Uploaded April 2026 | Updated September 2026, 1 week ago
The Investing Framework I Built Over 30 Years Just Broke.
Here's what I'm replacing it with.
For decades, I evaluated every company through nine lenses. Moat durability. Cash flow. Capital allocation. Founder excellence, etc.. The framework built fortunes in the pre-AI world.
Then I stress-tested it against 2026. Half of it is quietly obsolete.
The shift in one sentence:
We are moving from "durable competitive advantages" to "compounding adaptive advantages."
The old question — how strong is this moat?
The new question — does this business get stronger or weaker as AI gets 10x better next year?
That's not a hypothetical. That's the world we're already in.
What still sticks: Cash is oxygen. If a company can't turn a prompt into a dollar, it's a hobby — not a business. Capital allocation. Founder obsession. Relative value. Iron laws.
What quietly broke: Quarterly latency — if you're waiting on a 10-Q, you're 90 days late. Linear 10-year TAM runways — they collapse to three. Static moats aren't assets anymore. They're targets.
What's new: AI Leverage Ratio. Switching Costs 2.0 — no longer software lock-in, it's AI integration depth. Inference Economics. Speed to Adapt — AI-native, or AI-bolted-on?
The real story is the bottlenecks. Hyperscalers will spend roughly $650B on AI capex in 2026 alone — bigger than the Apollo program in real dollars.
I've organized my entire AI book around four bottlenecks:
Power. Vistra. Constellation. Talen. Cheniere. Cameco. GE Vernova. Caterpillar. Vertiv. Bloom Energy. Ormat.
Photonics — the "Speed of Light" play. Copper is dying inside the AI cluster. Lumentum. Corning. Applied Optoelectronics. Aehr. AXT. Sivers.
Efficient Compute. Every watt saved is a watt you don't have to generate. Navitas. Nebius.
Enabling Infrastructure. Sandisk. Texas Pacific Land. Sterling Infrastructure. Carpenter Tech. Teradyne. Clean Harbors.
Plus the core — Tesla, Nvidia, Lilly — and the more speculative cohort I size like venture, not core.
In the full session, I also walk through the most recent 13F of one of the sharpest AI-thesis funds out there — and the position-management moves worth studying.
I cover the two private bottleneck bets that could become the "Standard Oil" and the "Intel Inside" of the post-energy-wall AI era.
And the public-market wrapper that gives you indirect SpaceX exposure 60-90 days before what could be the largest IPO in history. The math here is genuinely asymmetric.
Watch the full video.
If this resonated and you want to be in the rooms where peers, founders, and the sharpest allocators in the world stress-test exactly this kind of thinking — that is what SFO Continuity is built for.
Not because of one idea. Because of the compounding of relationships and conversations across cycles. We share what we own. What we passed on. What we got wrong. What we're stalking next.
That second-order edge is the most undervalued asset in modern investing.
This era doesn't reward the patient observer. It rewards the tactical builder.
Apply to join → angelorobles.com/membership
For informational purposes only. Not investment advice. When you invest you can lose all your money.
#FamilyOffice #AIInvesting #InvestmentStrategy #AIInfrastructure #Photonics #NuclearEnergy #HedgeFunds #SpaceX #PrivateMarkets #UHNW #PortfolioStrategy #CapitalAllocation #SFOContinuity #AssetAllocation #WealthManagement #angelorobles
The Investing Framework I Built Over 30 Years Just Broke.
Here's what I'm replacing it with.
For decades, I evaluated every company through nine lenses. Moat durability. Cash flow. Capital allocation. Founder excellence, etc.. The framework built fortunes in the pre-AI world.
Then I stress-tested it against 2026. Half of it is quietly obsolete.
The shift in one sentence:
We are moving from "durable competitive advantages" to "compounding adaptive advantages."
The old question — how strong is this moat?
The new question — does this business get stronger or weaker as AI gets 10x better next year?
That's not a hypothetical. That's the world we're already in.
What still sticks: Cash is oxygen. If a company can't turn a prompt into a dollar, it's a hobby — not a business. Capital allocation. Founder obsession. Relative value. Iron laws.
What quietly broke: Quarterly latency — if you're waiting on a 10-Q, you're 90 days late. Linear 10-year TAM runways — they collapse to three. Static moats aren't assets anymore. They're targets.
What's new: AI Leverage Ratio. Switching Costs 2.0 — no longer software lock-in, it's AI integration depth. Inference Economics. Speed to Adapt — AI-native, or AI-bolted-on?
The real story is the bottlenecks. Hyperscalers will spend roughly $650B on AI capex in 2026 alone — bigger than the Apollo program in real dollars.
I've organized my entire AI book around four bottlenecks:
Power. Vistra. Constellation. Talen. Cheniere. Cameco. GE Vernova. Caterpillar. Vertiv. Bloom Energy. Ormat.
Photonics — the "Speed of Light" play. Copper is dying inside the AI cluster. Lumentum. Corning. Applied Optoelectronics. Aehr. AXT. Sivers.
Efficient Compute. Every watt saved is a watt you don't have to generate. Navitas. Nebius.
Enabling Infrastructure. Sandisk. Texas Pacific Land. Sterling Infrastructure. Carpenter Tech. Teradyne. Clean Harbors.
Plus the core — Tesla, Nvidia, Lilly — and the more speculative cohort I size like venture, not core.
In the full session, I also walk through the most recent 13F of one of the sharpest AI-thesis funds out there — and the position-management moves worth studying.
I cover the two private bottleneck bets that could become the "Standard Oil" and the "Intel Inside" of the post-energy-wall AI era.
And the public-market wrapper that gives you indirect SpaceX exposure 60-90 days before what could be the largest IPO in history. The math here is genuinely asymmetric.
Watch the full video.
If this resonated and you want to be in the rooms where peers, founders, and the sharpest allocators in the world stress-test exactly this kind of thinking — that is what SFO Continuity is built for.
Not because of one idea. Because of the compounding of relationships and conversations across cycles. We share what we own. What we passed on. What we got wrong. What we're stalking next.
That second-order edge is the most undervalued asset in modern investing.
This era doesn't reward the patient observer. It rewards the tactical builder.
Apply to join → angelorobles.com/membership
For informational purposes only. Not investment advice. When you invest you can lose all your money.
#FamilyOffice #AIInvesting #InvestmentStrategy #AIInfrastructure #Photonics #NuclearEnergy #HedgeFunds #SpaceX #PrivateMarkets #UHNW #PortfolioStrategy #CapitalAllocation #SFOContinuity #AssetAllocation #WealthManagement #angelorobles
