Recently, NVIDIA's newly launched Blackwell AI chip has encountered overheating issues in servers, raising concerns among customers about their ability to timely activate new data centers. According to The Information, the Blackwell graphics processing units (GPUs) are overheating when connected to server racks designed to accommodate 72 chips. Insiders reveal that NVIDIA's engineering team is actively addressing this issue, and the company's staff have already begun...

Recently, NVIDIA unveiled its brand-new Blackwell platform and showcased its initial performance in the MLPerf Training 4.1 benchmark tests. According to the test results, Blackwell achieved a doubling of performance in certain aspects compared to the previous Hopper platform, garnering widespread attention in the industry. In the MLPerf Training 4.1 benchmark tests, the Blackwell platform excels in LLM (Large Language Models).

Recently, NVIDIA's latest generation Blackwell GPU has sparked unprecedented demand, with CEO Jensen Huang revealing in an investor meeting that the supply for the next 12 months is completely sold out. This phenomenon is similar to the explosive demand for the Hopper GPU a few months ago, indicating a strong market desire for high-performance computing. This surge in sales is primarily driven by traditional clients, including AWS, Google, Meta, Microsoft, Oracle, and CoreWeave.

In the fiercely competitive artificial intelligence hardware market, AMD has introduced its brand new MI325x AI chip, aiming to challenge Nvidia's latest Blackwell series chips for dominance. This new chip represents another significant move by AMD in the field of artificial intelligence computing, striving to expand its share in this lucrative market. The MI325x AI chip is designed to directly compete with Nvidia's Blackwell GPUs, which are widely regarded as the leaders in AI.

Recently, Foxconn (Hon Hai) and Nvidia announced an exciting collaboration plan to build a massive supercomputer. This computer will use Nvidia's latest Blackwell chip architecture and is named the 'Hon Hai Kaohsiung Supercomputing Center'. This project was officially unveiled during the recently concluded Foxconn Technology Day event. Image credits: The image is generated by AI, and the image is licensed by service provider Midjourney. The scale of this supercomputer is enormous.

Recently, NVIDIA's CEO Jensen Huang revealed in an interview that OpenAI's latest o1 large language model (LLM) will be based on NVIDIA's Blackwell architecture, which has achieved an astounding improvement in inference capability. As the market gradually emerges from the hype of generative AI, developing LLM models capable of generating 'inference-focused' responses has become increasingly important, marking a significant step toward achieving general artificial intelligence.

Oracle has announced the launch of the world's first Zettascale cloud computing cluster. This cluster is equipped with up to 131,072 NVIDIA Blackwell GPUs, capable of delivering a peak performance of up to 2.4 ZettaFLOPS. This powerful AI infrastructure enables enterprises to process large-scale AI workloads with greater flexibility and sovereignty. Oracle Cloud Infrastructure (OCI) integrates with NVIDIA's advanced GPU architecture.

Nvidia CEO Jensen Huang emphasized in an interview with the Financial Times that despite encountering issues in the production of the next-generation AI processors, the company remains committed to its plan of launching a new flagship product every year. Although the release of the Blackwell chip has been delayed, Nvidia engineers are accelerating the development of the next two generations of chips. The production issues are mainly related to adjustments in the 'mask' process during chip design, and these changes have been completed without affecting product functionality. Nvidia is leveraging AI to enhance data processing and product design efficiency in response to investors' concerns about growth rates and profit margins.

This article outlines the latest developments in the AI inference chip field, particularly the results of the MLPerf Inference v4.1 competition, which gathered the latest chips from companies such as AMD, Google, UntetherAI, Cerebras, FuriosaAI, and Nvidia. Nvidia's H200 and GH200 chips maintain a lead in the 'Data Center Closed' category, but chips from other companies excelled in specific tasks, especially in terms of energy efficiency. AMD's Instinct Accelerator.

Nvidia has encountered engineering challenges during the launch of its next-generation AI chip series Blackwell, leading to delays in the product release schedule and affecting market expectations. The main issue arises from the need to redesign an AI accelerator for compatibility with the Hopper H100 chip, as well as a processor combined with a graphics card that cannot be mass-produced due to technical problems. These challenges highlight the complexity of the technology innovation process. As a result, Nvidia's stock price dropped by 6.4%, while competitor AMD's stock price rose by 1.8%. Despite the delays, analysts believe that this...

NVIDIA plans to advance the release date of the Blackwell B100 GPU to Q2 2024. NVIDIA has selected SK Hynix's HBM3e memory for the Blackwell B100 GPU. The early launch of the Blackwell B100 is aimed at meeting the rapidly growing demand for AI. NVIDIA received samples of SK Hynix's HBM3e memory in August of this year. The Blackwell B100 will help NVIDIA maintain its leading position in the AI market.