Tempus AI (TEM)

Type: Public multimodal healthcare data and AI company — NASDAQ: TEM

Tempus AI operates the largest US private multimodal oncology dataset: genomics, clinical records, imaging, pathology — connected and annotated. Its business model is to turn this data into an operating system for precision medicine, monetized through: (1) data licensing to pharma for drug discovery and clinical trial design, (2) clinical decision support tools (the "Tempus ONE" AI physician assistant), and (3) clinical trial matching.

Why it matters for AI-bio: Tempus is the most important source of real-world multimodal biological data for pharma AI programs. Unlike Recursion (cellular phenomics, preclinical) or Schrodinger (computational chemistry), Tempus's data links to human clinical outcomes — the hardest and most valuable data type. The AstraZeneca/Pathos deal (2024) is the most prominent example: pharma paying to access Tempus's dataset to train their own AI models.

Key metrics

• Net revenue retention: ~126% (2025) — customers expand spend once integrated
• Dataset: largest private multimodal oncology dataset in the US (exact size not public)
• Revenue model: data licensing + SaaS + clinical tools

Financials

• 10-Q filed FY2026; strong YoY revenue growth (SEC EDGAR)

Key relationship

• AstraZeneca/Pathos deal — Pharma licensing Tempus data for AI model training; validates the data monetization model

What to watch

• Whether 126% NRR sustains into 2026–2027 (the key metric for moat defensibility)
• Expansion from oncology into cardiology, neurology, and metabolic disease
• Whether CMS/payer adoption of Tempus genomic testing drives top-of-funnel data volume