At a Glance
- Company: Xaira Therapeutics
- Location: US (stealth origins; ARCH/Foresite incubated)
- Funding: $1B committed at launch (2024)
- Focus area: AI/ML for biology, protein design, data generation
- Next steps: Pipeline advancement across modalities
What happened?
Xaira Therapeutics has publicly detailed its AI-driven drug discovery platform on its website, marking a shift from stealth after securing $1 billion in committed funding in April 2024. The platform integrates advanced machine learning, expansive data generation and therapeutic development to predict targets, design molecules and select patients.
Why does it matter?
Xaira’s approach aims to engineer medicines for traditionally undruggable targets by combining generative AI from co-founder David Baker’s lab with massive functional genomics/proteomics datasets. This end-to-end AI could shorten timelines, cut costs and boost success rates in a field where empirical methods dominate, positioning Xaira among AI biotech leaders.
Who’s behind it?
Co-founded by Nobel laureate David Baker (protein design), incubated by ARCH Venture Partners and Foresite Labs, led by CEO Marc Tessier-Lavigne (ex-Genentech CSO, Stanford president). Team advances RFdiffusion/RFantibody models for novel proteins/antibodies.
How was it achieved?
Core elements include foundational AI models for protein design, a data platform generating multidimensional datasets (molecular to human scales) and product development workflows. Experiments feed back into model improvement, creating a virtuous cycle for causal disease insights and therapy design.
What’s next?
Xaira will advance an AI-generated pipeline across modalities, focusing on hard targets; public Perturb-seq datasets (e.g., X-Atlas/Orion) already released to accelerate field-wide progress. Expect first clinical candidates and further model unveilings.
What are people saying?
Marc Tessier-Lavigne: “AI will transform every step… two-, three-fold improvements in speed and success rates.” Bo Wang: “Scale and quality needed for virtual cell models.”