchestration: Scheduling logic to improve resource utilization and reduce idle training costs.
- Full-lifecycle monitoring and operations: Automated 24/7 alerting and operational mechanisms to reduce interruptions to long-running training workloads.
Stability Testing
PaleBlueDot AI ran a week-long continuous full-load stability test on the cluster. The non-stop simulation replicated production scenarios in which enterprise training jobs run for weeks, covering compute, networking, storage, and scheduling systems.
The company also deployed high-density power delivery and air-cooling infrastructure to support sustained full-load operation. A quality assurance framework applies hardware burn-in testing, single-node acceptance testing, and cluster-level long-duration stability testing.
“Performance at scale is determined by how well every layer of the infrastructure works together,” added Watts. “Our core focus is translating cutting-edge NVIDIA GPU hardware into standardized, production-ready and reliable computing capacity that enterprises can adopt efficiently at scale.”
The NVIDIA Exemplar Cloud achievement builds on PaleBlueDot AI’s collaboration with NVIDIA, which would continue around bringing next-generation NVIDIA architectures to training and inference workloads.
Source: PaleBlueDot AI
About PaleBlueDot AI
PaleBlueDot AI provides cloud computing services for artificial intelligence workloads. Founded in 2024, the company is headquartered in Palo Alto, CA. Its Token Factory platform provides reserved GPU clusters, dedicated capacity and GPU sourcing. The company also offers customized compute services for organizations with infrastructure requirements. Its PBD Token Router provides access to AI models through one platform. PaleBlueDot AI serves enterprise customers, AI-native companies and startups. Its primary colocation facilities are provided by Digital Realty and hold ISO/IEC 27001 certification, with SOC 2 and SOC 3 reports for facility operations. The company reports more than 58 GPU clusters, 89,160 connected GPUs and operations in 22 regions. It provides account access, cluster requests and pricing information through its website.
About NVIDIA
NVIDIA, founded in 1993 and headquartered in Santa Clara, CA, designs and manufactures graphics processing units, systems on chips, networking hardware, and AI intelligence software such as CUDA. Its products serve industries including gaming, data centers, autonomous vehicles, professional visualization, robotics, health care, and energy. The company introduced the GPU in 1999 and later expanded into accelerated computing and AI infrastructure. In gaming, its GPUs support high-performance rendering, while in AI and high-performance computing, its systems provide the infrastructure for training and deploying large-scale models. NVIDIA also develops tools for robotics and autonomous driving.
