1. What happened
Purpose-limited data license to Pathos within a three-party arrangement involving AstraZeneca.
2. Buyer / seller
Pathos AI is the AI buyer or model-development side. Tempus AI owns or contributes the proprietary corpus.
3. Economics
$200m data-license fees over three years, including $50m upfront payable. Up to 50% may be settled in Pathos preferred shares. Separate $35m flows run from AstraZeneca to Tempus and Tempus to Pathos. Economics scope: contractually identified data license fees within broader arrangement. Strategic Signal has not estimated undisclosed consideration.
4. Proprietary data
A de-identified multimodal oncology dataset licensed for a foundation-model development program. Historical depth: not_disclosed. Refreshability: not disclosed.
5. Rights and restrictions
License purpose is development and training of the specified model. Ownership: no disclosed transfer. Training: yes explicit purpose limited. Post-training: not disclosed.
6. Why the data matters for AI
Develop and train an oncology foundation model. unknown; the filing does not define longitudinal outcome fields
7. Asset Score — 82/100
Potentially valuable domain training corpus; uncertainty about contents, age and renewal materially limits coverage. Score is strategic analysis, not measured utility. Score coverage: 71%.
| Factor | Weight | Rating / 5 | Points |
|---|---|---|---|
| Uniqueness / scarcity | 12 | 4 | 9.6 |
| Historical depth | 8 | Unknown | — |
| Decision → outcome richness | 13 | Unknown | — |
| Decision Value Density | 13 | 4 | 10.4 |
| Real-world grounding | 10 | 4 | 8.0 |
| Domain economic value | 9 | 4.5 | 8.1 |
| Refreshability | 8 | Unknown | — |
| Proprietary advantage | 9 | 4 | 7.2 |
| Model-learning usefulness | 10 | 4.5 | 9.0 |
| Non-replicability | 4 | 3.5 | 2.8 |
| Rights usability | 4 | 3.5 | 2.8 |
8. Transaction Signal Score — 75/100
Explicit fee allocation is valuable pricing evidence. Equity settlement, reciprocal commitments and missing comparable bids prevent treatment as a clean cash market price. Score coverage: 85%.
| Factor | Weight | Rating / 5 | Points |
|---|---|---|---|
| Disclosed economics | 16 | 4.5 | 14.4 |
| Clean price discovery | 16 | 2 | 6.4 |
| Separability of data consideration | 15 | 4 | 12.0 |
| Explicit training / model-improvement use | 14 | 5 | 14.0 |
| Clarity of rights purchased | 12 | 3.5 | 8.4 |
| Competing bids | 10 | Unknown | — |
| Strategic buyer quality | 6 | 3 | 3.6 |
| Repeatability / precedent value | 6 | 4 | 4.8 |
| Independent model-uplift evidence | 5 | Unknown | — |
9. M&A / valuation implications
Analysis: useful license-fee precedent, not a clean cash sale valuation. Comparability depends on rights, payment form, scope and linked obligations; no annual run-rate or per-patient price is inferred.
10. Evidence boundary
This observation covers the data-license leg only. Payment completion, corpus renewal and many contractual restrictions remain unknown. Related fees and financing are context, not additional data transactions.