Most small molecule drugs and biologics miss a large group of important receptors in the body. These receptors, known as GPCRs, play big roles in major diseases, but they are unusually hard to target in a precise and safe way. Skape Bio is trying to change that by using artificial intelligence to design an emerging class of therapies known as mini proteins that can fit these receptors snugly and control them in ways traditional drugs often can’t. Christoffer Norn, co‑founder and CEO of Skape Bio, discusses how mini proteins share strengths of both small molecules and biologics while avoiding some of their weaknesses, why the company’s approach could open up new therapeutic opportunities, and how Skape Bio is building a broad pipeline of differentiated GPCR‑targeting therapies.

AI’s promise in drug discovery is being held back by a simple but stubborn problem: the field doesn’t have enough of the right kind of lab-generated data to train good models. Public databases and scattered studies give a patchy view of how proteins interact, mostly showing successful interactions and using different methods and conditions, which makes it hard for AI systems to learn broad rules or understand what doesn’t work. A-Alpha Bio tackles this data gap with AlphaSeq, a lab platform that can test around a million protein pairs in one experiment under the same conditions, producing rich, consistent information on both hits and misses that’s well suited for AI. David Younger, co-founder and CEO of A-Alpha Bio, talks about why far more lab data and infrastructure will be needed than most people expect, why current public datasets fall short, and why the company’s business model is built around providing data and services rather than developing its own drugs.

Multi-specific antibodies promise to unlock complex biology that conventional monoclonals can’t touch, but their added mechanisms of action also introduce safety and developability risks. These antibodies—especially T‑cell engagers—behave differently from traditional monospecific antibodies, and seemingly minor architectural tweaks can cause disproportionate shifts in potency, selectivity, and cytokine release. LabGenius is trying to turn multi-specific design from an intuition-driven art into a genuine engineering discipline by generating proprietary data at scale and feeding them back into machine learning models. Angus Sinclair, chief scientific officer of LabGenius, discusses why many safety failures in early solid-tumor T‑cell engagers were effectively locked in at design, how the company’s AI platform engineers multi-specific T‑cell engagers that are both potent and selective in solid tumors, and where AI is actually adding value in multi-specific design today.

Acute ischemic stroke is both ubiquitous and undertreated. Only a small fraction of patients currently receive clot-busting drugs or mechanical thrombectomy because of the small treatment window. That’s because existing therapies require rapid presentation to specialized centers and carry nonreversible bleeding risks that make clinicians hesitant to use them. Basking Biosciences is developing a first-in-class, reversible thrombolytic that targets von Willebrand factor to expand access to safe, effective treatment. Basking Biosciences CEO Julia Owens and co-founder and chief scientific officer Shahid Nimjee discuss the tension between restoring blood flow and causing irreversible intracranial hemorrhage that clinicians face in treating ischemic stroke, the narrow treatment window for existing therapies, and how modulating von Willebrand factor may open a much larger treatment window across a broader range of care settings.

Plasma gelsolin is an abundant, endogenous regulator of inflammation that is consumed during severe inflammatory insults. When levels fall too low, patients are at higher risk of organ damage and death, particularly in settings like acute respiratory distress syndrome where a dysregulated inflammatory response floods the lungs with fluid and leaves patients dependent on ventilatory support with no approved therapies today. BioAegis Therapeutics is working to turn recombinant human plasma gelsolin into a pipeline-in-a-product. Susan Levinson, CEO of BioAegis, discusses recombinant human plasma gelsolin as a potential first-in-class treatment for ARDS and other inflammasome-driven conditions, how it modulates cytokine storms without suppressing the immune system, and its potential in other conditions including neurodegenerative diseases.

A chronic shortage of high‑quality targets remains one of the biggest constraints in drug discovery, even as therapeutic tools become more powerful and diverse. Regeneron is tackling that problem with its Regeneron Genetics Center, which has built a genetics‑driven discovery engine that integrates human genetics with rich clinical data, large‑scale proteomics, and AI‑driven analytics. Aris Baras, head of the Regeneron Genetics Center, discusses how proteomics is reshaping RGC’s view of risk prediction, how AI helps his team sift through hundreds of millions of variants, and what it really takes to scale this kind of effort and translate it into more successful, transformative therapies for patients.

Epigenetics, the layer of chemical switches that controls how genes are turned on and off, can act like cancer’s operating system when a single epigenetic enzyme becomes essential for a tumor to survive. K36 Therapeutics is developing first‑in‑class medicines that block an epigenetic enzyme that helps certain multiple myeloma cells grow, change identity to escape treatment, and become resistant to today’s drugs. Terry Connolly, CEO of K36, discusses a new way to fight cancer by changing how cancer cells read their DNA instead of chasing one mutation at a time, how K36’s experimental therapies aim to re‑sensitize tumors to existing treatments, and the potential to create new options for people whose cancers have stopped responding.

Most marketed antibodies work as antagonists, simply shutting off a receptor, even though many immune, metabolic, and cancer pathways require more nuanced control. Metaphore Biotechnologies' function‑first platform combines live-cell experiments with machine learning to read how receptors and binding partners behave in living systems, distill those complex dynamics into the key functional features, and then design functional antibodies that agonize, bias, or multi-target pathways from the outset. Metaphore CEO Angela Hwang discusses how the company is using this approach to open up difficult or previously undruggable targets, generate medicines with better profiles, and give drug developers greater control over complex signaling than traditional drug development approaches allow.

Cells continuously sense their environment and in response to stressors, adapt, recover, or die. Soley Therapeutics uses its AI platform to capture thousands of intracellular features and map how cells sense, interpret, and respond to stress. The approach gives Soley the ability to pursue previously undruggable targets. It has generated more than 10 novel oncology programs in less than two years and advanced two first‑in‑class experimental small molecule therapies toward the clinic. Yerem Yeghiazarians, co-founder and CEO of Soley, discusses the science behind the company’s first-in-class lead candidate, the applicability of Soley’s platform to a broad set of diseases, and the capital efficiency of the company’s approach to drug development.

Promising drugs can become abandoned or underused because of tolerability issues, poor drug‑like properties, or other fixable limitations, even when there is already compelling human evidence that they work. PureTech Health starts with an unmet need and human pharmacology, then systematically dissects and solves the specific liabilities of discontinued drugs to unlock breakthroughs in an approach that has proved to be a highly efficient means of value creation. Eric Elenko, president and co‑founder of PureTech, discusses the company’s disciplined approach to drug innovation around rescuing promising but discontinued therapeutics, its hub‑and‑spoke structure, and how this model can neutralize emotional bias, enforce clear success criteria, and turn partially derisked assets into commercial successes.