EUR 850.00 For Business Accounts Only

Morningstar | Another Nvidia Analyst Day, Another Set of Gaudy Addressable Market Estimates; Maintain $120 FVE

On March 19, Nvidia hosted its analyst day at which management provided updates for the firm’s efforts in major segments such as gaming, data center, and automotive. In our view, the common theme of the presentations was that Nvidia is more than just a chip provider by offering a platform that layers software and support on top of its hardware offerings. While we applaud this focus on software and support, we maintain our view that major customers ranging from cloud vendors to automakers pursuing self-driving vehicles won’t put all their eggs in the GPU basket. Management once again outlined ostentatious total addressable markets for its data center and automotive opportunities at $50 billion (2023) and $30 billion (2025), respectively. We concede Nvidia’s first mover advantage in artificial intelligence and key applications will allow it to generate meaningful growth going forward (mid-teens growth through fiscal 2024 after a flat fiscal 2020). However, we continue to view shares of narrow-moat, positive-trend Nvidia as overvalued relative to our unchanged fair value estimate of $120 due to competitive forces that appear disregarded by the market narrative.

Following four explosive growth years (38% on average), gaming revenue growth decelerated to 13% in fiscal 2019 and is poised to be lower in fiscal 2020 due to excess channel inventories related to the fallout of cryptocurrency-mining GPU demand and weaker demand in China. Management noted the channel inventory is on track to clear during the first quarter, which bodes well for the rest of fiscal 2020 (though we don’t expect a major rebound). Coupled with a reinvigorated AMD, we don’t think Nvidia will replicate high-double-digit growth rates for its gaming business in our explicit forecast period.

For Nvidia’s GeForce PC gaming GPUs, the firm has enjoyed five-year CAGRs of 14%-ASP growth and 14%-unit growth, while fast-growing laptop gaming GPUs have seen 21%-ASP and 32%-unit growth. We continue to expect slower adoption of Nvidia’s latest Turing GPUs as the majority of games are unable to take advantage of ray-tracing that allows for realistic image rendering (the current GeForce installed base has 50% prior-generation Pascal GPUs and 48% GPUs older than Pascal).

Nvidia’s data center group remains a focal point for the firm’s growth engine, up 52% to $2.9 billion in fiscal 2019. We remain positive on Nvidia’s dominance of the training market for AI, which should spur average data center growth of 31% over the next five years, in our view. Management highlighted some interesting use-cases for Nvidia’s T4 inference GPUs by Linkedin, Snapchat, PayPal, Tencent’s WeChat, Twitter, and Pinterest, and the firm generated a few million in inference GPU revenue during fiscal 2019. That said, most of these workloads feature large batches of data processed concurrently, whereas we anticipate many inference workloads will feature smaller batches of data, many of which won’t occur in cloud environments and may have unique requirements beyond raw processing power and thus would utilize unique chip solutions beyond GPUs.

Automotive sales for Nvidia remain in a transition mode from legacy infotainment to more advanced semi- and fully-autonomous products. Management dissected its 2025 $30 billion TAM by $25 billion for level 2 through level 5 offerings (including robo-taxis), $3 billion for training and development solutions, and $2 billion for validation of autonomous models. We expect burgeoning competition in each of these realms (especially in the first) with Intel-Mobileye as the biggest challenger already boasting over 30 million cars on the road today with Mobileye EyeQ content and a worthy roadmap to match-up against Nvidia’s Drive PX systems.