Are there safe, defensible bets in AI investing?

Harry Stebbings · May 18, 2026

An investor's job is not to buy the obvious AI winners but to find the ancillary value created around them

You can agree with the thesis that AI value is real and still make more money by investing adjacent to it rather than in NVIDIA, AMD, Nebius or CoreWeave

81:42 20VC: Turning Peter Thiel's $100K into $10M Angel Portfolio | The One Man Accelerator at The Four Seasons | Why VCs Can Be Sharks and What Founders Need to Know | Why Stocks and Cash are BS and You Should Invest in Land with Josh Browder

Mike Mignano · Jul 6, 2026

USV did not miss the model layer; it deliberately played a different game by betting on energy as the layer that wins regardless of which model wins

If you believe in AI and intelligence, an energy layer is needed underneath to power it, and we have learned we need even more energy and more portability of energy than expected

Scope: USV invested in none of the big model-layer companies

28:48 20VC: Why Now is the Time for the Application Layer | Why OpenAI & Anthropic Won't Win the App Layer | Why Startups Should be TokenMaxxing | Why VCs Should Reduce Weighting on Price & Ownership in an Age of AI with Mike Mignano, USV

Jerry Murdock · Feb 28, 2026

Nothing in AI investing is safe right now; the only viable approach is judging who will execute best

No one can tell you what's safe — the returns come from a small minority of bets anyway

14:08 Nothing is safe back execution quality

Martin Casado · Jul 28, 2025

AI investing is paradoxical: you have to play at every layer of the stack because it's one of the fastest value creations in twenty years, yet it's very high risk because non-leaders have already been wiped out

The leaders in every layer keep compounding value while there have already been tons of wipeouts among non-leaders, so both abstaining and mis-picking are costly

13:44 Must invest across every layer but only leaders survive non leaders are wiped out

Cem Sertoglu · Nov 20, 2024

Every software company today is in some sense an AI company, so the right move is to play the game on the field rather than sit out AI — while staying out of hundred-million-dollar foundation model rounds

There is no software company that doesn't use machine learning and AI in some way to do its job better; accordingly all of the last five investments out of fund three are AI companies in one way or another

54:59 Every software company is now an ai company so invest broadly while avoiding mega foundation model rounds

Larry Aschebrook · Jun 16, 2025

The right AI strategy is a few concentrated bets in foundation model winners plus picks-and-shovels exposure, with the balance of capital deployed into other megatrends where AI is already embedded

AI is now embedded across cyber, SaaS, fintech and consumer, so exposure comes through the whole portfolio rather than only through AI-native bets

75:03 Concentrated foundation model bets plus picks and shovels plus broad ai embedded portfolio exposure