- CEO Marc Tessier-Lavigne
- Total raised More than $1 billion in committed capital
- Headquarters South San Francisco, CA
David Baker’s Seattle laboratory has become the place for the biopharma industry to push the limits of artificial intelligence in protein design. Xaira Therapeutics will see if some of his lab’s AI breakthroughs are ready to actually transform drug R&D.
While Xaira’s story began at Baker’s Institute for Protein Design, the startup is thinking far beyond a single AI model. Its A-list investor group has committed over $1 billion, while also hiring former Genentech CSO Marc Tessier-Lavigne as CEO to lead the effort.
It has some of the biggest names in biopharma VC backing it as well. ARCH Venture Partners’ Bob Nelsen and Foresite Labs’ Vik Bajaj led the launch, helping fill out a board that includes former FDA head Scott Gottlieb, the Nobel Prize-winning chemist Carolyn Bertozzi, and former Johnson & Johnson CEO Alex Gorsky.
But big names and a billion dollars don’t equal scientific validation. Xaira’s launch this spring, as reported by Endpoints News, has only poured gasoline on the already-raging debate over how much of biotech’s excitement for AI is hype, and how much is real.
Xaira has roughly doubled in headcount since this spring, and is now approaching 100 employees. Its two Bay Area sites will combine into a permanent South San Francisco headquarters later this year, Tessier-Lavigne said, while its Seattle site will carry on with its computational focus.
Xaira hopes to use AI across the drug R&D process, focusing on identifying targets, making molecules and designing the right trials with the right patients. Tessier-Lavigne said making molecules is the area most ripe for disruption in the short term, as Xaira is already using some AI models from the Baker lab to develop its first wave of drugs.
For now, these models — like RFdiffusion and RFantibody — are good enough to generate antibodies that are initial hit candidates, which are then improved with “old-school methods,” Tessier-Lavigne said. If the models keep getting better, he expects they’ll produce molecules that look increasingly like lead candidates, and eventually optimized lead candidates.
“We’ve already made a lot of progress compared to what’s been published in terms of the power of our models,” he said. “We see that accelerating also over time.”
Tessier-Lavigne declined to divulge initial targets or a timeline to the clinic, but said Xaira’s focus is on targets “poorly served” by today’s drugs and antibody screening methods, and moving faster, and with a better rate of success, than industry standards.
Its Bay Area labs are generating volumes of biological data from experiments using CRISPR to activate or knock out genes. It plans to train future models with these perturbation data that, in the long run, could identify new drug targets. Xaira expects to soon exceed the publicly available data on these kinds of perturbations, Tessier-Lavigne said.
“What will really revolutionize things is, not only if we can make drugs using AI, but if we can use AI to improve the success rate in choosing the right targets and in choosing the right patients,” he said. “The medium- and longer-term opportunity will be in target elucidation and patient stratification.”
Tessier-Lavigne declined to disclose any more specifics about the company’s funding, beyond saying that the $1 billion-plus in committed capital will be drawn in tranches. The goal of the large raise was to not return to capital markets until Xaira has a “rich pipeline” progressing in the clinic, he said.
“The plane has achieved liftoff,” he said. “The wheels are off the ground, but now we’re climbing with a lot of work to get to cruising altitude.”
Key backers: ARCH Venture Partners, Foresite Labs, F-Prime Capital, NEA, Sequoia, Lux Capital, Lightspeed Venture Partners, Menlo Ventures, Two Sigma Ventures, and SV Angel
Find the full list of 2024 Endpoints 11 winners here.