$1.7B for Takeda x Iambic last week. Now Merck skips techbio entirely and goes to Mayo Clinic.

TL;DR

$1.7B for Takeda x Iambic last week. Now Merck skips techbio entirely and goes straight to Mayo Clinic. The deal itself is interesting. The architecture behind it is what matters. Mayo isn’t licensing a static dataset — through Platform_Orchestrate, Merck runs its AI models (including virtual cell technologies) inside Mayo’s secure environment, against multimodal data: clinical notes, genomics, imaging, molecular data. Mayo’s first collaboration of this scale with a global biopharma, and a deliberate departure from healthcare’s culture of proprietary contracts and siloed data. Initial focus: IBD, atopic dermatitis, multiple sclerosis. The bigger question: which health system builds the next pharma-grade data platform?

$1.7B for Takeda x Iambic Therapeutics last week. Now Merck skips the techbio entirely and goes to straight to Mayo Clinic.

The deal itself is interesting. The architecture behind it is what matters.

Mayo isn’t licensing a static dataset. Through their Platform_Orchestrate program, Merck runs its AI models (including virtual cell technologies) inside Mayo’s secure environment, against multimodal data: clinical notes, genomics, imaging, molecular data. All de-identified, all connected.

This is Mayo’s first collaboration of this scale with a global biopharma company. And it follows a deliberate strategy. Mayo’s platform team explicitly designed this as a departure from healthcare’s deeply rooted culture of proprietary contracts and siloed data.

Initial focus: IBD, atopic dermatitis, multiple sclerosis. All areas where treatment response varies wildly and real-world multimodal data can sharpen target identification.

The bigger question: if this model works, which health system builds the next pharma-grade data platform?

Key takeaways

• Two pharma-AI deals in two weeks, two structurally different architectures: Takeda acquires Iambic’s techbio stack outright; Merck runs its own models inside Mayo Clinic’s data.
• Mayo isn’t licensing a static dataset. Merck runs its AI models — including virtual cell technologies — inside Mayo’s secure environment via Platform_Orchestrate.
• The data covered is multimodal: clinical notes, genomics, imaging, molecular data. All de-identified, all connected.
• Mayo’s first collaboration of this scale with a global biopharma, and a deliberate break from healthcare’s culture of proprietary contracts and siloed data.
• Initial focus areas (IBD, atopic dermatitis, multiple sclerosis) were chosen because treatment response varies wildly and real-world multimodal data can sharpen target identification.
• The bigger strategic question: if compute-inside-the-data works, which health system builds the next pharma-grade data platform?