Xaira Therapeutics Emerges from Stealth with $1B+ Funding, Aims to Redefine Drug Discovery Using AI

On April 23, 2024, local time, the renowned international media outlet Endpoints News reported that Xaira Therapeutics (hereinafter referred to as “Xaira”) had officially launched, emerging from stealth mode. Backed by over $1 billion in financing (approximately RMB 7.2 billion), a group of prominent venture capitalists and scientists are collaborating to establish a novel biotech company that aims to redefine drug discovery and development through the comprehensive application of emerging AI technologies.

According to incomplete statistics from VBInsight, this is the largest financing deal in the global pharmaceutical sector since the beginning of 2024. It ranks third in terms of funding amount within the healthcare and medical field for 2024, surpassed only by the bond financing of dental solutions provider MB2 Dental and the new round of capital raised by publicly listed company BridgeBio. Among startups, it was exceeded only by Altos Labs, an anti-aging company founded by multiple Nobel laureates, which secured $3 billion in funding, surpassing Xaira Therapeutics.

Xaira’s primary investors include ARCH Venture Partners and Foresite Capital, with other investors comprising F-Prime, NEA, Sequoia Capital, Lux Capital, Lightspeed Venture Partners, Menlo Ventures, Two Sigma Ventures, the Parker Institute for Cancer Immunotherapy (PICI), Byers Capital, Rsquared, and SV Angel.

The Transformative Vision of the Star-Studded Founding Team and Genentech Executives

Xaira was founded in May 2023 and remained in stealth mode for an extended period. Xaira was co-founded by two prominent venture capital experts in the biotechnology sector: Bob Nelsen of ARCH Venture Partners and Vik Bajaj of Foresite Labs.

Among the projects Bob Nelsen has been involved in are startups he led investments in, including 47 companies valued at over $1 billion, such as Illumina and Altos Labs, which boasts Nobel laureate backing, as well as a series of high-quality ventures acquired by multinational giants like Pfizer, BeiGene, and Bristol Myers Squibb.

Also worth mentioning is the scientific co-founder, Dr. David Baker, Director of the Institute for Protein Design at the University of Washington, who was once hailed as having the “Hand of God.” He is also an investigator at the Howard Hughes Medical Institute (HHMI) and a member of the U.S. National Academy of Sciences. Dr. Baker’s primary research focus is computational structural biology. His team developed Rosetta, a computational software suite for predicting and simulating the structures and interactions of biological macromolecules and for molecular design. Rosetta is widely used in universities, research institutions, and healthcare companies around the world and enjoys a stellar reputation. Currently, several Ph.D. students from the Baker Lab have joined Xaira Therapeutics full-time. Another co-founder, Dr. Hetu Kamisetty, previously worked at Meta and the Institute for Protein Design.

Furthermore, Xaira’s current leadership team brings a strong Genentech lineage.

First is Xaira’s CEO, Marc Tessier-Lavigne, President of Stanford University and former Executive Vice President and Chief Scientific Officer at the biotechnology giant Genentech. His appointment has brought along a group of friends and former colleagues from Genentech: Dr. Arvind Rajpal, who previously led large-molecule drug discovery at Genentech, and Don Kirkpatrick, who served at Genentech for over 13 years and currently serves as Vice President at Xaira.

According to information already disclosed by Xaira, the company currently has approximately 50 employees in Seattle and California. However, its board of directors alone comprises more than 10 members, all of whom are highly esteemed figures in the regulatory, scientific, and corporate sectors. Prominent professionals such as Dr. Mathai Mammen, former Executive Vice President of Johnson & Johnson’s Pharmaceutical Research and Development; Dr. Carolyn Bertozzi, Nobel Laureate in Chemistry and Professor of Chemistry, Systems Biology, and Radiology at Stanford University; Kaye Foster, Senior Advisor at Boston Consulting Group; Dr. Scott Gottlieb, former Commissioner of the U.S. Food and Drug Administration (FDA) and Partner at New Enterprise Associates (NEA); and Bryan White, MBA Chair of the Advisory Board at the Institute for Protein Design, have joined the company, signaling a strong momentum toward industry innovation.

Integrating Multi-Party Resources to Bet on AI Large Models in Three Core Areas

While still operating in stealth mode, Xaira was registered under the name Orion Medicines. Initially, it aimed to enter the field of antibody drugs, leveraging AI technology to introduce novel R&D concepts. Its development efforts primarily encompass three core areas: advanced machine learning research, large-scale data generation, and robust therapeutic development capabilities.

