YC AI Startup Bets
The Infrastructure Layer (Picks & Shovels)
YC bet: the tools and infra every AI company needs.
The 'picks and shovels' bet is the idea that some AI companies win by building the tools and infrastructure every other AI company needs — not the end apps. In a gold rush, the ones selling picks and shovels often win.
In simple terms
Build the infrastructure every AI company needs, rather than a single AI app.
How it works
- 1The AI development stack has layers: applications, frameworks/orchestration, data & context, model layer, compute & infrastructure.
- 2Picks-and-shovels companies build tools and infrastructure for every layer of the stack.
- 3They enable other companies to build AI, boosting developer productivity.
- 4They remove technical bottlenecks and power the wider AI ecosystem.
- 5As AI adoption grows, so does demand for this foundational layer.
Key points
- They don't build AI apps — they build what apps are built on.
- Examples span vector DBs, context management, model APIs, inference and compute.
- Every successful AI application is built on top of these tools.
- It's a large, long-term opportunity as AI scales.
Why it matters
The picks-and-shovels framing helps you spot durable AI opportunities: the infrastructure layer grows with the whole ecosystem, regardless of which apps win.
Frequently asked questions
- What does 'picks and shovels' mean?
- Selling the tools and infrastructure others need to build, rather than the end product — like selling shovels in a gold rush.
- Why is it a strong bet?
- Demand for foundational tools grows with the whole AI market, not just one app's success.