A living document for thinking through how AGI reshapes the investment landscape. The goal is to build durable principles and a decision-tree framework that helps identify which companies benefit and which get hurt — and to update it as new information arrives.
This is not a prediction. It is a structured way to think about multiple possible futures and position accordingly.
Core Assumptions
- Target horizon: 5 years from now (~2031)
- Recursive self-improvement (RSI) begins late 2027 or early 2028
- Investment style: Buy and hold. Find assets where the floor is high and the upside is asymmetric.
Contents
1. Durable Principles 2. Decision Tree 3. Crypto / Bitcoin Thesis 4. Company Screener Criteria1. Durable Principles
These should hold regardless of which path AGI takes. They are the bedrock for all downstream analysis.
Greed, fear, herd behavior, status-seeking, desire for experiences, shelter, food, health. None of this changes in 5 years. Markets will overshoot and undershoot. The transition period will produce the biggest mispricings of our lifetime. Having dry powder during panic is itself a strategy.
AGI is a cognitive breakthrough, not a physics breakthrough. Energy conservation, speed of light, thermodynamics — these constrain what can actually happen in the physical world. No matter how smart the AI, you can't build a power plant in a week.
Clinical trials, construction, permitting, growing crops, building trust, regulatory approvals — these have irreducible time components. AGI can optimize the cognitive parts but not the waiting parts.
The legal system is slow. Even if AGI arrives, property rights, contracts, and corporate structures don't get rewritten overnight. If this assumption breaks, almost nothing is predictable.
When cognitive labor is abundant, it's worth nothing. Capital, physical assets, regulatory permissions, energy — these become relatively more valuable. Basic economics holds.
Digital monopolies (software, platforms) are vulnerable because AGI can replicate software. You can't replicate a pipeline, a transmission line, a port, water rights, or spectrum licenses with intelligence alone.
Since we genuinely cannot predict which branch we're on, positions with capped downside and uncapped upside are strictly better than point-bet predictions. This is Buffett's core insight and it applies even more under radical uncertainty.
Earth has finite surface area. We are not going to Mars in 5 years. Land is the one asset class that is definitively, physically scarce. Pursue land acquisition separately and independently of all other analysis.
2. Decision Tree
Key branching questions. As real-world signals come in, we can determine which path we're on and adjust accordingly. Click to expand.
How fast do capabilities improve once recursive self-improvement begins? This is the single most important variable.
- Disruption is sudden; most companies can't adapt in time
- Winners determined by what you already own, not what you do
- Physical assets, energy, and land dominate
- Massive market volatility (Principle 1: human nature → panic)
- Cash and liquidity to buy during panic is critical
- Disruption is gradual; incumbents can adapt
- Execution and adoption speed matter more
- Companies that integrate AGI fastest win
- More like a normal technology transition, just bigger
Signals to Watch
Rate of AI benchmark improvement after RSI milestone. If monthly jumps exceed what previously took years, we are on the Fast path.
Investment Implication
If Fast: you must already own physical/scarce assets before the panic. If Slow: you have time to pick best-adapting companies.
The cost of running AGI determines who captures the value and how competitive dynamics play out.
- Only big companies can deploy at scale
- Power concentrates in Big Tech + large enterprises
- Compute and energy are the bottleneck → NVIDIA, energy, data centers win
- Inequality increases between companies that can afford AGI and those that can't
- Everyone can deploy AGI — no moat from "having AGI"
- Competition drives margins to zero in knowledge work
- Moats come from physical assets, brands, and regulatory positions
- Deflationary pressure across the entire economy
Signals to Watch
Cost per token / cost per unit of cognitive work over time. Are AI labs competing on price or on capability? Is inference cost dropping 10x/year or plateauing?
Investment Implication
If Expensive: own the compute/energy supply chain. If Cheap: own the things AGI can't replicate (physical assets, regulatory moats, trusted brands).
AGI can think, but can it act in the physical world? This determines whether physical-world businesses have a protective moat or not.
