Investors should prioritize NVIDIA (NVDA) as the essential infrastructure play, as its specialized chips remain the primary beneficiary of a "tornado of demand" that continues to exceed market earnings models. For foundational AI exposure, Google (GOOGL) is a high-conviction pick due to its technical superiority in document processing and its guaranteed spot in the emerging three-player model oligopoly. Look for "de-commoditized" hardware winners like Celestica (CLS) for its 60% share in AI Ethernet switches and Corning (GLW) for the massive fiber optic demand required to connect data centers. While Anthropic is the top private conviction for its $500 billion coding market opportunity, public investors should pivot away from traditional SaaS stocks like Salesforce (CRM), which face severe budget cannibalization from AI. We are currently at only 0.1% penetration of AI adoption, suggesting significant upside remains for companies where average selling prices are rising due to increased manufacturing complexity.

Anthropic

• Anthropic is currently the highest conviction position for WhaleRock Capital Management. • Originally viewed as a "dark horse" in the foundational model race, it has emerged as a leader alongside OpenAI and Google (Gemini). • The company has pivoted successfully to focus purely on the enterprise market, whereas OpenAI has a stronger consumer lean. • Coding is identified as the "true unlock" for the company. Their tool, Claude Code, has moved toward being "agentic," meaning it can perform complex tasks autonomously rather than just completing lines of text.

Takeaways

• Market Opportunity: The coding market alone is estimated to be a half-trillion dollar opportunity, with developers potentially spending $20,000–$30,000 annually on tokens to increase productivity. • Competitive Moat: Anthropic is building a "harness" (SDKs, orchestration layers, and tools) around its API, creating a sticky ecosystem similar to how AWS locked in cloud customers. • Investment Timing: WhaleRock invested when the valuation was around $18 billion, noting that the company's revenue growth (from $100 million to $1 billion and heading toward $9 billion) is unprecedented in software history.

NVIDIA (NVDA)

• Mentioned as a core play in the "infrastructure layer" of the AI S-curve. • The guest argues that even at high valuations, the stock has historically been "cheap" on a forward-looking basis because the market fails to model exponential earnings growth. • NVIDIA benefits from the "de-commoditization" of hardware; because AI workloads grow 10x annually, the demand for their specialized chips remains insatiable.

Takeaways

• Infrastructure First: In a new technology cycle, the "chips at the bottom" get the demand first and offer the most clarity on who the winners are. • Pricing Power: Because AI servers are mission-critical (costing $200k–$300k vs. $5k for old servers), NVIDIA and its partners have moved from being commodity suppliers to being essential, high-margin partners.

Google (GOOGL)

• One of WhaleRock’s largest positions. • Despite concerns about AI disrupting search, Google is viewed as a top-three player in the foundational model "oligopoly." • Specifically noted for its technical superiority in ingesting and processing large PDFs and complex documents.

Takeaways

• The Three-Horse Race: The AI model market is evolving like the cloud market (AWS, Azure, GCP). Google is expected to be one of the few companies with the capital and compute to remain at the "leading edge."

The AI Hardware Stack (Celestica, Corning, Elite Materials)

• Celestica (CLS): A contract manufacturer that transitioned from commodity work to high-end AI servers and Ethernet switches. They have a 50-60% share of the cloud Ethernet switch market. • Corning (GLW): Benefits from the massive amount of fiber optic cable needed to connect AI data centers. One Microsoft data center can require enough fiber to circle the earth four times. • Elite Materials: A supplier of "copper clad laminate" for high-layer printed circuit boards (PCBs). AI servers require 40-layer boards compared to 10 layers for standard servers.

Takeaways

• The "L-Curve": While most technologies follow an "S-curve," AI infrastructure demand is described as an "L-curve"—moving straight up because there is a global shortage of compute power. • De-commoditization: Investors should look for hardware companies where ASPs (Average Selling Prices) are rising because the technology is becoming harder to manufacture.

Enterprise Software (SaaS) Sector

• The guest expresses a bearish sentiment on traditional SaaS companies (e.g., Salesforce). • WhaleRock sold almost all of its software holdings and was "net short" the sector entering 2024.

Takeaways

• Budget Cannibalization: CIOs are shifting budgets away from traditional software seats toward AI tokens (like Anthropic) because the ROI is faster. • Disruption Risk: AI native startups may soon replace legacy ERP and CRM systems. The "Rule of 40" for software is being replaced by a new metric: What % of your sales are AI-driven? Most legacy players are only at 1-2%.

Investment Themes: The S-Curve Framework

• The Framework: Invest based on the S-curve (adoption cycle), competitive advantage (moats), and underappreciated earnings power. • The "Tornado of Demand": This occurs when barriers to adoption (price, complexity) are removed. For AI, the "light switch" went off in 2023/2024. • Current Phase: We are currently at only 10 basis points (0.1%) penetration of knowledge workers truly using AI. The move to 5-15% over the next four years represents a massive investment opportunity.

Takeaways

• Don't Fear "High" Charts: It is okay to miss the first 100% of a move if the "top" of the S-curve is trillions of dollars away. • Focus on the "Recursive Loop": Look for companies where the AI helps the company build better AI (e.g., using Claude to write the code for the next version of Claude).