Bending the Curve of Drug Discovery and Patient Care: A Conversation with Paulo Fontoura, Chief Medical Officer of Xaira Therapeutics

In 2024, Xaira Therapeutics was launched with a groundbreaking $1B capital raise. The company's generative AI capabilities span from developing fundamental computational methods to applying these techniques in biological discovery, molecule design, and clinical development. We recently sat down with Dr. Paulo Fontoura, an industry veteran and the newly appointed Chief Medical Officer of Xaira, to discuss what gets him excited at Xaira, and his vision for the future of drug discovery and development.

Paulo, congratulations on the launch of Xaira! What drew you to Xaira?

Paulo Fontoura: Thank you, it's great to be here. As many others, I firmly believe that AI is poised to revolutionize drug discovery, drug design, and the overall process of identifying new treatments for diseases and influencing healthcare. The speed of change in this revolution is accelerating exponentially, and I'm eager to be at the forefront of this revolution and really push the envelope.

What drew me to Xaira is its unique blend of talent—a world-class team that includes leaders like Marc Tessier-Lavigne and Hetu Kamisetty, innovators from David Baker's lab and Foresite Labs, alongside seasoned pharma veterans such as Debbie Law. This fusion of expertise perfectly reflects the company’s ambitious vision and makes the opportunity to harness AI for drug discovery incredibly compelling.

With many companies leveraging AI across many fronts, how does Xaira differentiate its approach?

Paulo Fontoura: At Xaira, we leverage generative AI to tackle three fundamental challenges in drug discovery and development. First, we focus on molecule design—developing antibodies and binders for historically difficult targets—and our technology is already yielding promising early results.

The other two things that we want to differentiate ourselves by are in building models. We are building generative AI models of biology by harnessing large-scale RNA-seq and Perturb-seq data to uncover new insights into disease pathophysiology. In addition, by integrating robust patient datasets with advanced analytical tools, we aim to determine why drugs work for specific patients, ensuring the right drug is matched to the right patient. That is the dream for many to advance precision medicine, but it has been very hard to make progress.

Xaira's recent $1 billion capital raise clearly reflects the strength of the company’s platform and vision. It must be exciting to pursue strategies with such strong backing. 

Paulo Fontoura: Indeed, this investment not only underscores the excitement surrounding AI but also validates the bold vision behind our company. With this significant capital, we have the freedom to pursue ambitious projects—not only in drug development but also in building generative AI platforms to tackle grand challenges in biology and patient selection. This funding empowers us to be both creative and deliberate in our approach.

Now that we have a very large blank canvas to play with, our strategy is to make sure every step we take is purposeful. When selecting new drug targets, we ensure that each candidate is truly developable, with a compelling target product profile and clear value for patients and society. We're going to do things that are important not just in terms of demonstrating the power of this technology, but also really creating real breakthroughs in medicine—whether by developing transformative drugs or creating AI models that offer deeper insights into new targets, endotypes, and patient groups.

Talking about patients, which disease areas stand to benefit most from AI-driven innovations?

Paulo Fontoura: At Xaira, we intentionally maintain a therapeutic-agnostic approach, focusing first on areas where our technology can achieve rapid breakthroughs. If we identify a target in a disease that has been notoriously challenging for drug development and we believe we can make a meaningful impact, we'll pursue it. Every candidate we select must add real value—a transformational breakthrough rather than merely a demonstration project. Our aim is to discover and design new targets that create substantial value for both patients and society.

When it comes to modeling biology and patient responses, we plan to be selective initially, concentrating on one or two therapeutic areas where we can see the quickest progress. Personally, I believe common chronic diseases—with significant patient heterogeneity and variable responses to standard therapies—offer some of the most promising opportunities for our generative AI models. That said, being selective at the outset doesn't preclude us from expanding later. As a new company, we are committed to deploying our technology, resources, and strategy in the most effective way possible, continuously learning and adapting as new opportunities arise.

