BioAsset

Insights / All articles

Research

BioAsset VC Dive: Xaira Therapeutics — Can AI Crack the Drug Discovery Code?

By Alay Galvankar, Lead Analyst & Vice President at BioAsset · 5 slides

Xaira Therapeutics launched with more than $1B in capital, one of the largest debuts in biotech history, to take on a core problem in healthcare. Despite advances in science, drug discovery remains expensive, slow and highly unpredictable. Many promising drugs fail late in development because biology is complex and often poorly understood.

Founded in 2024, Xaira took an unorthodox approach. Most companies today layer AI onto traditional biotech workflows. Xaira was built as AI-native from the ground up, bringing together leaders in biology, AI and medicine to create a unified platform where computation and experimentation continuously inform each other.

The leadership reflects that mix, with former Genentech executives and Nobel Prize-winning scientists. CEO Marc Tessier-Lavigne is a former Genentech executive and Stanford professor, and director Carolyn Bertozzi is a Nobel laureate and Stanford professor. The company is backed by top-tier venture firms including ARCH, Sequoia and NEA, and it has unicorn status. It is headquartered in San Francisco, with more locations in Seattle and London.

Making biology computable

By combining large-scale biological data with computational models, Xaira can simulate how cells, proteins and diseases behave. That allows the company to predict outcomes more precisely than traditional trial-and-error methods, and to keep improving its models through a feedback loop with experimental data. In other words, Xaira is making biology “computable” to design better medicines faster.

One of its most innovative capabilities is generative protein design. Using advanced AI models, Xaira can create new proteins and antibodies that conventional methods cannot reach, unlocking potential therapies for previously “undruggable” diseases. The platform builds on foundational science from the Institute for Protein Design at the University of Washington, which lets Xaira rapidly translate computational predictions into real-world experiments. With the help of Oracle Cloud Infrastructure, the company is also able to train large AI models.

The models behind the platform

X-Atlas/Orion is the largest Perturb-seq dataset and the largest single-cell dataset publicly available, at around 8 million cells. It covers all human protein-coding genes and powers AI that predicts cell behavior and disease biology. That gives Xaira and the broader scientific community an unprecedented foundation for predictive models.

Generative protein design uses AI models to create novel proteins and antibodies beyond natural possibilities, and it integrates prediction with experimental validation. It moves drug discovery from trial and error to engineered design, opening entirely new therapeutic possibilities.

Virtual cell modeling, on the AI-native platform, uses AI and large datasets to simulate cellular responses digitally. It predicts how genes, proteins or drugs behave before lab tests and supports end-to-end drug discovery, from target to molecule to validation. That dramatically speeds up discovery, reduces costs and increases success rates.

The outlook

Xaira is still a relatively young company, and the majority of its products and prototypes are in the developmental stage. But as funding increases, leadership strengthens and the technology advances, this company could be the next big thing. Releasing X-Atlas/Orion publicly also builds goodwill with the scientific community and positions Xaira as a data infrastructure leader, not just a drug company.

There is a bearish case too. Even with its rapid growth and optimistic forecast, Xaira has 189 employees, offices in San Francisco, Seattle and London, and zero products on the market. Funding of $1B+ depletes fast in biotech. Drug discovery is a graveyard for well-funded early-stage companies, so keep an eye on any future successes or failures in clinical trials.

The market around it is expanding. Experts project a 38.90% compound annual growth rate (CAGR) for the AI in healthcare market from 2026 to 2032. Grand View Research’s chart of that market, which is split into software solutions, hardware and services, shows it at $36.7 billion in 2025 and $50.7 billion in 2026, reaching $505.6 billion in 2033.

This article is the text version of a BioAsset slide post. See the original slides on bioasset.info.