The very thing venture capital is betting billions on is about to be worth as much as cloud storage.
The Summary
- General Catalyst's Madhu Namburi says AI intelligence will commoditize over time, fundamentally reshaping how investors approach dealmaking
- The thesis: raw AI capability becomes table stakes, not competitive advantage
- What matters next: what you build on top of the commodity, not the commodity itself
The Signal
Madhu Namburi, managing director and co-head of General Catalyst's creation fund, laid out a future where AI intelligence follows the same arc as computing power, storage, and bandwidth. All became cheap and ubiquitous. All stopped being the business model and started being infrastructure.
This isn't a hot take from an outside observer. This is coming from someone actively deploying capital into AI companies. When your own investors start calling your core technology a future commodity, the smart founders are already building the next layer up.
"Intelligence will become a commodity over time."
The commoditization pattern is predictable: early scarcity drives massive investment, competition scales supply, margins compress, differentiation moves up the stack. It happened to data centers. It happened to bandwidth. Now it's happening to intelligence itself.
For dealmaking, this changes everything. Investors are no longer betting on who can build the best model. They're betting on:
- Who can deploy AI agents that solve actual business problems
- Who can create moats in data, distribution, or domain expertise
- Who can build workflows that compound over time, not just inference runs
The signal here isn't that AI is overhyped. It's that AI is under-hyped in the wrong direction. Everyone's focused on foundation models. The real value is in what gets built when inference costs approach zero and every company has access to frontier-level intelligence.
The Implication
If you're building on raw AI capability as your moat, you're building on sand. The question is what proprietary data you have, what network effects you can create, what workflows you can own. Commoditized intelligence means intelligence becomes the floor, not the ceiling. The companies that win will be the ones who assume GPT-5 level reasoning is free and build from there.
For investors, this means the diligence shifts. Less "how good is your model?" More "what happens when everyone has a model this good?" The companies that have good answers to that question are the ones that survive the commoditization wave.