Industry landscape · October 2026
A short reading of the market: what large pharmaceutical companies are buying, which technical approaches are gaining ground and where capital is going. Every figure is attributed and linked to its source.
Market signals
Industry trackers count 114 AI/ML drug-discovery partnerships signed in 2025 (84 in 2024) and about $45.9 billion in headline value in the first half of 2026, of which roughly $1.9 billion was committed upfront.
Source: IntuitionLabs deal trackerTakeda’s collaboration with Iambic Therapeutics is valued at more than $1.7 billion, covering oncology and gastrointestinal targets.
Source: IntuitionLabs deal trackerBoltz-2, released openly by MIT and Recursion, predicts binding affinity at accuracy approaching free-energy calculations while running far faster. Recursion researchers then published a pipeline that feeds it into absolute binding free-energy methods.
Source: BioPharmaTrend; Bio-IT WorldGSK committed $50 million upfront to license Noetik’s cancer foundation models, trained on roughly 40 million cells from over 1,000 patient tumours to predict spatial gene expression.
Source: BioPharmaTrendLarge pharma is moving to longer-term arrangements that integrate data, models and workflows. Merck signed a $1 billion multi-year agreement with Google Cloud, and Lilly is building an AI supercomputer with NVIDIA.
Source: IntuitionLabs deal trackerIn May 2026 Incyte announced a collaboration to deploy Edison Scientific’s Kosmos AI scientist across its discovery and development work.
Source: IntuitionLabsGinkgo Bioworks now sells curated biological datasets designed for pre-training and fine-tuning AI models.
Source: IntuitionLabsAI antibody discovery startup Earendil Labs announced an $787 million private round in March 2026, and Astellas exercised its option on an AI-designed gene-delivery capsid from Dyno Therapeutics in April 2026.
Source: IntuitionLabs; Business WireBiotecnika reports a rise in computational biotech startups in several Indian metros, backed by Indian and global investors.
Source: Biotecnika, March 2026What it means for us
Fast learned models for exploration, rigorous simulation for the decisions that cost money.
Every ranking arrives with its method, its uncertainty and code a partner can rerun.
Pipelines, data engineering and models that plug into a partner’s systems.
Computational designs only matter once an assay agrees, so we plan the wet-lab loop from day one.
Next step
