The primary investment thesis is to buy the "picks and shovels" of the AI revolution, focusing on critical infrastructure like power and hardware ahead of a predicted 2027 boom. Consider a foundational position in Intel (INTC), a large-cap hardware manufacturer with strong government backing. A high-conviction bet is on the "NeoCloud" theme, where specialized providers like CoreWeave rent out massive AI compute power. For a more sophisticated trade, consider buying Core Scientific (CORZ), a key supplier to CoreWeave, as a leveraged play on the entire AI data center ecosystem. Lastly, view Bitcoin miners like IRON and Riot as undervalued infrastructure assets that can pivot from mining to providing high-demand AI compute services.

AI Infrastructure & "Picks and Shovels"

The core investment thesis discussed is not about picking which AI model will win, but rather investing in the essential infrastructure—the "picks and shovels"—required to build advanced AI. This strategy, employed by investor Leopold Aschenbrenner, focuses on identifying and investing in the key bottlenecks of the AI revolution.

• Investment Philosophy: The strategy is to find assets that are mispriced because the market has not yet grasped the sheer scale of the coming AI boom. This involves looking for "forced winners" based on a future where AI development requires trillions in spending.
• Key Bottlenecks Identified:
  • Power: The immense energy required to run AI data centers. This leads to investments in utilities, power generators, and Bitcoin miners (as a proxy for energy infrastructure).
  • Compute Hardware: The physical chips and components needed for AI. This includes chip makers like Intel and their competitors.
  • Secure, Well-Governed Labs: The elite research labs building the foundational models, such as Anthropic and Google.

Takeaways

• The primary insight is to invest in the supply chain of AI rather than the end products.
• This involves identifying companies that own critical, hard-to-replicate infrastructure like data centers, power contracts, and specialized hardware manufacturing capabilities.
• The main risk factor highlighted is that this entire thesis is based on a specific timeline, with AGI (Artificial General Intelligence) predicted around 2027. If this timeline is significantly delayed, these investments may be too early.

Intel (INTC)

Intel was presented as a major, successful contrarian bet made by Leopold's fund. The investment was made before it was obvious that the company would play a critical role in the US's AI strategy.

• The Trade: The fund took a massive position in Intel call options when the company was widely considered a "dying competitor to NVIDIA."
• The Catalyst: The bet paid off spectacularly when the US government took a 10% stake in Intel, causing the stock to skyrocket.
• The Result: This single trade was a primary driver in growing the fund from $1 billion to $1.5 billion in under three months.
• Portfolio Role: Intel is now considered the fund's "steady state investment"—a large, foundational position in the AI hardware space, backed by government support.

Takeaways

• Contrarian bets on legacy tech companies can yield massive returns if you correctly predict a strategic shift.
• Government intervention and industrial policy (like the CHIPS Act) can serve as powerful catalysts for undervalued hardware and manufacturing stocks.
• Intel is viewed as a safer, large-cap way to get exposure to the AI hardware build-out, especially given its strategic importance to the US government.

NeoClouds (Investment Theme)

"NeoClouds" is a term used to describe a new category of companies that provide hyper-specialized AI compute services, acting as a crucial layer between the big tech companies and the physical hardware.

• Business Model: Described as "AWS, but supercharged and specialized towards AI compute." These companies build and manage the data centers, allowing tech giants like Microsoft, Meta, and Google to rent massive amounts of AI compute power without having to build all the infrastructure themselves.
• The Bet: The thesis is that as demand for AI training explodes, big tech will increasingly rely on these specialized providers, signing multi-billion dollar contracts with them. This was once seen as a "crazy" bet but is now becoming an obvious trend.

Takeaways

• NeoClouds represent a direct investment in the surging demand for AI compute power.
• As big tech companies commit hundreds of billions of dollars to AI, these specialized providers are direct beneficiaries.
• Key companies to watch in this space are CoreWeave and Iron (IRON).

CoreWeave & Core Scientific

This is presented as Leopold's latest and most sophisticated trade, representing a nearly $1 billion combined bet on the neocloud ecosystem. It involves buying both the leading neocloud provider and its struggling supplier to control the outcome.

• CoreWeave:
  • A leading neocloud provider that recently signed a $19 billion partnership with Microsoft and has deals with OpenAI.
  • Leopold's fund has taken a new $563 million position in CoreWeave, making it the second-largest holding.
• Core Scientific:
  • A former "risky Bitcoin mining company" that is a key third-party infrastructure provider for CoreWeave.
  • The company recently failed to deliver promised power capacity to CoreWeave, causing a 10% dip in CoreWeave's stock, highlighting its supply chain risk.
• The "Levered" Trade:
  • In response, Leopold's fund took a nearly $400 million position in Core Scientific, increasing its stake to 9% of the company.
  • This is seen as a "levered long" on CoreWeave. By owning a significant piece of its key supplier, the fund can likely gain a board seat and influence Core Scientific to ensure it meets its obligations, thereby de-risking CoreWeave's growth.

Takeaways

• This is a high-conviction, activist bet on the entire neocloud vertical. The strategy is to not just own the winner (CoreWeave) but to also buy and fix its weakest link (Core Scientific).
• It demonstrates the hidden value in Bitcoin mining companies, which possess critical infrastructure (data centers, power contracts) that can be repurposed for the AI boom.
• This dual investment is a bet that CoreWeave will be the leading neocloud provider and that by taking an active role, the fund can ensure its success.

Bitcoin Miners (as an AI Infrastructure Play)

The podcast highlights an unconventional investment thesis: buying Bitcoin mining companies not for their exposure to Bitcoin (BTC), but for their valuable physical infrastructure.

• The Insight: Bitcoin miners are "improperly applied" assets. The market values them based on volatile crypto prices, but their real, durable value lies in their massive, power-hungry data centers, land contracts, and grid interconnections.
• The Pivot: As AI creates a "power scramble," these companies are perfectly positioned to pivot from mining Bitcoin to providing high-demand AI compute services.
• Portfolio Holdings: Besides Core Scientific, other companies mentioned in the portfolio that fit this theme include IRON, CIFR, Hutt, and Riot.

Takeaways

• Look beyond a company's primary business to identify undervalued, hard assets on its balance sheet.
• Bitcoin miners represent a potential arbitrage opportunity: buying crypto-priced assets that have the potential to be re-rated as high-demand AI infrastructure providers.
• This is a "picks and shovels" play on the single biggest constraint for AI: energy and power. Companies that have already secured power are at a massive advantage.