FDA Rejects Proposal to Ease Oversight of AI Medical Devices 🔗

The FDA declined a request from Harrison.ai, an Australian health AI company, to reduce premarket review requirements for AI-enabled diagnostic tools from developers with prior approval track records and robust post-market monitoring. The agency held to its existing framework, signalling that demonstrated safety history does not automatically justify lighter regulatory scrutiny for new AI systems making clinical decisions. With over 1,000 AI-powered medical devices currently authorized, this decision draws a clearer line between AI tools that touch administrative workflows and those that influence diagnosis and treatment. For developers attempting to distinguish themselves on compliance grounds, the message is that regulatory differentiation will be earned through evidence, not reputation.

Boehringer Ingelheim and Brainomix Take Lung Fibrosis AI Into Prospective Trial 🔗

Following results from REVISE-PPF, a retrospective study demonstrating that Brainomix's e-Lung software could reduce diagnosis times for progressive pulmonary fibrosis by over two years compared to standard methods, the partners have announced PROGRESS-PPF: a prospective clinical trial evaluating e-Lung in real-time care settings across U.S. pulmonary facilities. The software analyzes CT scans to detect early indicators of interstitial lung disease and carries FDA approval. Moving from retrospective validation to prospective trial is the step that separates AI diagnostic tools that show promise from those that can claim clinical evidence. As payers and health systems apply more scrutiny to AI diagnostic claims, that distinction is becoming commercially significant.

Viz.ai Launches AI-Powered Pulmonary Care Suite 🔗

Viz.ai has launched a Pulmonary Care Suite integrating acute and chronic pulmonary workflows into a unified platform connected to EHR systems. The suite covers COPD, pulmonary embolism, and lung nodule pathways, with AI-based pulmonary embolism detection from imaging and chart summarization linked to clinical guidelines. Viz.ai operates across approximately 2,000 hospitals, reaching around two-thirds of the U.S. population. The pulmonary suite marks a deliberate expansion beyond the cardiovascular and neurological workflows that anchored the platform, pointing toward a strategy of multi-specialty coverage within hospital systems already running its infrastructure — deepening the footprint rather than winning new accounts.

Merck KGaA Raises 2026 Outlook as Bioprocessing Demand Accelerates 🔗
Merck KGaA reported Q1 net sales of €5.1 billion with 2.9% organic growth, driven by its Life Science and Electronics divisions. The Process Solutions segment grew 16.2% organically, surpassing €1 billion in quarterly sales on the back of increased demand for downstream processing and single-use technologies. The company raised its full-year 2026 guidance to up to €21.4 billion in net sales. Electronics growth was tied explicitly to AI infrastructure demand through semiconductor materials.

The signal: Merck KGaA occupies the supply layer underneath the drug discovery platforms generating headlines. When its Process Solutions business grows at 16% organically, it reflects real production activity in biologics and cell therapy manufacturing, not announcements. Upstream reagent and single-use technology demand is one of the cleaner leading indicators of where actual manufacturing activity is heading. The AI infrastructure investment story has a physical substrate, and it is growing.

Novo Nordisk Partners with Cellular Intelligence to Apply AI to Parkinson's Cell Therapy Manufacturing 🔗

Novo Nordisk has entered a collaboration with Cellular Intelligence to apply the company's AI foundation model to its allogeneic pluripotent stem cell-derived dopaminergic progenitor therapy, currently in Phase I/II clinical trial. Cellular Intelligence's platform is designed to optimize cell therapy process development and reduce manufacturing timelines by learning from experimental data across cell types and conditions.

The signal: The manufacturing science for cell therapies has not kept pace with the biology. As allogeneic programs move through Phase II and Phase III, manufacturing scale-up will be the operational bottleneck for most companies in this space. Novo Nordisk choosing an external AI manufacturing platform for a clinical-stage program signals that process optimization is being treated as a specialist AI domain, not an internal process development extension. Platforms that can demonstrate reproducibility across manufacturing conditions at clinical scale will have a structural advantage as the pipeline matures.

Google Launches $10 Million Quantum AI Program for Life Sciences 🔗

Google Quantum AI and Google.org have launched REPLIQA (Research Program at the Intersection of Life Sciences and Quantum AI), a $10 million initiative funding research at Harvard, MIT, UC San Diego, UC Santa Barbara, and the University of Arizona. Research targets include simulation of protein folding and cellular processes at resolution beyond classical systems, quantum modeling of the P450 enzyme relevant to drug metabolism, and quantum sensor development for real-time metabolic observation. The program is led by Hartmut Neven and is positioned as long-duration foundational research with practical applications expected in the early 2030s.

The signal: The dollar amount is modest relative to the funding rounds tracked in this newsletter. The significance is different: Google is formally connecting quantum computing to molecular biology as a named research program with specific institutional partners and defined targets. Quantum AI in life sciences has existed as a theoretical category for several years. REPLIQA represents an institutional commitment to working out the use cases at foundational level, with a realistic timeline. The platforms being built now will determine what the AI drug discovery field runs on a decade from now. Worth tracking as a long-duration story.

Anthropic Partners with the Gates Foundation on $200 Million Global Health AI Initiative 🔗

Anthropic and the Gates Foundation have announced a four-year, $200 million partnership targeting AI applications in health, education, and economic mobility in low- and middle-income countries. The health component is the most directly relevant for this audience: the initiative will apply Claude to drug and vaccine discovery workflows, disease surveillance, and clinical decision support in settings where those capabilities are currently unavailable at scale. Open datasets for underrepresented languages are part of the plan, addressing a known gap in AI accessibility for African health systems.

The signal: Frontier AI labs are no longer limiting their life sciences ambitions to commercial pharma. Anthropic has now committed meaningful capital and technical resources to health AI in LMIC settings — a market that global pharma has historically underinvested in. If that work produces usable infrastructure (validated models, open datasets, clinical decision tools), it will eventually feed back into the broader life sciences AI stack. The partnership is worth tracking less as a Gates Foundation story and more as a signal that the leading AI labs view healthcare as a primary domain, not a vertical application.