Investors should consider Eli Lilly (LLY) as a primary play on the AI "application layer," with potential to become the largest U.S. company by the end of the decade through its proprietary AI-driven drug discovery. For infrastructure exposure, Marvell Technology (MRVL) offers significant upside in optical networking and custom chips, with some analysts projecting a long-term path toward a $1 trillion market cap. While Bitcoin (BTC) remains in a technical downtrend, long-term investors can "nibble" at current levels near the 200-week moving average to capture its future role as a settlement layer for AI agents. To hedge against geopolitical volatility and energy-intensive data center growth, look toward defensive energy giants ExxonMobil (XOM) and storage specialist Fluence Energy (FLNC). Conversely, it is wise to reduce exposure to memory stocks like Micron (MU), as the trade has become crowded and AI efficiency gains may eventually reduce demand for high-bandwidth memory.

Based on the analysis of the podcast transcript with Jordi Visser, here are the key investment insights and asset mentions:

Eli Lilly (LLY)

• Visser identifies Eli Lilly as potentially the most important AI company in the world, moving beyond just a pharmaceutical play.
• The company is utilizing Lillipod, a specialized AI model trained on their own proprietary data using over 1,000 NVIDIA Blackwell GPUs, rather than general internet data.
• He views "peptides" as this decade’s version of API keys—a fundamental building block for value creation.

Takeaways

• Bullish Sentiment: Visser predicts Eli Lilly could become the largest company in the U.S. by the end of the decade, surpassing even NVIDIA.
• Actionable Insight: Look at LLY as the leader of the "Application Layer" of AI. While hardware is currently volatile, LLY represents the shift toward ROI-driven AI utilization in healthcare and longevity.

Marvell Technology (MRVL)

• Mentioned as a key player in the "optical" side of AI infrastructure.
• Visser highlights that as data centers move toward custom ASICs (Application-Specific Integrated Circuits), the demand for optical networking and "co-packaged optics" (CPO) will explode.
• He notes that NVIDIA’s Jensen Huang has suggested Marvell could eventually reach a $1 trillion market cap (currently around $250 billion).

Takeaways

• Bullish Sentiment: Visser prefers Marvell over memory stocks like Micron because the optical cycle is in its "very early innings."
• Actionable Insight: Marvell is viewed as a "double or triple" opportunity. It was recently added to the S&P 500, which may provide further institutional support.

Bitcoin (BTC)

• Visser views Bitcoin as the "only high-quality stock that will exist for certainty in a decade."
• He aligns with the thesis that AI agents will need a "crypto rail" (like Bitcoin or major Layer 1s) to transact, settle, and coordinate value at machine speed without human intervention.

Takeaways

• Long-term Bullish / Short-term Bearish: He admits Bitcoin is technically in a "bear market" until it crosses back above its 200-day moving average.
• Actionable Insight: He is "nibbling" and buying small amounts on the way down, specifically noting that Bitcoin is currently at its 200-week moving average, which historically has been a disciplined entry point for long-term investors.

Memory Stocks: Micron (MU), SK Hynix, Samsung

• Visser is notably cautious on the memory sector despite strong consensus from Wall Street (e.g., Goldman Sachs).
• He warns of "Recursive Self-Improvement" (RSI), where AI models may eventually learn to solve their own memory bottlenecks, drastically reducing the amount of High Bandwidth Memory (HBM) needed.

Takeaways

• Bearish/Neutral Sentiment: He exited his Micron positions in the $600-$700 range (referring to market cap/pricing levels in his specific tracking).
• Risk Factor: Retail and institutional positioning is "all-in," creating a crowded trade. He believes the risk-reward is no longer attractive for "doubles or triples."

Energy Sector: ExxonMobil (XOM) & Chevron (CVX)

• Visser is buying these as a defensive hedge against geopolitical instability, specifically the Strait of Hormuz and tensions with Iran.
• He views energy as a "stuck market" that provides safety while AI hardware names experience high volatility.

Takeaways

• Bullish/Defensive Sentiment: These are not expected to be "explosive" growers but are necessary for portfolio balance.
• Actionable Insight: If oil prices migrate higher due to supply chain disruptions, energy stocks will likely outperform the broader "AI rotation."

Fluence Energy (FLNC)

• A play on the massive energy requirements of AI data centers.
• The company recently signed master supply agreements with two major "hyperscalers" (large cloud providers) and has a partnership with Siemens and NVIDIA.

Takeaways

• Bullish Sentiment: Visser believes batteries and energy storage are a necessity for the "Mad Max phase" of data center build-outs where companies are forced to go off-grid for power.

Investment Themes & Sector Rotation

• The "Fireworks" are Over: The period where any AI-related stock would automatically gap up 10-30% is finished. We are moving from "Discovery" to "Digestion."
• Rotation, Not Bubble: Visser argues this is not a 2000-style bubble burst because earnings and profit margins are still hitting all-time highs. Instead, money is rotating out of "Chips" and into Healthcare and Financials.
• Compute vs. Energy: The old economic cycle (Labor vs. Capital) is being replaced by a new one (Compute vs. Energy). Digital "employees" (AI agents) consume tokens and electricity, not wages and houses.
• Application Layer: The next phase of the market will favor companies that can prove Return on Invested Capital (ROIC) from AI, rather than just those selling the "shovels" (chips).