OpenAI said on X that it has acquired Torch, a health-tech startup it described as unifying lab results, medications and visit recordings, an announcement that comes days after OpenAI rolled out ChatGPT Health. OpenAI did not disclose financial terms, but The Information reported the deal was worth about $100 million in equity.
Why Torch fits ChatGPT Health
The acquisition aligns with the product problem OpenAI highlighted in its ChatGPT Health launch: health information is often scattered across portals, apps, wearables, PDFs and clinical notes, making it hard for patients, and providers, to see a coherent longitudinal picture. OpenAI said Health runs as a separate space with added privacy protections, and that conversations in Health are not used to train its foundation models.
Torch’s stated focus is that same “data fragmentation” layer, pulling lab results, medications, visit histories and other records into a unified profile that an AI system can reference consistently. Axios described Torch as a platform that integrates health data and includes visit recordings, positioning it as infrastructure OpenAI could fold into a records-connected ChatGPT workflow rather than a standalone consumer app.
Enterprise lens and deal signals
For U.S. healthcare CIOs, CMIOs and digital leaders, the Torch highlights the role of integration and governance in regulated environments.
PwC’s 2026 health services deals outlook argues AI is becoming a “genuine differentiator” in dealmaking and that acquirers are placing valuation premiums on platforms that can deliver measurable productivity gains, not “bolt-on” automation. In this context, OpenAI’s sequencing, launching Health, then acquiring data unification infrastructure, puts focus on connecting systems, standardizing data and making outputs usable in real workflows.
A separate biopharma deal was announced this week. AstraZeneca has acquired Modella AI, expanding a collaboration the companies announced in 2025. In its announcement, Modella said the acquisition will integrate its multimodal foundation models and AI agents across AstraZeneca’s global oncology portfolio to support clinical development and biomarker discovery.
Together, the Torch and Modella transactions reflect acquisition focused on proprietary data, domain workflows and validation loops, then embed them into larger product portfolios.
McKinsey has estimated generative AI could create $60 billion to $110 billion a year in economic value for pharma and medical-product industries, largely by accelerating discovery and development and improving productivity across the value chain. That value potential is pushing companies to invest more heavily in internal AI capabilities where advantage depends on data access, model adaptation and repeatable deployment.
What changes for buyers
For enterprise buyers, the shift is that Torch and Modella are no longer independent vendors. Torch and Modella were independent vendors; now their roadmaps become tied to OpenAI and AstraZeneca priorities.
For health systems evaluating GenAI partners, key questions include integration plans, data-handling assurances, and support models as capabilities move from standalone products into broader platforms, at the same time that more healthcare AI offerings are being packaged as end-to-end stacks rather than point solutions.