Picks & Shovels
Selling tools to AI builders
The Pattern
In every gold rush, the most reliable winners sell picks and shovels. NVIDIA doesn’t build AI apps — it sells the GPUs everyone needs to train them. TSMC doesn’t design chips — it manufactures everyone else’s designs. ASML doesn’t make or design chips — it builds the $350M lithography machines that make manufacturing possible.
Each layer of the AI infrastructure stack extracts value from the boom without betting on which specific applications win. NVIDIA captured 90%+ of AI training chip spending — the single greatest supply-chain monopoly in modern technology.
Key Metrics & Benchmarks
Who Uses This Pattern
Strengths & Weaknesses
STRENGTHS
- Profits regardless of which AI applications win
- Massive demand with limited supply creates pricing power
- Hardware/infra moats are harder to replicate than software
- Every AI gold rush participant is your customer
WEAKNESSES
- Enormous capex required ($10B+ fabs, $1B+ data centers)
- Technology risk — new architectures could displace incumbents
- Customer concentration (hyperscalers are biggest buyers)
- Cyclical risk if AI investment slows
How AI Is Transforming This Pattern
The AI picks-and-shovels opportunity is the largest infrastructure buildout since cloud computing. Over $200B/year is being spent on AI infrastructure. The question: will NVIDIA’s concentration persist, or will custom chips (Google TPU, Amazon Trainium, AMD MI300) dilute the monopoly? History suggests the monopolist retains 60-70% share even as alternatives emerge.
Business Engineer Insight
Picks & Shovels is the lowest-risk, highest-margin way to participate in any technology wave. The strategic insight: position at a chokepoint where demand is inelastic and alternatives are scarce. NVIDIA’s CUDA ecosystem is a bigger moat than silicon — millions of developers trained on CUDA create software lock-in that transcends any single hardware generation.
Related Patterns
Understand the strategic architecture behind this business model pattern — and how the best companies deploy it for competitive advantage.