[REPLAY] Gavin McCracken: A Journey To 2,000x Returns

Gavin McCracken describes finding Andean Precious Metals at 70 cents, initially exiting at 75 cents to pursue other opportunities like Velour Energy (which returned 10x). After a negative 2024 due to uranium exposure and a Liberation Day margin call that wiped 40% of his account, he reinvested heavily into Andean with 3.33x leverage, riding it to approximately 100x returns in 2025 when the company released a buyback filing. He emphasizes that success came from understanding the commodity macro thesis (gold/silver debasement), identifying low-cost operators with tight floats and strong insider ownership, and deploying capital incrementally while monitoring technical support levels and chart patterns.

McCracken's investment process relies heavily on commodity cycle identification rather than fundamental analysis alone. He explains using the concept of "self-avoiding random walks" from computer science—the idea that once a strategy or asset class has performed well for an extended period, one should recognize the cycle is mature and rotate to new opportunities. He successfully exited oil in 2020 (having made 100-150x over three years) into uranium, then out of uranium into gold. He values companies based on enterprise value-to-free-cash-flow ratios and break-even costs per commodity unit, specifically targeting operators that remain profitable even in severe commodity downturns.

On management and capital allocation, McCracken emphasizes that insider buying is a critical validation signal—if executives won't personally invest their own capital into the thesis they're pitching, it raises red flags. He cites PetroTal as a negative example where despite ludicrous free cash flow generation, management chose dividends over buybacks, resulting in range-bound share price performance. Conversely, he highlights Andean's CEO ownership (52%) and aggressive insider buying at higher prices as validation of his position.

McCracken discusses his current macro thesis centered on a global fuel crisis driven by geopolitical tensions between Iran and Saudi Arabia. He argues that critical pipeline infrastructure (specifically the Yanbu pipeline handling 7 million barrels daily) represents existential risk to global oil supply. This supports his overweight position in Suncor—a fully integrated Canadian producer with refining assets, low break-even costs ($42/barrel, targeting $38), and limited exposure to windfall taxes or export controls (due to Alberta's potential separatist movement and U.S. Constitutional constraints on Trump-era export restrictions). He's deployed approximately 65% net worth in Suncor across common equity (30% margin requirement) and June 2025 call options, providing 120% of net worth in underlying exposure.

On artificial intelligence, McCracken presents a contrarian bear thesis despite his AI PhD background. He argues that Large Language Models trained via next-token prediction on existing text cannot generate truly novel ideas—only combinatorial recombinations of existing concepts. He predicts the current GPU supercycle and AI enthusiasm represents bubble-peak conditions similar to 2007 finance, driven by venture capitalists and company executives who overstate progress (six-month AGI claims) to raise capital. Model compression and efficiency improvements will likely reduce compute demand rather than increase it. He uses AI tools only for code translation (English-to-Python) and verification, as he doesn't trust LLM outputs for numerical calculations or fundamental analysis.

McCracken contrasts his concentrated, rotating-commodity approach with more diversified strategies like Calvin's (which holds 25+ positions sized 1-2.5% of NAV). He acknowledges Calvin is the "better investor" despite lower returns, but explains his concentrated approach suits his personal goal of launching a technology research company without excessive venture capital control. He aims to rotate out of positions in the "late third inning" rather than hold through fourth inning to maximize returns, as extended hold periods would conflict with his entrepreneurial timeline.

On broader market dynamics, he criticizes "metric hijacking" across society—the tendency for institutions (universities, governments, corporations) to optimize for easily-measurable metrics (GPA, h-index, debt-to-GDP) rather than actual value creation. This incentivizes cookie-cutter thinking over creative breakthroughs. He expresses interest in dining with John von Neumann to understand how historical polymaths approached problem-solving before metrics-based optimization dominated academia and finance.