Foundation models
ModerateFrontier labs and challengers building LLMs, multimodal, reasoning, and agentic systems.
Capability defines what the rest of the stack can charge for. But monetization, margin durability, and the long-run number of viable frontier labs are all live questions.
Companies
16
6 pub · 10 priv
Public mcap
$12.76T
Public revenue
$1.91T
Wtd YoY
20%
Wtd GM
62%
Wtd OM
33%
Avg fund
62
Avg tech
49
Market structure
Current TAM
$25.00B
5y TAM
$200.00B
5y CAGR
50.0%
Companies
16
Margin structure
Mixed
Concentration
Concentrated
Capex intensity
Extreme
Commoditization
High
Drivers
What grows the pie
- ↑Token volume
- ↑Reasoning model adoption
- ↑Agent workloads
Constraints
What can break the thesis
- ↓Capex per training run
- ↓Model commoditization
- ↓Open-weights pressure