The importance of safeguarding in AI

Key Concepts:

• Intermediary Producers: Companies like Nvidia that produce core components for AI.
• Hyperscalers: Large cloud providers and data companies (e.g., Alphabet, Meta) that consume AI components and offer AI services.
• TPUs (Tensor Processing Units): Custom AI accelerators developed by Google (Alphabet) for machine learning workloads.
• Inference (AI): The process of using a trained AI model to make predictions or decisions on new data.
• Model Training (AI): The process of teaching an AI model by feeding it large datasets.
• Sovereign AI: A concept emphasizing control and ownership over one's own AI data and infrastructure, rather than just buying chips.
• Data Sovereignty: The idea that data is subject to the laws and governance structures of the nation in which it is collected or processed.
• Data Privacy: The protection of personal data from unauthorized access and use.
• Cradle-to-Grave Solutions: Comprehensive services covering the entire lifecycle of a product or system, from inception to decommissioning.
• Trough Year: The lowest point in an economic or business cycle.
• Through Cycle Earnings Power: A company's ability to maintain and improve its profitability across different phases of an economic cycle.
• Durable Franchise Business: A company with a strong, lasting competitive advantage (moat) that allows it to generate consistent profits.
• Competitive Advantage (Moat): A structural feature of a business that protects it from competition and allows it to earn above-average profits.

Tech Market Dynamics and AI Competition

Christopher Davis, partner at Hudson Value Partners, discussed the regaining momentum in tech, particularly highlighting a bounce back for Nvidia, which was up in the $180 range. He emphasized a crucial "separation" within big tech stocks, distinguishing between "intermediary producers" like Nvidia and "end hyperscalers" such as Alphabet.

Davis noted that the market is realizing Nvidia's gross margins, which have been in the "80% range," are not sustainable indefinitely due to impending competition. This competition is notably emerging from companies like Alphabet, specifically with their TPUs (Tensor Processing Units). These TPUs are particularly useful for "inference" applications in AI, rather than "model training." Davis views this competition as positive for the broader AI group, suggesting it could lead to lower AI spending for hyperscalers like Meta, making their investment dollars "go farther."

The Imperative of Sovereign AI and Data Protection

Davis stressed that Sovereign AI is one of the most misunderstood terms currently. He clarified that it's not merely about governments purchasing large quantities of chips but fundamentally about "control and ownership over your data." He used an analogy: businesses wouldn't allow competitors free rein to access their sensitive data in filing cabinets without safeguards like NDAs or conflict reviews. Similarly, in the age of AI, investors, companies, and enterprises must prioritize "privacy and data protection within AI."

He highlighted that the terms of use for many AI services are "not exactly favorable to us," but rather to the hyperscalers, as every user query helps "sharpen all those models." Consequently, businesses with sufficient resources will increasingly seek to build their "own private solutions," reducing reliance on public cloud services or external computers for sensitive data. While some private and public companies (e.g., big banks developing private AI versions and chatbots) are already moving in this direction, Davis anticipates significant future development in this area.

Dell Technologies: Enabling Private AI Solutions

Davis identified Dell as a key player in facilitating private AI solutions. Dell's business encompasses enterprise server and storage, alongside its PC division. While AI servers haven't been a high-margin business for Dell due to intense competition in large data centers, Davis sees a significant opportunity. He believes that companies capable of affording it will deploy "Dell servers, Dell towers, Dell storage solutions for their own private AI systems and solutions within their companies." This caters to Dell's traditional "sweet spot" of Fortune 5000-type businesses, offering potentially "a lot more profitable" avenues than the highly competitive, multi-gigawatt data center projects.

Amentum Holdings: A Niche in Nuclear Services

Davis also highlighted Amentum Holdings, a US-based government and defense contracting business specializing in mission-critical solutions. The company, a spin-off, recently reported "fantastic earnings," with its stock up about 18% before a slight pullback. Amentum has a market cap of $7.3 billion, 243 million shares outstanding, and trades at approximately 12 times forward earnings.

The "gem within their business" is their nuclear services business. Amentum provides "cradle-to-grave solutions for operating nuclear power plants" in the US and UK, and also works with the US Navy on submarines and other nuclear-powered vessels. Davis emphasized that nuclear energy is a "hottest theme" for energy security, powering AI, and data centers. He considers Amentum "one of the best companies that people out there simply don't own in a very expensive sector and a hot theme," noting that the market is beginning to recognize its value, evidenced by price target raises and increased investor involvement. The company boasts a substantial "$47 billion order backlog," indicating "plenty of runway" for future growth.

Deere & Co.: A Durable Franchise in Industrials

Despite a recent earnings report that disappointed investors due to its forecast, Davis expressed a bullish outlook on Deere & Co. He grounded this perspective in value investing principles, identifying Deere as a "textbook franchise" with a "strong competitive advantage," a "wonderful dealer network and supply chain," and a commitment to "continually innovating and reinvesting."

Davis noted a market shift towards cyclicals and industrials, positioning Deere favorably for 2026, which management has guided as the "trough year" for both earnings and the agricultural cycle. He praised CEO John May's leadership in improving the "through cycle earnings power" of the enterprise. Notably, Deere's margins are currently "4 and a half percent higher" than at a similar point in the cycle in 2016, indicating improved efficiency and less inventory.

Deere's future prospects are bolstered by a strong product lineup, upcoming investor day on December 8th, and new products at CES. Crucially, Deere is a "real implementer of AI data," applying industrial technology to farming to "improve efficiency, increase crop yields." Davis recommended Deere as an investment, particularly within the industrial sector, for those interested in "food security" or making "smart investments towards the bottom of the cycle."

Conclusion

Christopher Davis provided a detailed overview of current market dynamics, emphasizing the evolving competitive landscape in AI, the critical importance of data sovereignty for businesses, and specific investment opportunities. He highlighted the shift from Nvidia's dominance to a more competitive AI hardware market with Alphabet's TPUs, the growing need for private AI solutions benefiting companies like Dell, and the undervalued potential of Amentum Holdings in nuclear services. Finally, he presented a contrarian bullish view on Deere & Co., citing its strong franchise, improved operational efficiency, and integration of AI in agriculture, positioning it as a compelling investment at the bottom of its cycle. The overarching theme is a move towards greater control over data and a recognition of durable, innovative businesses in both tech and traditional industrial sectors.