Wells Fargo notes of investor meeting with NVDA CFO
Our Call: Overweight for $370 target
Yesterday, (1/5) we had the opportunity to participate in an investor meeting with NVDA's CFO, Colette Kress, in conjunction with this week's CES 2022. Following this week's GeForce RTX (3050 desktop GPU; 3070 Ti & 3080 Ti for mobile), GFN, Omniverse, and Auto (Hyperon 8 & DRIVE) announcements (see note), we think the most incremental takeaways from our discussion included:
1) Supply chain constraints to persist at least thru 1HC2022; seasonality not relevant,
2) Reiterated $8B auto pipeline (thru 2027) with more to come; rev. ramp commencing in 2HC2022,
3) Omniverse now has >100,000 creator downloads; rev. oppy beginning in 2022.
4) Expect materializing BlueField DPU ramp to be a 2H2022 story; underappreciated incremental data center growth driver.
Software / Subscription Story.
NVIDIA's ability to derive materializing revenue contributions from an expanding portfolio of software / subscription offerings remains an important piece of the company's forward growth thesis. Our discussions supported our belief that we should expect to see some rev. ramp materializing in 2022. This includes offerings such as the company's Enterprise AI Software suite, AI Launchpad, GeForce Now (GFN), and Omniverse platform (40M creators x $1,000/yr. license a fair est. of opportunity). We continue to believe that investors should consider the ramp of these opportunities showing up on the company's balance sheet / off-balance-sheet disclosures—i.e., deferred revenue growth and Remaining Performance Obligations (RPO). While still small today—deferred revenue exited F3Q22 at $489M; RPO (includes deferred) totaled $620M—we believe a ramp in these figures can be considered a leading indicator of NVDA's software / subscription opportunities.
Automotive Opportunity — Inflection in Late-2022 / 2023?
While NVDA has noted that it is still early in terms of the company's opportunities in the automotive sector (currently 2%-3% of total rev.), our discussions continue to leave us to believe that the company's $8B pipeline (thru 2027) will begin to show an inflection as we move into the late-2022 / 2023 time frame. Ms. Kress also noted that there is likely to be further updates to this pipeline amid ongoing design win momentum looking forward. This week NVIDIA highlighted the rollout of the DRIVE Hyperion 8 compute architecture / platform for full self-driving systems—available for 2024 vehicle models. This platform integrates two NVDA DRIVE Orin SoCs with support for 12 cameras, nine radars, 12 ultrasonics, and a front-facing Lidar sensor. The company also demonstrated the integration of DRIVE Chauffeur and DRIVE Concierge (first introduced in Nov '21 - see blog) — Chauffeur for autonomous driving and Concierge for AI-enabled assistance (Omniverse Avatars).
In addition to a flagship partnership with Mercedes, NVDA continues to highlight a strong positioning with Chinese EV vendors—e.g., NIO, Xpeng, Li Auto, IM Motors (SAIC), and R Auto (SAIC). As a reference, below we include some charts illustrating the inflection we are seeing in Chinese EV shipments over the past few months / quarters—total pure electric vehicle production stood at ~14% of total China vehicle production in Nov '21, up from sub-6% a year ago and reflecting YTD pure EV production at +180% y/y. Management noted that it will begin to monetize its ~$8B automotive pipeline in 2HC2022 with autonomous solutions starting to become a larger percentage of auto revenue.
GeForce RTX Upgrade Cycle; Supply Dynamics.
This week NVDA noted that GeForce GTX (pre-Ray Tracing) GPUs still accounted for ~75% of the company's total GeForce GPU installed base. As illustrated below, Steam Hardware Survey data showed that GeForce RTX GPUs accounted for approx. one-third of the surveyed GeForce installed base exiting 2021; Ampere-based GeForce RTX 30xx-series GPUs accounted for ~15% of the installed base. Although this week's announcements suggest the company is nearing the completion of the build-out of their Ampere portfolio rollout (commenced in Sept '20), we believe this suggests a continued strong installed base upgrade opportunity ahead—especially as we continue to see the rollout of new gaming titles w/ Ray Tracing support.
From a supply perspective, NVIDIA noted that it saw a strong holiday season (continues into January) and that overall channel inventory levels are expected to remain very lean. Supply improvements are expected to be a 2HC2022 story; the company noted that its historical seasonality is unlikely to apply through calendar 2022. As a reminder, NVIDIA exited F3Q22 (Oct) with $6.9 billion in inventory purchase commitments of which $2.944B was to be used in F4Q22 and $4.637B used in F2023. This included $1.606B in prepaid supply agreements vs. having no prepaid agreements disclosed exiting prior quarters.
Omniverse Ramp a 2022 Story.
NVIDIA's Omniverse platform strategy spans across the company's product segments—from gaming and content creators to enterprise and automotive. Our discussions highlighted this week's announcements of Omniverse being offered free for individual creators, as well as the role it plays in the automotive segment. NVIDIA noted continued adoption of Omniverse especially at the enterprise level — reminder that the company noted that it had 700 enterprises evaluating Omniverse during its F3Q22 earnings call; today also noting that it has seen 100,000 downloads. We expect to hear more metrics on the adoption / use cases as we move through 2022. NVIDIA continues to see large companies such as BMW adopt Omniverse tools that allow collaboration in virtual design, while also pointing out Nike, Gucci, and others. Management highlighted the ability to move physical testing (ex. automotive) to the virtual world to improve testing speed / efficiency. While it is still hard to quantify the full breadth of revenue opportunity associated with Omniverse - i.e., pull-thru of hardware to software licensing - the company did note that it is fair to use a TAM est. based on ~40 million content creators x $1,000/yr. license.
BlueField-2/35 Data Processing Units (DPUs):
NVIDIA reiterated that its BlueField DPUs expected to be a major driver of its datacenter story moving forward with the potential for DPU insertion to every server especially in multi-tenant environments. NVIDIA expects a BlueField ramp to begin through calendar 2022; we continue to believe VMware's full release of Project Monterey and cloud adoption will be key drivers. We continue to believe that DPUs will become an increasingly important piece of the datacenter as the move toward disaggregation accelerates.
Price Target Basis and Risk
Price Target for NVDA: $370.00 from NC
Our $370 price target is based on a ~68x P/E multiple on our C2023 estimates. We continue to believe a premium multiple for NVIDIA is warranted given what we believe is a multi-year strong competitive positioning for data center growth driven by cloud and AI, gaming, next-generation autonomous vehicle applications, and an expanding ecosystem of products/applications (e.g., Omniverse).
Risk for NVDA
Risks include: (1) increased competition in the PC gaming, cloud data center markets, and high-performance computing, (2) the continued development of new markets, including artificial intelligence/machine learning and autonomous driving, (3) delays in product introduction due to use of third party process technology, component availability, etc.
Investment Thesis
NVDA
Our Overweight rating is based on our positive stance on NVIDIA’s competitive positioning in gaming GPUs and expanding growth opportunities in data center (GPUs and the potential leverage of acquiring Arm), HPC, and emerging / expanding AI opportunities (autonomous vehicles, healthcare, robotics, etc.). We see sustainability in NVIDIA's CUDA software platform differentiation, as well as the company's continually deepening system software capabilities. We see NVIDIA as one of the most attractive secular growth stories in large-cap semis. |
