Life Sciences / Regulatory Review 🧬

📌 Navigate

📋 Exec Summary

Q3 2025 is the quarter AI crossed from experimental tool to operational infrastructure in life sciences. The FDA's AI/ML clearance pace kept accelerating — later 2025 totals reached 295 devices cleared and 1,451 cumulative, with median clearance times holding at 142 days. The EU AI Act's second enforcement wave hit in August (GPAI rules). The UK MHRA moved aggressively on digital reform with Early Access service, international reliance routes, and AI Airlock phase two. Isomorphic Labs raised $600M and expanded its Novartis partnership. Digital health funding showed selective recovery at $9.9B YTD. AI clinical trial tools entered production deployment with 65% enrollment improvement and 30-50% timeline acceleration.

📊 What Moved

📈 Trend Arcs

1. FDA AI/ML Clearance Acceleration — Velocity: Sustained High

295 clearances in 2025 set a new record, up from the prior year. Cumulative total reached 1,451 devices. Radiology CAD (product code QIH) still accounts for 25% of all clearances, but cardiology, pathology, and ophthalmology AI are gaining share. The 62% SaMD classification rate signals that standalone software — not just embedded device algorithms — is the dominant form factor. Median clearance time of 142 days means the FDA pathway is becoming predictable enough to build business models around.

2. Multi-Jurisdictional Regulatory Convergence — Velocity: Accelerating

Three parallel regulatory frameworks are advancing simultaneously. The EU AI Act enforced GPAI rules in August 2025, with high-risk medical device AI obligations phasing in August 2026-2027 (further shaped by the proposed Digital Omnibus package). The UK MHRA launched international reliance routes accepting FDA, TGA, and Health Canada approvals for streamlined UK market access. This creates a regulatory arbitrage opportunity: companies with FDA 510(k) clearance can now access the UK market faster while EU compliance timelines remain longer. Builders must plan for three parallel submission tracks.

3. AI-Native Drug Discovery Reaches the Clinic — Velocity: Inflecting

Isomorphic Labs raised $600M from Thrive Capital (March), expanded its Novartis collaboration, and is preparing to dose first patients in AI-designed oncology compounds — four years from founding to clinic. The $3B milestone potential across Eli Lilly and Novartis partnerships validates the commercial model. Meanwhile, AI clinical trial optimization tools hit production: BEKHealth and Dyania Health demonstrated 3x faster patient identification with 93-96% accuracy, and platforms are achieving 30-50% trial timeline compression. The gap between AI-assisted and traditional drug development is now measurable in years.

🗺️ Landscape Shift

🧭 Regulatory Direction of Travel

💰 Funding

🔍 Counter-Narrative

• The consensus: AI in life sciences is still mostly hype — real clinical impact is years away. The reality: The FDA cleared 295 AI/ML devices in a single year with a predictable 142-day median timeline — this is operational, not experimental. AI clinical trial tools are in production at major sites, demonstrating 65% enrollment improvement and 30-50% timeline acceleration. Isomorphic Labs is preparing to dose first patients in AI-designed compounds.
• The counter-counter: Most cleared AI/ML devices are decision-support tools, not autonomous diagnostic systems. Full autonomy in clinical settings remains genuinely years away. But the infrastructure layer — clearance pathways, trial optimization, manufacturing monitoring — is already delivering measurable ROI. Builders should focus on this infrastructure layer rather than waiting for the autonomous-diagnosis endgame.

📐 Builder's Benchmark

👀 What to Watch

• FDA Predefined Change Control Plans (PCCP) framework finalization — enables adaptive AI/ML devices without full re-submission
• EU Digital Omnibus legislative progress — potential 12-month extension for MDR-aligned AI systems
• Isomorphic Labs first-in-human trial initiation — validates AI-to-clinic pipeline economics
• Anthropic Claude for Life Sciences (post-quarter October launch) — signals frontier lab vertical strategy
• UK MHRA AI Airlock phase 2 outcomes — four AI technologies in regulatory sandbox
• Real-world evidence infrastructure investments — RWE becoming standard regulatory submission component
• AI clinical trial platform consolidation — expect M&A as pharma builds vs. buys
• CGT manufacturing AI adoption curve — watch for first fully AI-monitored commercial manufacturing run
• ISCT guidance document on AI/ML in CGT — industry-wide standards for digital manufacturing adoption
• Japan PMDA SaMD framework development — potential new fast-track for FDA-aligned AI devices
• Multi-modal AI diagnostics — combining imaging, genomics, and EHR data in single clearance applications

📎 Sources