1. What happened
A statement of work dated October 14, 2026 licenses oncology records to BMS, grants access to Tempus Lens, and includes analysis and research services through December 31, 2031.
2. Buyer / seller
Bristol Myers Squibb is the AI buyer or model-development side. Tempus AI owns or contributes the proprietary corpus.
3. Economics
$137.5 million in aggregate fixed payments, invoiced quarterly over the term; realized cash receipts are not verified. Economics scope: Contractual consideration covers the oncology data license, Lens access, and related analysis and research services; no allocation among those components is disclosed.. Strategic Signal has not estimated undisclosed consideration.
4. Proprietary data
Certain multimodal records from Tempus AI's de-identified oncology dataset, licensed to Bristol Myers Squibb for analysis and use with Pathos Foundry in therapeutic development. Historical depth: not_disclosed. Refreshability: The agreement runs through December 31, 2031 with quarterly invoicing; exact delivery cadence and update rights are not separately disclosed..
5. Rights and restrictions
Use is governed by the Onco SOW under the existing master services agreement and tied to defined oncology data, software access, analyses, and services. Ownership: No ownership transfer is disclosed.. Training: not disclosed. Post-training: not disclosed.
6. Why the data matters for AI
Analysis of de-identified oncology data with Tempus Lens and Pathos Foundry to support therapeutic product development; model training is not disclosed. The filing does not identify a decision-to-outcome field structure.
7. Asset Score — 86/100
The licensed de-identified oncology corpus is strategically valuable and renewable over the term, while historical depth and decision-to-outcome field structure remain undisclosed. Score coverage: 79%.
| Factor | Weight | Rating / 5 | Points |
|---|---|---|---|
| Uniqueness / scarcity | 12 | 4 | 9.6 |
| Historical depth | 8 | Unknown | — |
| Decision → outcome richness | 13 | Unknown | — |
| Decision Value Density | 13 | 4.5 | 11.7 |
| Real-world grounding | 10 | 5 | 10.0 |
| Domain economic value | 9 | 5 | 9.0 |
| Refreshability | 8 | 4 | 6.4 |
| Proprietary advantage | 9 | 4 | 7.2 |
| Model-learning usefulness | 10 | 4 | 8.0 |
| Non-replicability | 4 | 4 | 3.2 |
| Rights usability | 4 | 3.5 | 2.8 |
8. Transaction Signal Score — 87/100
The filed agreement isolates substantial fixed data-license economics and a defined term, but no competing process or independent model-uplift evidence is disclosed. Score coverage: 85%.
| Factor | Weight | Rating / 5 | Points |
|---|---|---|---|
| Disclosed economics | 16 | 5 | 16.0 |
| Clean price discovery | 16 | 3.5 | 11.2 |
| Separability of data consideration | 15 | 5 | 15.0 |
| Explicit training / model-improvement use | 14 | 3.5 | 9.8 |
| Clarity of rights purchased | 12 | 4 | 9.6 |
| Competing bids | 10 | Unknown | — |
| Strategic buyer quality | 6 | 5 | 6.0 |
| Repeatability / precedent value | 6 | 5 | 6.0 |
| Independent model-uplift evidence | 5 | Unknown | — |
9. M&A / valuation implications
The filed consideration is a useful commercial precedent for proprietary oncology data access, but bundled software and services prevent treating the full amount as a clean corpus-only price.
10. Evidence boundary
This record covers the Tempus-to-BMS oncology license only. The separate Tempus-to-Pathos neurology license is a distinct transaction. Training rights, competing bids, realized payments, and independent model uplift are not disclosed.