Gavin McCracken is one of the most interesting investors (and people) I’ve ever met. He’s an AI PhD, self-taught physics enthusiast, and I think he’s a polymath … like an actual genius type that just-so-happened to find commodity investing in his spare time.
What makes me think that? Well, over a roughly six-year stretch, Gavin compounded a small account into a 2,000x return by concentrating almost entirely into two commodity equities, layering on margin, and refusing to sell either until the thesis broke.
I don’t have the exact numbers, but he turned ~$8K into $1M+ YOLO’ing into two stocks the Collective knows well: Valeura Energy (VLE) and Andean Precious Metals (APM).
Executive Summary
I don’t need to rehash the VLE and APM theses, but to understand Gavin’s investing philosophy, it’s important to find the common denominators between both positions:
They were cheap, cash-generative, operationally unlevered, and run by insiders who were buying. That is basically his entire investment checklist.
What follows is a synthesis of the five pillars of his process, drawn from our 2+ hours of conversation.
We also end the Podcast Notes with Gavin’s contrarian thoughts on AI and how he uses it in his investment process.
Let’s dive in!
How Gavin Finds Ideas
Gavin relies on a low-tech idea generation model. He runs very few stock screeners. He does not subscribe to bulge-bracket research. Instead, he reads a handful of Substack posts, watches what insiders are doing with their own money, and spends an enormous amount of time in the weeds of balance sheets that nobody else is reading, and privately messages other high-signal investors.
Gavin admitted that he found many of his ideas from friends pitching him stocks on WhatsApp. Patagonia Gold (PGDC), a small position he’s held since $0.20, came straight from Calvin Froedge (@calvinfroedge on X). But the filter after that is entirely his own.
PGDC is a great example. Apparently, Calvin called Gavin at like 2 AM, imploring him to buy the stock. I love ideas like that.
“You find the idea from someone else. But then you have to do the work yourself. Because if you don’t do the work yourself, you won’t have the conviction to hold it when it’s down 40 percent.”
Insider buying as the primary signal
If Gavin had to pick one variable, it would be insider buying. Specifically, insider buying with personal cash, at market prices.
Andean Precious Metals (APM) is a great case study.
Management had been patient, capital-disciplined, and, critically, buying stock in the open market at depressed prices. When Gavin stress-tested the balance sheet at $30 silver and realized the cash on hand was approaching the market cap, the combination of cheap optionality and insider conviction was enough.
Read what isn’t being read
Nobody wants to read a 200-page Thai offshore oil operating report. Nobody wants to research small-cap Bolivian silver producers. Nobody wants to comb through hundreds of Spanish-language documents and Spanish-language YouTube videos to channel-check PDGC’s thesis. Gavin does, and that’s the advantage he (and we as individual investors) have against institutions.
“Retail investors actually have an advantage. You have a small portfolio. You can go anywhere the institutions can’t.”
That’s always been the Macro Ops advantage … go anywhere, invest in anything that we deem to have highly asymmetric risk/reward … to do the work that others won’t to achieve returns others can’t.
Position Sizing & Leverage
Gavin constructs his portfolio the way you’d expect an ex-lover to construct your portfolio after a nasty breakup.
He’s highly concentrated in small-cap commodity producers.
At various points he has run 3.33x gross exposure through margin, with two positions accounting for the overwhelming majority of his book.
Size to your edge, not to a diversification target
He doesn’t think in terms of diversification buckets or benchmark-relative sizing. He thinks in terms of: where is the asymmetry, how wrong can I be, and how much of a drawdown can I stomach before I capitulate. When the answers to all three aligned for VLE in 2022 and APM in 2023-4, he went very, very big.
“The SpongeBob meme where he’s walking with his wallet out to the cash register — that was me with Valeura. I just kept thinking, why don’t I take all my money and move it to Valeura.”
Leverage: used consciously, sized for the margin call
The downside, obviously, is that you can get margin called, which Gavin did on Liberation Day. The tariff-driven volatility event that forced a broad drawdown in commodity equities, forcing him to sell large chunks of VLE.
I get Gavin’s logic, though. If the underlying equity is a cash-rich, debt-free producer with low break-even costs, the margin call is painful but not fatal. If it’s a high-operating-leverage junior explorer, you’re dead on the first 20% move.
