Psychology, Value, and Data: How Santiago Santos Navigates The Markets

Santiago Santos sold the top of the crypto market in 2021 and is a former partner at Parify. He traces his original interest in crypto to the remittance problem, noting that moving money from the US to Mexico costs at least five to six percent in fees and that remittances are the first or second source of income for Mexican families. Bitcoin represented a faster and cheaper solution, but he now believes stablecoins have inherited that killer use case regardless of Bitcoin price. His broader view is that finance outside the US and developed countries is inefficient by design, with friction built into the system intentionally.

Santos founded Inversion as a holding company rather than a fund specifically to avoid the pressure of deploying committed capital on a deadline. He evaluated more than 20 sectors where stablecoins could meaningfully transform traditional businesses and ultimately concluded that filtering for crypto led Inversion toward businesses that were more fragile than businesses he would want to own. Remittance companies like Remitly and Wise were examined closely and found to have poor unit economics, no pricing power, and agents and fulfillment partners functioning as the real intermediaries. Larger remittance companies publicly claimed stablecoins would release working capital, but deep investigation could not confirm the business could be made more efficient. Crypto native businesses like Binance, Tether, and Hyperliquid were phenomenally good businesses on their own terms, but the broader thesis did not hold.

The core lesson Santos drew is that technology is orthogonal to a good business rather than central to why you buy one. He now plans to pivot Inversion toward buying good businesses regardless of crypto fit, applying the Buffett and Munger model of finding good operators, providing capital, and compounding. A good business in his framework produces meaningful cash flow through any market environment, has moats and pricing power, and faces manageable disruption risk. If no such business can be found, he believes the right move is to buy the benchmark rather than force an investment.

One sector Santos identified as fitting this framework is critical power servicing. Hospitals, schools, and hyperscalers are federally mandated to maintain and regularly service critical power generators, and failure to do so can result in jail time for hospital administrators. The industry is highly fragmented with geographic constraints and labor shortages. Emergency dispatch can be billed at three to four times the standard rate, and Santos sees AI applied to routing, scheduling, and report generation as a way to achieve margin uplift and take on more clients without proportional cost increases.

Santos chose not to charge a two percent management fee, which he believes preserved his ability to avoid deploying capital into a thesis he no longer believed in. He argues that fund managers charging that fee face incentive distortions that push them to deploy regardless of conviction, and that growing a fund from 20 million to 2 billion dollars transforms a manager psychologically from an investor into an asset manager. He cites Benchmark as a firm that maintained discipline by keeping its fund size at approximately 300 to 400 million dollars for an extended period. He warns that fundraising skill and actual investing skill can be dangerously decoupled, and cites Long-Term Capital Management as a case where miscalculated risk and a misjudged probability of Russian bond default caused a blowup.

On crypto more broadly, Santos holds that the technology has not failed but that valuations and non-technology issues including fraud and scams are legitimate concerns. He views meme coins as a manifestation of an affordability crisis and a primitive desire to escape financial hardship, comparable to buying lottery tickets during recessions. He believes the excuse of blaming regulators for crypto's lack of progress has largely expired and that the industry should now self-regulate, citing the Token Transparency Act as one example.

Santos is very long on pharma companies with large proprietary data sets, arguing that AI models can make cross-domain relationships across those data sets that no single scientist could replicate. He says academic defensiveness, where scientists protect theories they have built careers around, has handicapped drug discovery in the same way ego handicaps investing, and that AI breaks through this because it does not care about credentials. He identifies compute infrastructure, energy, and healthcare as AI-resistant sectors regardless of how AI develops, while flagging hallucination risk as real and noting that model weights and the information being served by AI companies deserve close attention. He argues that hyper-connectivity creates an illusion of being informed while algorithms serve self-serving content, making markets more inefficient rather than less, and that the current moment is the best time in history to be a generalist because AI accelerates cross-domain discovery.