Beta
EconomyReading · ~3 min · 62 words deep

Customer Concentration

Customer concentration measures how much revenue depends on a few customers · AI labs often sit at 40-60% from top 10 vs SaaS at <20%.

TL;DR

Customer concentration measures how much revenue depends on a few customers · AI labs often sit at 40-60% from top 10 vs SaaS at <20%.

Level 1

If a company's top 10 customers generate 60% of revenue, customer concentration is high · a risk factor. In classic SaaS, top-10 concentration of <20% signals product-market-fit across a broad customer base. AI labs trend high because enterprise deals are massive · one Azure/Google contract can equal thousands of SMB subscriptions. Regulators and investors scrutinize concentration in diligence.

Level 2

OpenAI reportedly has Microsoft as >50% of revenue via revenue-share + Azure commitments. Anthropic has Amazon + Google as major customers post-Series E · concentration >40%. Mistral's enterprise tilt gives it 60%+ from top 10. High concentration = fragile (losing one customer hurts) but also validating (enterprise buyers staked their AI strategy). Low concentration = robust but usually means lower absolute revenue.

Level 3

Concentration math: measure revenue from top N customers / total revenue. Typical cuts: top 1, top 5, top 10. SaaS diligence benchmarks: top 10 < 20%. AI: top 10 < 50% is typical. Public filings require disclosure of any customer >10% of revenue · OpenAI, Anthropic are private but leaks suggest Microsoft/Amazon both cross 10% for respective labs. Mitigation: diversifying across enterprise, API, and consumer channels · trend for frontier labs is to dilute concentration by scaling API + consumer.

The takeaway for you
If you are a
Researcher
  • ·Revenue share from top N customers
  • ·SaaS: top 10 < 20% · AI: 40-60%
  • ·Public filing threshold: any single customer >10%
If you are a
Builder
  • ·Not actionable at team level
  • ·Signals partner risk when choosing an AI provider
  • ·Diversify to avoid being a vendor's 10%+ customer
If you are a
Investor
  • ·Primary risk factor in AI-lab diligence
  • ·High concentration = fragile but validating
  • ·Watch trajectory · concentration should decline as revenue scales
If you are a
Curious · Normie
  • ·How much money comes from just a few customers
  • ·Risky if too much · one bad deal could hurt the business
  • ·AI companies tend to have more concentration than other tech
Gecko's take

AI customer concentration is the biggest unhedged risk most frontier labs carry · watch how they diversify as API and consumer scale.

SaaS: top 10 < 20%. AI: 30-50% at current stage is acceptable given enterprise scale.