Gavin has explicitly said he would not run leverage against anything other than a low-cost, balance-sheet-strong producer.
Against diversification (for him, not necessarily for you)
Gavin’s approach is not replicable for most investors and is not even what he would recommend to a friend.
We discussed Calvin’s portfolio construction, which runs ~25 positions at 1–5% each. He was up ~50% in Q1 2026 with far less stress.
Gavin’s concentration is a function of three personal factors: he has a high pain tolerance, he is extremely confident in his research and convictions when he thinks he’s right, and he wants to build a research business and needs the money from trading to fund it.
“I’m not even trying to maximize my gains as much as I’m just trying to not die when the market’s been going bull for too long.”
Like most things, the right answer is somewhere in the middle. You don’t want 50 different positions at 1-2% of NAV … it takes too much time, and you won’t get the true bang for your commodity buck doing that.
In commodity investing, you’re better off going heavy into your favorite 2-3 plays, and then sitting in them until something changes.
That’s also what I love about our MO trading process. We can get huge notional size in some positions for very little actual risk by using stops and trade management.
Markets as Phase Transitions
Almost everything Gavin said about investing psychology boils down to one competence: the ability to sit on your hands while a position is working against you in the short term.
Gavin developed this skill through a mix of formal training in computer science and statistical physics, and a deliberate philosophy about what markets actually are: a series of phase transitions.
Gavin’s intellectual toolkit comes from a world that most investors don’t touch. He talks fluently about P vs NP, self-avoiding random walks, and phase transitions in statistical systems. The framework he uses most often is the idea that markets spend most of their time in one regime, then rapidly transition to another — and that the moments of transition are where all the returns (and all the losses) actually live.
“A self-avoiding random walk doesn’t go back on itself. Once a market moves into a new regime, you don’t get the old regime back for free. You have to wait for the next phase transition.”
Reflexivity and the refusal to predict
As Gavin explained his Phase Transitions framework, I couldn’t help but think of Soros’s Reflexivity Theory.
Markets reflect back on themselves; prices influence fundamentals; narratives drive flows, which drive prices, which drive narratives. Gavin’s response to this is not to try to predict the feedback loop, but instead to position himself where a feedback loop will work for him: oversold, under-owned, fundamentally profitable, insider-supported businesses in a commodity with structural supply constraints.
The small-portfolio superpower
“You gotta be, I don’t know, dumb enough to think you can do anything. And then you gotta be smart enough to actually do the work.”
The psychological asymmetry he keeps returning to is that retail investors have structural advantages that professionals have traded away: no quarterly marks to defend, no benchmark to track, no LP letter to write when a position is down 40%.
Again, this is a superpower we should extract as much as we can in markets.
Gavin’s Valuation Framework
Gavin has a simple valuation framework. He wants to buy a business for roughly two times the free cash flow it will generate over the next 12 to 18 months, with a balance sheet that can withstand any reasonable drawdown in the underlying commodity. If those two conditions are met, almost everything else is a distraction.
The 2x EV/FCF hurdle
Here’s the important part. Gavin doesn’t “screen” for stocks trading at an EV/FCF of 2x. Instead, he finds companies that are either about to or ramping up production. Then he determines whether he can buy that stock for ~2x (or less) of its run-rate production, assuming a conservative commodity price.
The APM Example
I love and hate my APM thesis because it was so damn simple. I didn’t need to do anything except understand this one paragraph:
“APM trades at a $119M market cap with a Net Current Asset Value (or NCAV) of $107M. The company has $161M in net cash and equivalents for an Enterprise Value of -$42M.
In other words, the market values APM’s two producing mines and its silver mill at -$42M.
Combined, we think those two mines can generate over $30M in annual free cash flow by 2025 using realistic gold/silver price assumptions.
Finally,management owns over 50% of the business and has repurchased stock over the past year ($0.02 above our cost basis).”
That is Gavin’s playbook to a tee:
- Strong balance sheet: $161M net cash ➢ Cheap price: -$42M EV
- Run-rate production cash flow: $30M
- Realistic metals prices: $30 silver and $3,500 gold
- Insider ownership: 50%+ with insider buying
That’s it, nothing more needed.