In the realm of large-model data generation, Xaira has integrated vast amounts of data, with datasets comprehensively describing biology-related characteristics from molecules to the human body and diseases, covering a wide range of scales. Given that Xaira’s team includes leading technologies and personnel cultivated through Illumina’s long-term functional genomics R&D efforts, and has integrated Interline Therapeutics’ top-tier proteomics team, the company is confident in its ability to generate, integrate, and learn from extensive multidimensional datasets.

In the development of innovative therapies, Xaira stated that its assembled team is capable of designing and targeting previously undruggable targets to develop a pipeline of therapeutic products with differentiated characteristics. Furthermore, Xaira’s existing resources enable the standardization of laboratory and clinical results, integrating them into its proprietary data assets; thus, the process of advancing its product pipeline also drives the improvement and development of AI models and foundational biological models.

Xaira’s flagship large AI model integrates RFdiffusion, a large AI model developed by Dr. Baker’s team. Xaira stated that it will use the financing proceeds to bolster research conducted by Dr. Baker’s team.

RFdiffusion is an innovative generative AI system that combines structure prediction networks with generative diffusion models, capable of using deep learning to generate novel proteins distinct from any known naturally occurring proteins. While traditional design methods may require testing thousands of molecules to identify one with the desired performance, the RFdiffusion approach enables research teams to test only a small number of molecules for each design challenge. The model was published in the prestigious journal Nature in July 2023. The associated press release highlighted that this AI model can design biomolecules on demand, holding promise for revolutionizing vaccine and drug development.

Dr. Baker’s team has continued to achieve breakthroughs in AI-driven antibody development, recently reporting another significant advance: by fine-tuning their generative AI system, RFdiffusion, they successfully performed de novo design of single-domain antibodies based on the heavy-chain variable region (VHH) against arbitrarily selected antigen epitopes. The effective binding of these antibodies to their target antigens was validated by cryo-electron microscopy (cryo-EM) structures at atomic resolution.

However, some industry experts have pointed out that there is still room for improvement in Baker team’s RFdiffusion model. According to the paper, only 1% of the generated antibodies achieved the expected efficacy against bacteria and viruses, and those with moderate affinity are difficult to develop into drugs. In response, Xaira stated that RFdiffusion, through continuous intensive training, will rapidly improve over time. Currently, Xaira has established a more comprehensive industrial platform integrating computational and experimental workflows, enabling traditional antibody engineering methods to gradually optimize new therapies generated by large AI models.

Xaira has not yet disclosed its product pipeline, but investors have pointed out that the significance of Xaira’s training of the large AI model RFdiffusion is akin to the advent of gene-editing technology: it will open another door for fields such as oncology and broaden the horizons for therapeutic development across other disease areas. Company leaders believe that in the future, Xaira may also apply AI to other therapeutic modalities, such as small-molecule drugs.

Confronting the Industry’s Two Major Pain Points: “Data + Talent”

Currently, biopharmaceutical companies have widely adopted AI as a routine tool, regardless of whether it is directly applied to drug R&D. Although no new drug developed entirely by AI has yet reached the market, the “AI + new drug” sector has moved beyond the conceptual stage and entered a period of rapid growth.

VCBeat pointed out in its “2022 Industry Research Report on AI-Driven New Drug Development” (hereinafter referred to as the “Report”) that data and talent have become the primary challenges facing the AI-driven new drug development industry.

Among these, data issues represent the biggest barrier limiting the current development of the AI-driven new drug industry. The data in this sector faces dual challenges in both quality and quantity, making self-generated data one of the key solutions to break through this bottleneck.

This further corroborates Xaira Therapeutics’ efforts in large-model data generation and the development of innovative therapies. In particular, by establishing its own integrated wet-and-dry laboratory/smart robotics laboratory, Xaira is poised to create new competitive barriers, especially in real-time tracking and updating of its pipeline product data and in integrating data resources from multiple pharmaceutical companies. The establishment of such an integrated wet-and-dry laboratory/smart robotics laboratory enables the generation of proprietary data, and only with more abundant, high-quality drug R&D data can algorithmic models be effectively optimized and iterated.

To address the significant supply-demand gap for interdisciplinary talent in the AI-driven drug discovery industry, Xaira Therapeutics has attracted and integrated a professional team since its inception. Further developing a cross-disciplinary talent management system may well be the company’s next strategic focus.

The “cold hard truths” distilled on the cover of the Report point out that the pharmaceutical industry remains one of the least efficient sectors and serves as the last bastion resisting technological disruption. If this stronghold can be breached, the moment of true industry disruption will no longer be distant; Xaira Therapeutics may have already taken the first step.