ASUS has announced that its AI POD, featuring the powerful NVIDIA GB200 NVL72 platform, will begin shipping in March 2025. With production ramping up, ASUS is reinforcing its role as a leader in AI innovation by offering a comprehensive range of AI solutions, from supercomputing to cloud services. The company aims to drive AI adoption across various industries, helping businesses speed up their time to market with tailored solutions.
Proof of concept funded by ASUS
ASUS is launching a proof of concept (POC) for the ASUS AI POD, powered by the cutting-edge NVIDIA Blackwell platform. This POC will provide a select group of innovators and enterprises with the chance to explore large-scale AI and deep learning capabilities. The POC initiative, fully funded by ASUS, allows participants to access state-of-the-art AI solutions while being guided by NVIDIA GB200 specialists.
Through this programme, ASUS aims to demonstrate the potential of its technology and provide enterprises with a first-hand look at the benefits of AI innovation.
ASUS AI POD with NVIDIA GB200 NVL72
ASUS AI POD, equipped with the NVIDIA GB200 NVL72 platform, delivers outstanding performance for AI applications. By integrating GPUs, CPUs, and NVIDIA’s latest interconnect technologies, the system is designed to handle demanding AI workloads, including training trillion-parameter language models and real-time inference.
Key features include:
- Exceptional AI performance: The system includes 72 NVIDIA Blackwell GPUs, providing up to 30 times faster real-time inference performance.
- Full-rack solution: It integrates GPUs, CPUs, and NVLink Switches for fast, low-latency communication, which accelerates training and inference of large language models (LLMs).
- High-speed interconnect: NVIDIA’s fifth-generation NVLink technology ensures high throughput and smooth communication between components.
- Advanced cooling: The system supports both liquid-to-air and liquid-to-liquid cooling solutions, enhancing energy efficiency and heat management.
The combination of advanced cooling and cutting-edge hardware makes the ASUS AI POD highly efficient, even for power-hungry AI applications.
Transformative AI infrastructure
ASUS has developed advanced liquid-cooling systems to improve energy efficiency and minimise power usage in data centres. These include components like cold plates for CPUs and GPUs, coolant-distribution units, and cooling towers. The goal is to reduce overall power consumption while optimising power usage effectiveness (PUE).
To further accelerate AI adoption, ASUS offers a full-service model that covers design, construction, implementation, testing, and deployment. This holistic approach allows clients to seamlessly integrate AI solutions into their operations. ASUS’s tailored infrastructure also works hand-in-hand with cloud application services, ensuring businesses can deploy AI projects quickly and effectively.
The ASUS AI POD with NVIDIA GB200 NVL72 represents a key step forward in AI infrastructure, giving enterprises the tools they need to succeed in a rapidly evolving technological landscape.
