Life Sciences / Regulatory Brief 🧬

📌 Navigate

📊 Exec Summary

FDA opened 2026 with its largest digital health deregulation in a decade — revising CDS and general wellness guidance to remove premarket review from single-recommendation tools and non-invasive wearable monitoring — while JPMorgan coverage recycled 2025 deal tape as historical comparison, not 2026 conference-native M&A.

Five things moved in the January opening window:

The pattern: FDA is deregulating the front door (CDS, wellness wearables, TEMPO reimbursement linkage) while capital is building the AI execution infrastructure (NVIDIA/Lilly, JPM M&A) — the regulatory and investment cycles are converging on the same 12-month window.

1️⃣ FDA Revises CDS and General Wellness Guidance

TL;DR: On January 6, CDRH issued two revised final guidances expanding non-device classifications for CDS software and wellness wearables — the most significant digital health deregulatory action since the 21st Century Cures Act, announced by Commissioner Makary at CES 2026.

What happened

• January 6, 2026: FDA issued revised final guidance on Clinical Decision Support Software and revised General Wellness Policy for Low Risk Devices
• CDS Criterion 3 (HCP recommendations): Enforcement discretion now applies to single-recommendation tools previously requiring 510(k); risk score functions and predictive tools newly covered without premarket review
• CDS Criterion 2 (medical information): Broadened — relevance no longer requires "normally communicated between HCPs"; now "supported by well-understood and accepted sources"
• CDS Criterion 4 (transparency): Reduced to "summary of general approach" — detailed algorithm disclosure no longer required
• General wellness: Non-invasive sensing (optical) estimating blood pressure, SpO2, glucose, HRV — now non-device if wellness-only intended use, no disease claims, no clinical accuracy claims
• Exceptions still requiring premarket review: Variant genomic data lacking established relevance; predictions within 24-hour timeframe
• Manufacturer constraints: Cannot label outputs as "abnormal"; cannot include clinical thresholds; must document clinical relevance of inputs
• Commissioner Makary announced at CES 2026 in video on X and public remarks; guidance issued without public comment period

📊 Key facts

🔗 Primary source → FDA Law Blog: CDS and General Wellness Guidance Updates

🔍 The non-obvious point

Issuing without a public comment period is the tell — this is policy signaling, not completed rulemaking, and the ambiguities are intentional.

• "Wellness-appropriate ranges" vs. "clinical thresholds" lacks definition — FDA can tighten enforcement on this grey zone without issuing new guidance
• The variant genomic data carve-out is specifically scoped to block genomics AI from using this deregulatory pathway — a deliberate choice that narrows SaMD relief for genomics applications
• The "forthcoming new regulatory framework for AI" (Makary's CES remarks) implies these guidances are a bridge, not a final state — build for the bridge, but plan for the framework

👀 What to watch

• Q2 2026: FDA's forthcoming "new regulatory framework for AI" — Makary's three-pillar framework (deregulate low-risk, post-market monitoring, forward-looking AI guidances) will determine whether the Jan 6 guidance expansion sticks or gets superseded

2️⃣ NVIDIA and Lilly: $1B AI Drug Discovery Lab

TL;DR: NVIDIA and Eli Lilly announced a joint co-innovation lab in South San Francisco with up to $1B invested over five years, creating a continuous learning system that connects wet laboratories and computational dry labs for 24/7 AI-assisted drug discovery.

What happened

• Announced January 12 at the JPMorgan Healthcare Conference
• Joint investment: up to $1 billion over five years
• Location: South San Francisco Bay Area; expected to open end of March 2026
• NVIDIA contributes: BioNeMo platform, Vera Rubin compute, Omniverse and RTX PRO servers for digital twins, Clara open foundation models, physical AI and robotics
• Lilly contributes: 150 years of proprietary biology/chemistry/medicine data, drug R&D expertise, AI supercomputer infrastructure, manufacturing and supply chain expertise
• Scope extends beyond drug discovery to clinical development, manufacturing, medical imaging, and commercial operations
• Model: Continuous learning loop with scientist-in-the-loop framework; 24/7 AI-assisted experimentation
• Jensen Huang distributed DGX Spark systems signed by hand to 12+ AI drug discovery leaders at JPM

📊 Key facts

🔗 Primary source → NVIDIA Newsroom: NVIDIA and Lilly Co-Innovation Lab

🔍 The non-obvious point

The "scientist-in-the-loop" framing is deliberate regulatory positioning, not just workflow design.

