Skip to content
    AI Pharma Pulse · Oct 16, 2026

    Issue 42

    Open on LinkedIn

    Sign in to LinkedIn, then select Subscribe.

    Issue 42 · Week of October 10–16, 2026

    Foundational AI Updates

    OpenAI offers U.S. academic labs free access to GPT-Rosalind for life sciences research

    On October 15, OpenAI introduced the Rosalind Academic Plan, giving eligible U.S. university labs up to 30 Business Standard seats at no cost for up to 12 months, with access to GPT-Rosalind subject to trusted-access approval. GPT-Rosalind is OpenAI’s specialized life sciences model, paired with the Rosalind workbench for tool use across medicinal chemistry, genomics, wet-lab assistance, and related scientific workflows. OpenAI said research data shared under the plan is not used to train models by default, and that API access is not included. Named labs already exploring the model include Zhao Lab at UCSF, the Bourdenx Lab, and Dr. Brian Hie at Stanford University and the Arc Institute. Access requires verifying the institution and principal investigator, reviewing intended use, and naming an institutional safety contact, with monitoring after activation.

    Why it matters: Frontier life-sciences models are moving into academic discovery workflows that feed sponsor pipelines and partnerships. Teams evaluating academic collaborations or internal “trusted access” tools should treat model version, approved use case, and lab-lead accountability as part of the evidence package—not as afterthoughts once a design or analysis shows up in a diligence or IND narrative.

    Source: OpenAI — Bringing GPT-Rosalind to academic research labs

    https://openai.com/index/introducing-the-rosalind-academic-plan

    Google DeepMind shows an agentic workbench narrowing antimicrobial peptide search to wet-lab candidates

    On October 15, Google DeepMind published a technical walkthrough of Google Antigravity, an agentic science workbench that orchestrates database search, ProtNLM function prediction, AlphaFold structural checks, genomic-neighborhood reasoning, and experimental-plan drafting in a single session. In the demo, a researcher steered the agent to mine Streptococcus pseudopneumoniae sequences and surfaced a previously uncharacterized 97–amino-acid Lactococcin 972–family bacteriocin (UniProt A0A0U0D591) with no meaningful Swiss-Prot homolog by BLAST, structural similarity to a known bacteriocin fold, and flanking genes consistent with a bacteriocin system. DeepMind said the lead and several others from the same pipeline are now in wet-lab validation and are not yet experimentally confirmed. The post frames rising antimicrobial resistance as a driver, citing an estimate of 10 million annual deaths globally by 2050 if novel therapeutics are not found.

    Why it matters: Agentic discovery tools are collapsing days of fragmented database triage into auditable, human-steered sessions—but the molecule still has to survive the bench. Discovery, CMC, and biosecurity teams should ask how session logs, model versions (including ProtNLM and AlphaFold outputs), and go/no-go criteria are retained when an AI-prioritized peptide enters synthesis and assay planning.

    Source: Google DeepMind — Accelerating the search for antimicrobial peptides with Google Antigravity

    https://deepmind.google/blog/accelerating-antimicrobial-discovery/

    Drug Development & AI

    AstraZeneca pays insitro $40 million upfront for AI target discovery in MASH and cardiomyopathies

    On October 15, insitro announced a worldwide exclusive research collaboration with AstraZeneca to discover novel targets and therapeutics in metabolic dysfunction-associated steatohepatitis (MASH) and cardiomyopathies. Terms include $40 million upfront, up to about $1.46 billion in potential discovery, development, and commercial milestones, and single-digit royalties on potential global sales. insitro will apply its Virtual Human platform to identify causal disease biology and advance AstraZeneca-nominated targets through development candidate; AstraZeneca will lead clinical development, regulatory work, and commercialization. Daphne Koller, Ph.D., founder and CEO of insitro, and Sharon Barr, AstraZeneca’s Executive Vice President and Head of BioPharmaceuticals R&D, both framed the diseases as heterogeneous under a single label—exactly where high-content human data and AI target selection are meant to help.

    Why it matters: Big Pharma is still writing large, back-loaded checks for AI platforms that claim causal biology, not just faster screening. Sponsors and partners should watch how target-rationale packages from Virtual Human–style systems are documented for governance, diligence, and eventual regulatory questions about why a program moved forward.

