GB300 NVL72
PulseAugur coverage of GB300 NVL72 — every cluster mentioning GB300 NVL72 across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
NVIDIA details AI factory strategy for maximizing ROI
NVIDIA is detailing its strategy for building AI factories designed for maximum return on investment, focusing on productivity, durability, and fungibility. These factories aim to maximize earning capacity by delivering…
-
Nvidia CEO Jensen Huang accused of underpromising hardware performance
A report from SemiAnalysis claims that Jensen Huang, CEO of Nvidia, is significantly underpromising the performance of their new hardware. While Huang stated at GTC 2026 that the Rubin NVL72 would offer 3x better perfor…
-
NVIDIA Vera Rubin NVL72 system debuts with leading MLPerf Inference v6.1 performance
NVIDIA has announced leading performance for its new Vera Rubin NVL72 system in the MLPerf Inference v6.1 benchmarks. The system demonstrated up to 3.7x higher throughput than its predecessor, the GB300 NVL72, on demand…
-
GB300 NVL72 offers 13x better performance per dollar than Hopper for agentic inference
SemiAnalysis reports that the GB300 NVL72, utilizing an innovative copper backplane and NVLink, offers a 13x improvement in performance per dollar compared to NVIDIA's Hopper architecture for agentic inference. This adv…
-
AI rack density capped by power and cooling, not chips
The density of AI server racks is increasingly limited by power delivery and cooling infrastructure rather than the number of chips. As AI hardware like NVIDIA's GB300 NVL72 and Vera Rubin NVL72 push power requirements …
-
Cursor's MoK redefines MoE execution, optimizing GPU kernel for faster AI training
Cursor has developed and open-sourced Mixture-of-Kittens (MoK), a new software layer designed to optimize the execution of Mixture-of-Experts (MoE) models. MoK addresses inefficiencies in token scheduling, inter-GPU com…
-
GPU rental costs rise as AI infrastructure complexity outpaces standardization
The cost of renting GPUs, even older models like the NVIDIA H100, has not decreased as expected due to the complex systems required for AI tasks. Unlike traditional computing, AI workloads depend heavily on the integrat…
-
Nvidia Vera Rubin boosts agent throughput 30x with heterogeneous system architecture · 1 source tracked
Nvidia is optimizing its Vera Rubin NVL72 system for agent workloads, achieving up to a 30x increase in throughput per megawatt compared to previous generations. This performance boost is not solely due to faster GPUs b…
-
NVIDIA launches NVLink Fusion to integrate custom XPUs into AI factories
NVIDIA has introduced NVLink Fusion, a technology designed to integrate custom-designed XPUs (eXponential Processing Units) with NVIDIA's existing AI infrastructure. This solution aims to accelerate the development and …
-
NVIDIA Vera Rubin NVL72 boosts AI agent efficiency by 30x, expands ecosystem with MediaTek deal · 10 sources tracked
NVIDIA has unveiled its Vera Rubin NVL72 system, which reportedly offers up to 30 times greater throughput per megawatt for AI agent workloads compared to previous NVIDIA GB300 NVL72 systems. This significant efficiency…
-
Anyscale Ray optimizes NVIDIA GB300 NVL72 with NVLink Domain placement
Anyscale has introduced NVLink Domain-Aware Placement Groups for its Ray framework, designed to optimize performance on NVIDIA's GB300 NVL72 systems. These new placement groups ensure that tightly coupled actors are sch…
-
Cursor releases open-source MoE training megakernel, Mixture-of-Kittens
Cursor Research has open-sourced Mixture-of-Kittens (MoK), a specialized training kernel designed for Mixture-of-Experts (MoE) models. This megakernel fuses MoE communication and computation into a single deterministic …
-
AI servers demand massive memory: 20TB HBM and 17TB RAM per unit
A single 72-GPU GB300 NVL72 AI server requires over 20 terabytes of high-bandwidth memory, primarily for GPUs and other AI chips. Additionally, these servers utilize 17 terabytes of LPDDR5X RAM, similar to that found in…
-
Anyscale to join Nscale, boosting Ray integration with infrastructure
Anyscale, a company known for its distributed computing framework Ray, has announced its definitive agreement to join Nscale. This acquisition aims to deepen the integration between Anyscale's software optimizations for…
-
NVIDIA GB300 NVL72 sets world record in AI pre-training
NVIDIA has set a new world record for Mixture of Experts (MoE) pre-training using its GB300 NVL72 system. This advancement pushes the boundaries of large-scale AI training capabilities.
-
Kimi K3's massive scale demands intensive networking despite optimizations · 8 sources tracked
SemiAnalysis reports that the Kimi K3 model, with its 2.8 trillion parameters, requires significant network bandwidth despite optimizations like Kimi Delta Linear Attention (KDA). The model's architecture necessitates t…
-
Kimi K3 2.8T model requires advanced hardware beyond NVIDIA DGX B200
The Kimi K3 2.8T model is exceptionally large, requiring specialized hardware beyond a single NVIDIA DGX B200 system. To accommodate its size, configurations involving GB300 NVL72, B300, or MI355X systems are necessary …
-
NVIDIA touts Blackwell platform's performance per watt for AI infrastructure
NVIDIA is emphasizing performance per watt as the critical metric for AI infrastructure, especially with the rise of agentic AI and Mixture-of-Experts (MoE) architectures. The company highlights its Blackwell NVL72 plat…
-
NVIDIA GB300 NVL72 firmware bug requires frequent reboots
A firmware bug in NVIDIA's GB300 NVL72 system requires racks to be rebooted every 66.5 days. This issue highlights ongoing problems with NVIDIA's driver and firmware quality, despite the company generally being perceive…
-
NVIDIA Blackwell platform dominates MLPerf Training 6.0 benchmarks · 4 sources tracked
NVIDIA's Blackwell platform has achieved top performance across all seven benchmarks in the MLPerf Training 6.0 industry standard tests. The platform demonstrated the fastest training times and enabled the largest-scale…