Weekly Tech+Bio Highlights #57: AstraZeneca Signs $555M AI-CRISPR Pact
Immune tolerance wins 2025 Nobel, AI-driven robotic IVF raises $50M after 18 births, & new out of stealth with a universal patient foundation model
Hi! This is BiopharmaTrend’s weekly newsletter, Where Tech Meets Bio, where we explore technologies, breakthroughs, and cutting-edge companies.
If this newsletter is in your inbox, it’s because you subscribed, or someone thought you might enjoy it. In either case, you can subscribe directly by clicking this button:
🤖 AI x Bio
(AI applications in drug discovery, biotech, and healthcare)
🔹 Patching DNA order screens for the AI era — Microsoft-led team with four DNA synthesis providers red-teamed open-source protein design tools (published in Science), found that AI-redesigned sequences can evade current checks, and deployed software patches that improve sequence-level detection of variants of proteins of concern at order intake.
🔹 Peptide binder design — Researchers at the University of Washington’s Institute for Protein Design unveil RFDiffusion2 for Molecular Interfaces, an all-atom diffusion model that co-designs phospho-specific peptide binders with atomic precision.
🔹 Insilico Medicine maps vast chemical space for NLRP3 — generating ~110 million molecules and over 34k scaffolds around the NLRP3 inflammasome target, integrating generative AI, 3D scaffold extraction, and pharmacophore-aware filtering.
🔹 Xaira posts its 8M-cell, genome-wide Perturb-seq atlas to Hugging Face to speed training/benchmarking of virtual-cell and GRN models. Xaira launched in 2024 with $1B in funding, led by Marc Tessier-Lavigne and co-founded with David Baker to fuse high-throughput biology and AI for drug discovery.
🔹 Adaptyv opens Proteinbase for AI protein design — a public repository of over 1,000 experimentally validated AI-designed proteins, linking sequences, results, and design methods to benchmark model performance and accelerate open, data-driven protein engineering.
🔹 Quantum drug discovery — DARPA backs PolarisQB to commercialize its quantum-AI platform for designing small-molecule drugs, which screens up to 10^30 molecules in hours and has shown success against hard targets like Ebola-related protein interactions.
🔹 BigHat Biosciences launches an AI-driven platform that combines ML with automated wet-lab testing to rapidly design, test, and optimize therapeutic antibodies across hundreds of iterations per week.
🔹 How PLMs think about proteins — Stanford’s InterPLM uses sparse autoencoders plus an LLM agent to surface thousands of interpretable concepts (binding sites, motifs, domains) from protein language models, revealing superposition and scaling effects—then applies them to flag missing database annotations and steer sequence design (Code, Visualization).
🔹 Forecasting lifelong disease trajectories — Nature study describes a generative transformer trained on 400k UK Biobank and validated on 1.9M Danish patients to model 1,000+ conditions, predicting mortality (AUC 0.97) and time-to-next diagnosis from routine history and risk factors; caveats include cohort bias and non-causal associations.
🚜 Market Movers
(News from established pharma and tech giants)
🔹 Bristol Myers Squibb, Takeda, and Astex join AbbVie and J&J in the Federated OpenFold3 Initiative, pooling thousands of protein–ligand structures via a federated platform to train advanced models for predicting molecular binding and accelerating small-molecule discovery.
🔹 Apple researchers publish a preprint introducing SimpleFold, a 3B-parameter protein folding model based solely on standard transformer layers and a generative flow-matching objective, achieving competitive accuracy while removing domain-specific architectural components.
🔹 Illumina launches BioInsight, an AI-driven informatics division integrating genomic, proteomic, and single-cell data to accelerate drug discovery and disease mechanism insights through large-scale multi-omic analysis.
🔹 Olympus launches the Olysense AI platform to assist colonoscopy by automatically detecting precancerous polyps from video data, aiming to improve early colorectal cancer diagnosis and integrate real-time analysis into clinical endoscopy systems.
💰 Money Flows
(Funding rounds, IPOs, and M&A for startups and smaller companies)
🔹 AstraZeneca signs a $555 million milestone-based deal with Algen Biotechnologies to develop AI-guided CRISPR therapies for immune diseases, leveraging Algen’s single-cell datasets and foundation models to map causal gene interactions.
🔹 AI in $1.5B SPAC deal — Singapore’s Nanyang Biologics to go public via merger with RF Acquisition Corp II, leveraging an AI platform to screen natural molecules for drug discovery, with a reported 27% boost in predictive performance over traditional methods.
🔹 Conceivable Life Sciences raises $50M in Series A funding to commercialize its AI-driven robotic platform that automates the full IVF workflow, following early trials producing 18 healthy births and a 120-patient study nearing completion.
🔹 CuspAI raised $100M in Series A funding led by NEA and Temasek to advance its AI foundation models for chemistry, enabling the design of new materials—such as CO2-capturing compounds and sustainable alternatives.
🔹 GHO Capital acquired Scientist.com, an AI-powered platform connecting pharma and biotech companies with pre-vetted research suppliers, aiming to expand its global reach and further automate R&D procurement and compliance.
⚙️ Other Tech
(Innovations across quantum computing, BCIs, gene editing, and more)
🔹 Precision Neuroscience publishes first human study of its minimally invasive brain-computer interface, showing its 1,024-electrode Layer 7 implant can safely record high-resolution neural signals and deliver stimulation without penetrating brain tissue.
🔹 Skin cells to eggs, proof-of-concept — an OHSU-led team used somatic cell nuclear transfer and engineered chromosome halving to create 82 human oocytes from skin cells that were fertilizable and formed embryos.
🚀 A New Kid on the Block
(Emerging startups with a focus on technology)
🔹 Standard Model Biomedicine emerges from stealth with a multimodal AI foundation model (and three papers) integrating genomic, imaging, clinical, and molecular data to support drug development, trial optimization, and clinical decision-making, aiming to reduce costs and timelines across the biomedical R&D lifecycle.
Founded by former leaders from BMS, Siemens Healthineers, and JPMorganChase, the team has published work across radiology, genomics, pathology, and EHR modeling, and is backed by Arkitekt Ventures and Virtue to build the infrastructure layer for next-generation biomedical AI.
🏛️ Bioeconomy & Society
(News on centers, regulatory updates, and broader biotech ecosystem developments)
🔹 Nobel Prize 2025 — Mary E. Brunkow, Fred Ramsdell, and Shimon Sakaguchi receive the Nobel Prize in Physiology or Medicine for discovering regulatory T cells and the Foxp3 gene, revealing how the immune system prevents attacks on the body’s own tissues and paving the way for new autoimmune and cancer therapies.
🔹 White House doubles down on AI for pediatric cancer — pledges $100M to expand the National Cancer Institute’s AI-driven Childhood Cancer Data Initiative.