• FDA's AI/ML-based SaMD framework requires human oversight for adaptive algorithms; building the lab with scientist-in-the-loop by design embeds regulatory defensibility at the architecture level
• Lilly's 150 years of proprietary data is the structural moat — NVIDIA provides compute and models, but Lilly's data cannot be replicated; the deal value to NVIDIA is access to ground-truth biology data to improve BioNeMo
• This partnership sets a capital threshold for AI drug discovery that most biotechs cannot match — the productivity gap between AI-native and traditional discovery will widen measurably in 2026-2027

👀 What to watch

• End of March 2026: Lab opening in South San Francisco — watch for first published results or preprint outputs as a signal of actual throughput vs. announcement

3️⃣ JPMorgan Healthcare 2026 — M&A and AI Infrastructure

TL;DR: The 44th annual JPM conference confirmed AI as core healthcare infrastructure, while the deal table below reflects January 2025 comparator activity used in coverage — not 2026 conference-native M&A announcements.

What happened

• 44th Annual JPMorgan Healthcare Conference: January 12-15, 2026, San Francisco; 500+ companies, 8,000 attendees
• M&A: J&J announced $14.6B acquisition of Intra-Cellular Therapies (neuropsychiatric); Eli Lilly / $2.5B Scorpion Therapeutics (oncology); GSK / up to $1.15B IDRx (oncology); AbbVie / up to $1.06B Simcere license (blood cancer)
• AI transactions: Hippocratic AI acquired Grove AI (pharma R&D and clinical trial agentic AI); NVIDIA distributed DGX Spark systems to 12+ drug discovery leaders
• NVIDIA expanded BioNeMo platform: nvMolKit (GPU-accelerated cheminformatics), RNA structure prediction models, BioNeMo Recipes for foundation model training/customization
• Note: the M&A table below is historical comparator tape from January 2025 coverage, not a record of 2026 conference announcements.

📊 Key facts

🔗 Primary source → Advisory.com: JPM 2026 Top 5 Takeaways

🔍 The non-obvious point

Hippocratic AI's acquisition of Grove AI is the sleeper move — it bridges patient-facing agents with clinical trial operations AI, creating a vertical stack that spans consumer and enterprise regulated workflows.

• Hippocratic AI has been operating in chronic care and post-discharge follow-up (consumer-adjacent, low-risk); Grove AI operates in pharma R&D and clinical trial operations (enterprise, higher-stakes)
• Combining the two creates an agent company that can serve the full patient journey from consumer to clinical trial participant — a regulatory classification question FDA has not yet answered for agentic AI
• The M&A volume at JPM 2026 signals strategic repositioning before FDA's forthcoming AI framework revalues target categories; acquirers are locking in positions now

👀 What to watch

• Q1 2026: FDA classification guidance for agentic healthcare AI — whether FDA treats patient-facing agents (like Hippocratic AI) as CDS software or consumer wellness will materially affect the Hippocratic/Grove AI integration

4️⃣ FDA TEMPO Pilot — Reimbursement-Linked Digital Health Review

TL;DR: On January 2, FDA began accepting statements of interest for the TEMPO pilot — a limited joint FDA-CMS pathway that coordinates regulatory review and reimbursement eligibility for select digital health devices targeting Medicare/Medicaid.

What happened

• January 2, 2026: FDA began accepting statements of interest for TEMPO pilot participation
• Approximately March 2, 2026: FDA will send follow-up requests to selected potential participants
• Purpose: Joint FDA-CMS initiative to accelerate market access for certain digital health devices targeting Medicare/Medicaid patients while protecting safety
• Structure: Reimbursement eligibility and regulatory review are coordinated for selected participants, not as a blanket rule — a limited pathway linking access and coverage logic
• Significance: First FDA-CMS coordinated pathway for digital health; narrows the standard sequence of 510(k) clearance followed by separate Medicare coverage analysis for selected participants

📊 Key facts

🔗 Primary source → IntuitionLabs: FDA Digital Health Guidance 2026

🔍 The non-obvious point

TEMPO's coordinated FDA-CMS structure is a structural change to the market access pathway, not just a faster version of the existing process.

• Standard pathway: 510(k) clearance (FDA), then separate Medicare coverage analysis (CMS), then contractor local coverage determination — often 2+ years sequentially
• TEMPO: FDA and CMS coordinate for selected participants, narrowing the burden around access and review rather than replacing the clearance process — compressing part of what was a multi-year sequence
• For founders targeting Medicare/Medicaid populations with digital health devices, TEMPO is a high-leverage pathway in 2026; the window for statements of interest is open now

👀 What to watch

• March 2, 2026: FDA follow-up request deadline — which categories of digital health devices receive follow-up will reveal FDA/CMS priority areas and signal where the pilot will generate clearances first

📊 The pattern

FDA opened 2026 by deregulating the front door — removing premarket barriers for single-recommendation CDS, risk-score tools, and non-invasive wearable monitoring — while simultaneously opening TEMPO as a reimbursement-linked fast lane. Capital at JPMorgan is building AI-native infrastructure, and the older deal tape in this brief is serving as historical comparison rather than live 2026 activity. The regulatory and investment cycles are running on the same 12-month convergence: FDA is clearing the pathway just as the tools to traverse it are arriving. Builders who wait for the AI framework to finalize before moving will miss the window.

👀 Watchlist

📎 Sources

Sources of truth

Also consider reading