Nvidia presented its 2023 GTC keynote on Wednesday, showcasing its dominant position in AI/accelerated computing and its position as a one-stop solution provider with a comprehensive portfolio of silicon, software/managed cloud services, hardware systems, and full-stack ecosystems for complex model training/deployment. Judging from the share price, the market is truly excited for Nvidia’s platforms and software solutions potential highlighted in the event!
Market leadership well-positioned for generative AI!
Nvidia's leadership in the field of AI and accelerated computing was highlighted during its recent event, where the company showcased its extensive portfolio of silicon, software, managed cloud services, and hardware systems. The company has established itself as the front-runner in providing off-the-shelf semiconductor solutions and platforms, such as GPU compute acceleration, that facilitate the complex training and deployment of cutting-edge AI-based frameworks like generative AI and large language/transformer models. With its next-gen H100 GPU platform being widely adopted by CSP customers, Nvidia is also launching managed cloud services that allow businesses to create, refine, and operate custom LLMs and generative AI models using their own proprietary data for specific domain-specific tasks.
Economic moat is stronger than ever before
Nvidia is adding value for its customers by enhancing the software solutions stack, delivering value to customers across the entire value chain, and providing various services such as platform-as-a-service and AI infrastructure-as-a-service for customer needs.
Nvidia now offers a range of services, including platform-as-a-service and AI infrastructure-as-a-service, to cater to a variety of customer needs - from specific tasks to custom AI services. Furthermore, Nvidia is continuously expanding its accelerated library offerings (consisting of 300 libraries and 400 AI models) to facilitate rapid market entry and extend their reach to other domains like chip manufacturing and quantum computing. NVIDIA has launched a range of cloud services, including the Nvidia AI Foundation, which helps customers monetize their datasets through training, testing, and deployment. Additionally, services like DGX Cloud and Omniverse Cloud are aiding in the advancement of generative AI and digitalization/simulation.
Nvidia has also expanded its ecosystem into new markets/applications, such as chip manufacturing and quantum computing, by launching new/updated acceleration libraries like CuLitho, RAFT, CV-CUDA, Spark RAPIDS, and cuQuantum. Nvidia 's ecosystem/partnerships now span across various industries such as software (e.g., Adobe), major cloud service providers, medical equipment providers (Medtronic), automotive (e.g., BMW Group, Geely Lotus, and Jaguar Land Rover), and manufacturing (TSMC, Synopsys, and ASML). The company is focused on maintaining its competitive edge by consistently delivering software updates, expanding its cloud offerings, and building strong partnerships to keep its customers engaged.
New platform unveiled!
Four inference platforms were launched by Nvidia, namely L4 for image/video compute, L40 for AI video/cloud compute, H100 NVL for LLM deployment, and Grace Hopper for recommendation models. These platforms are designed to cater to various complex AI applications that cannot be supported by general-purpose CPUs due to their limited compute power. The team is also on track to roll out Grace (ARM CPU), Grace Hopper, and BlueField 3 in the latter half of the year as per their technology roadmap.
Conclusion
The Nvidia GTC event is indeed an eye-opener to me, showing me how Nvidia is indeed a very promising company. For this year, despite the weaker first half of the year, Nvidia is expected to benefit from solid demand in the PC gaming industry, which will offset the secular decline in PC OEM. Furthermore, the data center segment is poised for strong growth, as more hyperscale customers adopt GPU-accelerated deep learning. Nvidia 's potential in the automotive and enterprise segments is also encouraging, though investors should note that the pace of adoption of autonomous driving remains uncertain. Also do note that due to the recent surge in share price, Nvidia is currently trading at a PE multiple of 61x FY24 EPS. Not cheap!
Comments