Investment Thesis Matrix
Bull/bear thesis per company in one synthesized view. The matrix is opinionated; treat it as a starting framework rather than a verdict.
Thesis matrix
| Company | Bull case (short) | Bear case (short) |
|---|---|---|
| CoreWeave | Infrastructure giant; multi-hundred-billion market cap | Customer concentration; debt-financed model strains in downturn |
| Crusoe | Stargate-anchored growth; energy advantage scales | Stargate execution risk; ESG narrative limits |
| Nebius | European Platinum-tier; engineering depth compounds | Geopolitical heritage; customer concentration |
| Lambda | Durable mid-market; brand compounds; IPO upside | Squeezed by both ends; capital ceiling |
| Together.AI | Open-source platform of record; lifecycle bundling | Hyperscaler absorption; margin compression |
| RunPod | DX-led mid-market platform; Serverless captures inference | Squeezed by specialists; sub-scale dedicated |
| Hyperbolic | Series-B scaling validates dual-product thesis | Sub-scale on both fronts; brand gap persists |
| Vast.AI | Marketplace network compounds; profitable indefinitely | Segment ceiling; GPU price normalization compresses value prop |
| TensorDock | Niche curation persists; small but durable | Vast dominance; small-scale moat |
Risk-return shape
- High return, high risk: CoreWeave, Crusoe, Nebius. Big upside if execution; big downside if customer concentration or capital structure breaks.
- Moderate return, moderate risk: Lambda, Together. Less explosive but more defensible.
- Modest return, low-to-moderate risk: RunPod, Hyperbolic, Vast, TensorDock. Bounded upside; capital-efficient business models.
Takeaway
The cast spans a meaningful risk-return spectrum. No single thesis applies. The final chapter looks at where the broader market is going.