    Source: Fierce Biotech — AstraZeneca inks $1.5B insitro deal for AI-enabled target discovery of liver, heart disease drugs

    https://www.fiercebiotech.com/biotech/astrazeneca-inks-15b-insitro-deal-ai-enabled-target-discovery

    Zealand Pharma teams with Generate Biomedicines on AI-engineered proteins for obesity and metabolic disease

    On October 13, Generate Biomedicines and Zealand Pharma announced a multi-target collaboration to design multispecific protein therapeutics for obesity and other metabolic diseases using The Generate Platform, which pairs generative AI with high-throughput experimental validation. Generate will lead research through lead-candidate selection; Zealand will handle clinical development, regulatory submissions, and worldwide commercialization. For two initial programs, Generate receives $30 million upfront and up to $35 million in early preclinical milestones, and is eligible for total milestones of nearly $1.2 billion plus tiered royalties, with an option to add a third program. Zealand CSO Utpal Singh cast the deal as expanding beyond peptides into proteins; Generate CEO Mike Nally said the teams will first define what a great medicine looks like for each program, then design candidates to that profile.

    Why it matters: Metabolic and obesity developers are buying AI-native modality expansion—proteins designed to a predefined product profile—while keeping clinical and regulatory ownership in-house. Expect partner diligence to probe how design specifications, experimental validation loops, and candidate-selection criteria are versioned before first-in-human packages are written.

    Source: BioSpace — Zealand pens up to $1.2B deal with Generate for AI-designed obesity, metabolic drugs

    https://www.biospace.com/deals/zealand-pens-up-to-1-2b-deal-with-generate-for-ai-designed-obesity-metabolic-drugs

    Regulatory, Policy & Industry Moves

    FDA launches a Rare Disease Framework Initiative aimed at fit-for-purpose evidence in small populations

    On October 14, FDA announced the Rare Disease Framework Initiative to build a holistic, fit-for-purpose approach to rare-disease drug and biologic development, outlined in a new white paper. Acting Commissioner Kyle Diamantas, J.D., and Rare Disease Innovation Hub Director Amy Comstock Rick, J.D., said the goal is greater consistency and predictability while maintaining rigorous standards, noting that only about 5% of rare conditions have an FDA-approved treatment. Commitments include earlier patient voice, external expertise, risk-based CMC and pre-approval inspection approaches, and novel statistical methods. A hybrid public meeting is set for October 28 at White Oak with the Reagan-Udall Foundation; a Federal Register notice on October 16 opened a request for information and comments. Possible later outputs include revised guidance, regulation changes, staff training, and organizational shifts.

    Why it matters: Rare-disease sponsors live or die on what evidence is realistically attainable. If FDA hardens “fit-for-purpose” into clearer expectations for endpoints, natural history, CMC flexibility, and patient engagement, protocol design, medical writing, and inspection readiness plans need to track the framework as it moves from white paper to operable review practice.

    Source: FDA — FDA Launches the Rare Disease Framework Initiative to Develop a Holistic, Fit-For-Purpose Framework for Rare Disease Drug Development

    https://www.fda.gov/news-events/press-announcements/fda-launches-rare-disease-framework-initiative-develop-holistic-fit-purpose-framework-rare-disease

    Moderna’s Noubar Afeyan urges regulation of commercialized AI agents, invoking thalidomide

    On October 13, at the Bloomberg New Economy Forum in New Delhi, Moderna co-founder and chairman Noubar Afeyan called for careful regulation of AI agents being commercialized and transmitted beyond a company’s own walls, distinguishing large language models (“capability”) from agents (“intentionality”). He compared the need to prove agent safety before market release to drug regulation after thalidomide caused birth defects. Coverage noted agents’ promise for accelerating drug discovery alongside risks of dangerous compounds or unreliable results, and placed his comments alongside broader industry debate on autonomous systems. Afeyan also cautioned that replacing clinical trials with digital simulations remains a distant prospect given the complexity of human biology.

    Why it matters: When a major biotech founder frames agent deployment in the same breath as drug safety failures, governance conversations inside R&D organizations gain outside pressure. Clinical development, pharmacovigilance, and digital teams using agentic tools for protocol drafting, data review, or compound design should assume auditors will ask who authorized the agent, what it was allowed to do, and how unsafe outputs were caught before they left the house.

    Source: Quartz — Moderna co-founder Noubar Afeyan calls for AI agent regulation

    https://qz.com/moderna-cofounder-noubar-afeyan-ai-agent-regulation-101326

    Originally published on LinkedIn. Read the original

    The Pulses

    Stay close to what’s changing.

    AI Vendor Pulse and AI Pharma Pulse, published on LinkedIn.

    About 1,070 newsletter subscribers · October 2026

    AI Vendor Pulse

    Mondays

    Open on LinkedIn

    AI Pharma Pulse

    Fridays

    Open on LinkedIn

    To subscribe, sign in to LinkedIn and select Subscribe on the newsletter page.