The most recent NViDiA earnings call gave us the latest developments from Nvidia and what they mean for the broader AI landscape. With the industry rapidly maturing, we're seeing new opportunities emerge, particularly in the disaggregation of AI infrastructure. As we navigate these changes, it's crucial to stay ahead of the curve and understand how these shifts could impact our investment strategies. I’ll be writing a longer form version dissecting these trends and where I see capital flowing so be on the look out!
The shift in the industry's direction is evident: as the cost of ownership decreases, barriers to entry will lower, inviting new end users into the accelerated computing market. This won't be limited to LLM training; it will extend into new geographies and applications. The miniaturization of what Jensen refers to as the "AI Factory" is inherently bullish for our positions (MT, Ai, Connect, etc.). A smaller, more specialized AI data center means more opportunities for value-added products and services, creating numerous avenues for value extraction. The more complex the process, the more chances there are to capitalize on it.
However, we view this as a relatively mundane quarter for Nvidia, with gross margin headwinds likely persisting until Blackwell production ramps up. This situation might lead to an accelerated pace of Hopper deliveries, benefiting the topline and serving as a potential blessing in disguise as Nvidia works through its remaining inventory. Demand for Blackwell remains strong, which should help navigate the current challenges.
The maturation of the AI industry is significant, marked by the evolution from large, brute-force data centers to a smarter, more agile network of smaller ones. This shift represents a crucial phase in the adoption of Generative AI and aligns with my multi-phase approach to AI investing—a framework we believe will remain relevant for years to come. As the industry continues to evolve, unexpected situations will arise, but we see these as opportunities. Consider the sovereign AI business, for example. Nvidia's customers, who once joked about printing money due to robust free cash flow, are now literally printing money because they are sovereign nations.
As I have emphasized, the bottleneck isn't in the data center. If it were, the solution would simply be to build the biggest one possible. Instead, the focus has shifted to making data centers smarter and more agile, reflecting a significant maturation of the industry. This evolution supports my investment strategy and reinforces our confidence in the framework we've established for navigating AI's continued growth.
Thanks for reading,
-BSR
Disclaimer:
Remember, this is not financial advice. You are fully responsible for your own decisions!
The information provided on this website/Substack is for informational purposes only. While it is believed to be reliable, Blackshore Research does not guarantee its completeness or accuracy. This content is not an offer or solicitation to buy any financial instrument. The information and terms on these pages can change without notice. Unauthorized use of Blackshore Research’s websites and systems is strictly prohibited. This includes data scraping, unauthorized access, password misuse, or improper use of posted information. Your eligibility for certain services will be determined by Blackshore Research and its affiliates. Investment services are not FDIC-insured and carry investment risks, including potential loss of principal. The proprietary information from Blackshore Research or third-party providers is for individual use only and must not be redistributed or transferred without prior written consent from Blackshore Research.