Although AI holds great promise, we have yet to realize its full potential into meaningful therapies. What are the key hurdles in your view?

Paulo Fontoura: Numerous companies are using AI, and initial failures are a natural part of the learning curve rather than a bad sign. One major challenge lies in the concrete deliverables of these models. While applications such as optimizing existing molecules are relatively straightforward, the real value of AI emerges when it creates fundamentally new therapeutics and discovers novel targets. Despite tremendous advances in protein folding and drug design, no entirely new molecules developed through these methods has yet progressed through clinical development, so the technology still has to prove itself. However, progress often appears slow at first until a quantum leap accelerates the process. It is not surprising that we have not yet seen a complete clinical success, as this is a normal phase in the evolution of new technology.

Additional hurdles exist in the clinical development process, which is highly regulated and involves health authorities, payers, and patient groups worldwide. AI has the potential to transform various aspects of drug development—from documentation and safety reporting to regulatory interactions and communication strategies—but adapting to these changes will take time. I believe the greatest challenge—and opportunity—lies in not only designing new drug molecules but also in pioneering innovative approaches to drug development. This aligns with the vision of precision medicine: establishing a clear chain of evidence linking the right target, reliable biomarkers, and effective surrogate endpoints can accelerate development dramatically, transforming the landscape for many chronic diseases.

What are you most enthusiastic about the field in the next five to ten years?

Paulo Fontoura: Ideally in the coming years, we would significantly improve the efficiency of the R&D process by addressing one of its biggest cost drivers—late-stage failures. We kill drugs too late because the data is insufficient. If we can change that, we could halt projects with much greater confidence and at a much earlier stage. And if we achieve this, the traditional timeline and expense of drug development—taking 10 to 15 years and nearly $3 billion per drug—could be reduced dramatically. That would be a fantastic revolution and everybody would benefit.

I do anticipate rapid and, at times, chaotic progress. Some approaches will succeed while others may fail, and we will learn valuable lessons from both outcomes. Generative AI offers a quantum leap in our understanding of biology, patient selection, and drug design. If it fulfills its promise, every stage of drug development—from target identification and lead generation to candidate selection and clinical development—stands to be fundamentally transformed. Companies at the forefront will generate critical knowledge, and Xaira is determined to be one of those pioneers. Our goal is to push the envelope, learn quickly, and iterate rapidly to continuously evolve and deliver greater value.

With this AI-driven revolution coming, how should the broader industry anticipate changes and adapt?

Paulo Fontoura: This revolution is not merely about adding a new tool to our toolbox—it represents a fundamentally new way of operating. We are approaching an era where advanced, super-agentic AI will serve as cognitive assistants, working alongside humans to accomplish tasks beyond our natural capabilities. In our highly regulated industry, where patient health, safety, and sustainability are paramount, this shift will require new strategies and a readiness to adapt. For these advancements to be fully realized, it is essential to foster an inclusive dialogue that goes beyond pharmaceutical and technology companies. Engaging regulators, payers, healthcare providers, patients, and caregivers will be key to shaping a future that meets everyone's needs. This is an exciting time, and I am hopeful that, together, we will make significant progress soon.

Thank you, Paulo. As we wrap up, do you have any final thoughts for our audience?

Paulo Fontoura: For me, the next decade offers the same opportunity in terms of revolutionizing pharmaceutical R&D as when we expanded therapeutic modalities. All of a sudden, we had biological drugs, gene therapies, RNAi and antisense technologies. Those abilities to manipulate biology at a fundamental level really open new fields of medicine and deliver drugs for many, many patients.

Generative AI offers much more than rapid drug design and iteration; it provides unprecedented insights into biology, disease mechanisms, and patient treatment. This technology could represent a quantum leap in R&D efficiency. Although the transition may be chaotic and marked by some failures, I am confident that by the end of the decade, we will be in a fundamentally different place. Given the pace of change, the level of investment, and the emerging talent, this revolution is inevitable—a true bending of the curve.