- Physical labor also gets disrupted
- Construction, manufacturing, agriculture, logistics all automated
- Very few safe havens — only truly irreplaceable assets matter
- Land, energy sources, water, mineral rights
- Physical world businesses have a protected window
- Companies with skilled physical labor forces have a temporary moat
- Infrastructure, construction, maintenance, logistics retain value
- Window to position before robotics catches up
Signals to Watch
Humanoid robot deployment numbers. When Tesla/Figure/others go from hundreds to tens of thousands of deployed units. Cost per robot-hour vs. human-hour.
Investment Implication
If Fast: only own the irreplaceable (land, energy, rights). If Slow: physical-world businesses with skilled labor are temporarily undervalued by a market panicking about AGI.
AGI is clearly transformative. How do governments react? This shapes which companies keep their gains.
- Government controls or heavily taxes AGI deployment
- Regulated industries (healthcare, finance, defense) benefit from barriers
- Incumbents protected by regulation do well
- Innovation slows but existing structures persist
- Fastest deployers win, creative destruction at max speed
- Big tech + fast-moving startups dominate
- Massive disruption but also massive GDP growth
- Inequality explodes
- Government takes from winners, gives to population
- Consumer demand sustained despite job losses
- Consumer staples, housing, experiences retain value
- Likely funded by money printing → inflation hedge assets benefit
Signals to Watch
Political rhetoric around AI, executive orders or legislation within 6 months of AGI being obvious. Which countries move first and in what direction.
Investment Implication
Redistribution + money printing → inflation hedges (real assets, possibly crypto). Laissez-faire → fastest AGI adopters. Regulation → incumbents in regulated industries.
AGI disrupts employment and traditional business models. Human nature (Principle 1) means the market will likely overreact.
- Market slowly adjusts; some stocks drift down, others up
- Less opportunity for bargains
- Rational transition
- Market sells off indiscriminately
- Baby thrown out with bathwater
- Physical-asset-heavy companies sold off even though they're resilient
- This is the opportunity. Companies with real physical assets trading below replacement cost.
- Buy during maximum fear
Signals to Watch
VIX, credit spreads, unemployment claims rate of change, narrative in financial media. When CNBC starts running "Is your job safe?" segments daily, the panic is here.
Investment Implication
Have a watchlist of physically-durable companies ready before the panic. Know your buy prices in advance. Maintain cash reserves specifically for this scenario.
3. Crypto / Bitcoin Thesis
The Case For (strongest in Panic + Redistribution branches)
- Everything is being disrupted → people don't know where value lives anymore
- Bitcoin is a known, artificially scarce asset with global Schelling-point status
- If governments print money for redistribution → inflation hedge narrative strengthens
- Self-reinforcing loop: price rises → more people pile in → price rises more
- When everything else trends toward zero marginal cost, an asset with no marginal supply has unique appeal
The Case Against (strongest in Slow RSI + Laissez-faire branches)
- No cash flows — in a world where real assets generate 50%+ returns via AGI deployment, opportunity cost of holding BTC is enormous
- Government could ban or heavily regulate crypto as part of controlling capital flows during transition
- If AGI makes the real economy enormously productive, non-productive stores of value become less attractive
Positioning
Crypto makes sense as a hedge on the panic/redistribution branch, not as a core holding across all scenarios. Size the position proportionally to your probability estimate of that branch.
4. Company Screener Criteria
For screening all US public companies, these are the factors that hold up across multiple branches of the decision tree.
Physical Asset Intensity
Does the company own scarce physical assets or infrastructure that AGI cannot replicate? Durable across most branches.
Low Cognitive Revenue Exposure
What % of revenue comes from selling cognitive labor? High = existential risk. The company's product becomes free.
High Cognitive Cost Exposure
What % of costs are cognitive labor the company buys? High = margin expansion opportunity when AGI replaces it.
Competitive Position
Oligopoly/monopoly keeps the margin improvement. Competitive market passes savings to customers.
Energy / Compute Exposure
Does the company benefit from increased demand for compute and energy? Structural tailwind regardless of branch.
Low Debt
Survives any transition volatility. Companies with high leverage may not make it through a panic even if the underlying business is sound.