Evinova Signs AI Deals with Astellas, BMS, and AstraZeneca 🔗

AstraZeneca's digital health spinout Evinova signed agreements with three major pharma companies to deploy its AI platform across clinical trial design and documentation. Astellas, Bristol Myers Squibb, and AstraZeneca will use Evinova's systems for study optimisation and document management, while providing operational data back to Evinova for benchmarking. President Cristina Duran framed these as long-term transformation commitments, not pilot projects. The fact that three top-20 pharma companies chose the same external AI platform for clinical development in a single week is a strong signal that the industry is converging on dedicated platforms rather than building in-house.

RadNet Acquires Gleamer for €230M 🔗

RadNet announced the acquisition of Paris-based radiology AI company Gleamer in a deal valued at €230 million. Gleamer brings four FDA-cleared devices and six CE-marked products across X-ray, CT, and MRI, with roughly $30 million in annual recurring revenue. Combined with RadNet's prior acquisitions of DeepHealth, iCAD, and See-Mode, this creates one of the broadest AI imaging portfolios in the world. CEO Howard Berger noted the deal will drive 45% to 55% growth in digital health sales this year and yield around $7 million in cost synergies. Consolidation in radiology AI is accelerating.

Veeva's $836M Quarter Confirms Life Sciences IT Spend Is Alive 🔗

Veeva Systems posted Q4 revenue of $836 million, up 16% year-on-year, beating Wall Street estimates. The company announced a $2 billion share buyback and reported over 125 customers on its next-generation Vault CRM, including major biopharma companies. More telling than the financials: CEO Peter Gassner said the primary growth driver is not AI adoption directly, but modernisation of legacy systems in safety and R&D. AI features are being rolled out as agents embedded into existing Veeva workflows, not as standalone products. For 2027, Veeva is guiding revenue between $3.585 and $3.6 billion.

The Signal: Veeva is the closest thing the life sciences software market has to a bellwether. When its numbers are strong, it means pharma IT budgets are flowing. The nuance here matters: Veeva's growth is coming from system modernisation first, with AI layered on top. That sequencing tells you where most of the industry actually is in its AI journey — upgrading the plumbing before turning on the intelligence.

Domino Data Lab Ships Agentic AI for Regulated Life Sciences 🔗
Domino Data Lab released a major platform update introducing what it calls an agentic development lifecycle for life sciences. The update includes universal tracing across every step of agentic AI creation, structured evaluation tools, and governed deployment interfaces. The pitch is specific: traditional ML workflows were not built to track and govern autonomous AI agents in regulated environments. Domino's update addresses this gap with compliance and traceability baked in.

The Signal: Agentic AI is moving fast, but in pharma and biotech, the question was never whether agents could do the work. It was whether they could do it with an audit trail that satisfies regulators. Domino is betting that the governance layer is the bottleneck, and that solving it first creates a moat. Worth watching as agentic deployments scale across drug discovery and clinical operations.

Data Sovereignty Becomes a Product Category 🔗

KALA BIO dominated headlines this week with its announcement of Researgency, an on-premises AI infrastructure platform licensed from Younet AI and designed specifically for biotech. The pitch: secure, data-sovereign AI that lets companies use advanced models without uploading proprietary biological data to third-party servers. With over 3,200 US biotech firms generating proprietary data but lacking in-house AI capabilities, KALA is targeting a real gap. In a parallel move, Insilico Medicine and Liquid AI launched a compact 2.6-billion-parameter drug discovery model that runs on private servers, delivering results that rival much larger systems while keeping sensitive data on-premise.

The Signal: Two very different companies, same thesis: the next wave of AI in pharma will be defined not by capability alone, but by who controls the data. The split between centralised and sovereign AI architectures is no longer theoretical. It is now a product category with real companies, real funding, and a clear buyer profile: any biotech sitting on proprietary datasets it cannot afford to expose but cannot afford to leave unanalysed.