Guotai Junan: NVIDIA's (NVDA.US) performance exceeded expectations again, accelerating computation and GenAI have reached a critical breakthrough point.

Guotai Junan released a research report stating that Nvidia's FY4Q24 revenue was $22.103 billion, YoY +265%, QoQ +22%, with data center revenue reaching $18.404 billion, YoY +409%, QoQ +27%, and a Non-GAAP gross margin of 76.7%. Total revenue for the fiscal year reached $60.9 billion, YoY +126%. Revenue guidance for FY1Q25 is $24 billion (2% fluctuation range), with a Non-GAAP gross margin range of 76.5%-77.5%. Overall revenue guidance is for continued high-speed growth. Nvidia's performance exceeded expectations once again, with hardware and software complementing each other, and accelerated computing and generative artificial intelligence reaching a critical point of outbreak. Key points from Guotai Junan include: FY4Q24 performance and FY1Q25 guidance exceeded expectations once again, with continued explosive growth in AI revenue such as data centers, where inference demand accounts for 40% of revenue. Nvidia's FY4Q24 revenue was $22.103 billion, YoY +265%, QoQ +22%, with data center revenue reaching $18.404 billion, YoY +409%, QoQ +27%, and a Non-GAAP gross margin of 76.7%. Total revenue for the fiscal year reached $60.9 billion, YoY +126%. Revenue guidance for FY1Q25 is $24 billion (2% fluctuation range), with a Non-GAAP gross margin range of 76.5%-77.5%, and overall revenue guidance for continued high-speed growth. It is worth mentioning that 40% of Nvidia's revenue comes from inference, mainly due to the rapid application deployment and the rapid transition of recommendation systems from CPUs to GPUs. Nearly 80 global automobile OEMs, Siasun Robot & Automation, Tier1 practitioners are using Nvidia's AI infrastructure to train models. Accelerated computing and the rapid promotion of GenAI are expected to drive Nvidia's performance to continue growing after 25 years. Its growth mainly comes from: General computing is gradually transforming into accelerated computing, significantly improving computing efficiency; Generative AI is driving new software applications, and AI Factory is generating a series of new world-changing AI applications, including ChatGPT, Midjourney, Sora, etc. In addition, the company is diversifying its layout with various GPU-specific CSPs. In various countries, including France, Japan, etc., are establishing their own sovereign AI infrastructure, promoting vertical industry applications such as healthcare, financial services, AI developers, autonomous driving, Siasun Robot & Automation, and more. Supply is gradually improving, with H200 starting to take over, and the new product code B100 expected to accelerate its launch, further boosting the efficiency of training large models. The overall supply chain, including wafers, packaging, memory, etc., is gradually improving, with the overall delivery cycle shortening. It is expected that CoWoS capacity will no longer be a core limiting factor for AI development. The company's new product code H200 is continuing to be promoted, with delivery expected to start in Q2, with performance twice that of H100. Additionally, the market's core focus on B100 is also expected to accelerate its launch, greatly improving training efficiency. Risk warning. Recurring tensions in US-China technology; product progress falling short of expectations.