Reinventing large-scale AI computing through hyperheterogeneous innovation, accounting for nearly 80% of the GPU market. The data center business grew rapidly and became the growth engine of the world's AI.
Key points of investment
The first coverage gave an “increase in holdings” rating. As an industry leader, Nvidia is unrelenting. Considering its outstanding performance in 1QFY2024 revenue, including a record of 4.28 billion US dollars in data center revenue, and Nvidia's own optimistic expectations that 2QFY2024's revenue outlook will reach 11.0 billion US dollars, we expect the company's FY2024E/FY2025E/FY2026E revenue to be 400.0/516.26/62.0 billion US dollars, an increase of 48.29%/29.07%/20.09%, FY2024E/FY2025E /FY2026E's adjusted net profit was US$15196/223.07/28.579 billion, respectively, up 247.89%/46.80%/28.12% from the same period.
Nvidia uses hyperheterogeneous innovation to build a systematic competitive advantage for large-scale AI computing. Nvidia has optimized its full-stack system for large-scale parallel computing in the AI era. The NVLink performance of Nvidia's chip interconnection communication technology was rapidly iterated, and the combination of GPU+Bluefield DPU+Grace CPU pioneered high-speed communication interconnection between chip systems. At the same time, CUDA acts as a general platform, introducing Nvidia software services and the entire ecosystem. We believe that the implementation of chip and system coupling enabled Nvidia to truly achieve hyperheterogeneous innovation.
The GH200 superchip is the culmination of Nvidia's products and technology. We believe that the GH200 incorporates the most advanced Grace Hopper architecture and uses a fourth-generation Tensor Core to improve computing performance and optimize models. NVLink achieves high-speed transmission. In particular, NVLink changed the complex transmission process of traditional PCIe, met the requirements for seamless high-speed communication between each GPU, and built a high-speed interconnection system between chips, which will further form a competitive barrier for Nvidia.
As a leading company, Nvidia will enjoy a large percentage of the dividends brought about by the high overall demand in the AI chip industry. IDTechEx predicts that the global AI chip market will grow to 257.6 billion US dollars in 2033; JPR predicts that the compound growth rate of global GPU sales will remain at 6.3% in 2022-2026. As a well-known leading company in the industry, Nvidia's product ecosystem has remarkable scarcity. It will fully benefit in the field of computing power and enjoy the increase in customer demand brought about by the market explosion.
Risk warning: AI application development falls short of expectations; the company's R&D progress falls short of expectations; geopolitical conflicts affect product sales.