Low break-even, not high torque
He is explicit that he does not want operational leverage. He doesn’t want the marginal producer whose cost structure is right at the commodity price. He wants the second-quartile cost producer with room underneath … the business that survives $50 oil or $30 silver and thrives at $95 or $77.
Buybacks over Dividends
A small anecdote, but revealing. Gavin recounts a conversation with the Petrotal CEO in which he demanded a buyback program to juice returns. The CEO’s reply: “No, my wife wants her dividends.”
Gavin eventually sold his entire position after that, and PTAL has been a dog since (despite oil’s price rise!).
Commodity Cycle Timing — Bottoms and Tops
Gavin is not a cycle caller in the FinTwit sense. He just wants to find “the second inning” of commodity moves.
“I am for second inning. I don’t try to catch the bottom. I want to see a bounce off support at least like two times.”
He wants confirmation that the commodity has stopped going down, that the marginal-cost producers have been wiped out or absorbed, and that the equity tape is no longer selling bad news. In practice, it means looking for stocks that are bouncing off their lows, but have a solid support level.
Signals that the bottom is in
- Mine or well shut-ins are being reported, not projected. The supply side only tightens for real when producers are actually cutting — not when they’re warning.
- Insider buying is broad-based across multiple issuers, not just one. Concentrated insider buying at one company is noise; industry-wide insider buying is signal.
- The marginal producer is being recapitalized at a discount to NAV. Distressed financings at the bottom are how the capital structure resets.
- Generalist coverage has left. When sell-side notes are down to one analyst per name, the shareholder base has been fully purged.
Signals that the top is in
Gavin’s exit framework inverts everything from entry.
- New entrants pricing perfection. When venture-style money is building new issuers at multiples the operating asset cannot support, the top is close. He referenced NeurIPS as the AI version of this: champagne towers, venture excess, social proof over underwriting.
- High-operating-leverage producers printing money. The marginal producer outperforming the low-cost producer is a late-cycle signal. It’s why he sold gold miners in 2012 and why he’s watchful now.
- Insider selling, not buying. Reliable, mechanical. Management teams that bought at the bottom will be the first to tell you when they think the top is in.
- Retail “easy trade” narrative. When the name becomes consensus on Fintwit, and every newsletter is long, the repricing is done.
Gavin’s Contrarian AI Take
Gavin holds a PhD in AI, yet has no AI exposure inhis portfolio. Why? In his words, he “has thought harder about the underlying technology than most people who are long it.”
Credentials matter
Gavin’s academic path ran through biochemistry, neuroscience, and computer science, with exposure to the actual machinery of modern machine learning. His critique is not just that AI is in a bubble, and it will eventually pop. It’s that they can’t create anything new … they’re glorified iPod shufflers.
“LLMs can’t create something new. They compress and reshuffle. Compression isn’t valuable. It’s not the same thing as intelligence.”
The GPU glut risk
His investment-side concern is GPU supply. If model-compression techniques continue to improve at their current pace, the compute required to serve a given capability drops by an order of magnitude, and the hyperscaler capex cycle overshoots the revenue cycle. The result is a glut of depreciating hardware and a repricing of the entire AI value chain.
Helium, not semiconductors
The one piece of the AI supply chain Gavin is constructive on is helium: a small, irreplaceable input into advanced chip manufacturing, with structural supply constraints and no substitute.
Gavin hasn’t found any helium ideas to invest in yet, but he’s actively looking for them (as am I).
How he uses AI in his own process
Gavin uses AI narrowly, mainly to translate English into Python for modeling scripts and some web scraping applications. He explicitly does not trust it with financial numbers.
“I’m terrified of a hallucination. I would not be caught dead with any exposure to AI in my portfolio.”
Closing Thoughts: So Much To Learn
There were two reasons I loved this conversation with Gavin:
- Gavin and I had already fished in the same ponds before we connected and did the podcast.
- It exposed a major weakness in my investing process, namely, not holding positions long enough to see real rewards.
- It forced me to reflect on how I can improve going forward and “squeeze more juice” out of the ideas I find.
