Few leaders in biotechnology have successfully willed an entirely new class of medicine into existence, but Oak’s newest advisor, John Maraganore, did exactly that. As the founding CEO of Alnylam Pharmaceuticals, he spent nearly two decades translating the raw biological discovery of RNA interference from a risky lab concept into a multi-billion-dollar global commercial reality. Along the way, he pioneered a corporate blueprint for the modern biotech era.
Today, Maraganore is one of the life science industry’s most influential architects and mentors.
Q: When it comes to pharma services and biotech startups right now, what are some of the most exciting trends you are seeing, and is there anything that is getting a little too much attention?
A: One of the most exciting things happening across our industry – both on the services side and the pharma/biotech side – is the integration of AI. Because nobody really knows exactly how far or how big it will ultimately go, it can sometimes feel like overhyping. But it is very real. It is already having a meaningful impact on drug discovery and development, and it is here to stay.
To me, it feels a bit like the early days of the web – business practices and ways of working are changing, and we can discover things more quickly. What’s important right now is understanding what is what: who is using it for real, practical applications, and who is just mentioning it because they think it's a popular buzzword.
Q: In drug development, everyone seems to be moving at a breakneck speed, yet this is an industry where you historically cannot rush the process or sacrifice safety. How do you see AI playing a role there?
A: There are certain things we can do quicker without losing quality or sacrificing safety. At the early stages of drug discovery, there are already clear applications. The most notable is in antibody engineering, where AI has enabled the design of antibodies much more rapidly than traditional approaches.
Companies exemplify this in different ways. Some, like Chai Discovery, work collaboratively with pharma and biotech on a platform+service model. Others, like Generate Biomedicines, use a mixed model, collaborating with big pharma while also developing their own internal pipelines.
Beyond discovery, we are seeing impactful applications of AI in patient identification and clinical trial enrollment – identifying specific participants and activating trial sites with high patient volumes to speed up the process. We also see it in manufacturing, where models use data parameters such as temperature, oxygen consumption, and media conditions in a fermenter to optimize cell line productivity and achieve higher yields of a biologic product. AI is helping across the value chain.
Q: To step back for a moment, for those less versed in the space, what is the ultimate benefit of using AI to design antibodies more rapidly?
A: It means speed; you can get to your drug candidate much more rapidly. In other cases, there are antibody targets that have historically been incredibly challenging–whether that means finding high-affinity antibodies toward a specific target or being able to discriminate between isoforms of the same target where you want to inhibit one but not the other. These engineering tools enable the refinement of an antibody drug, which can significantly improve the quality of the development opportunity.
Q: Beyond manufacturing and discovery, you also mentioned AI being deployed in documentation and review. Does the acceleration of these cycles raise any concerns about compromising the quality or efficacy of the final drugs?
A: It will not compromise quality. In fact, it’s going to improve it. It is a little bit like the question people ask about autonomous vehicles. Are these self-driving cars going to be more dangerous than human drivers? It turns out that Waymos are proving to be much safer than taxi or rideshare drivers because they aren’t prone to human error. Even the FDA has discussed using AI in its drug review process to accelerate the approval cycle for new medicines.
Q: You have previously spoken about needing a "Polar Express" level of faith – believing in a technology even when the rest of the market is deeply skeptical. In today's cautious environment, how can you tell the difference between a visionary who is ahead of their time and someone who is just blowing smoke?
A: It can be hard sometimes. The Polar Express analogy comes from the children's story where only those who truly believe can hear the jingle bell. There are times when the investment community and stakeholders write off a technology and refuse to invest further.
I experienced this firsthand in 2010 when we were developing RNAi (RNA interference) as a whole new class of medicines. The world essentially gave up on it. Even when we showed people amazing data, they just couldn't fathom that it was real or believe it would ever work. They turned out to be very wrong, but we had to show them the way.
As an entrepreneur, when you are faced with that obstacle, you can't run away and hide under a blanket. You have to get out in front of it and give people a reason to believe by showing real, meaningful data – clinical data for therapeutics, or disruptive workflow data for a platform. That is the entrepreneur's obligation when facing doubt.
Q: Do you think AI can help an entrepreneur in that specific situation?
A: I do, because it can help accelerate aspects of discovery to get that data faster. But it still requires patience, persistence, and steadfastness – things I don't think AI itself can give you. AI can be a tool to advance the business, but the core persistence has to come from the founder.
Q: Does having a hands-on CEO who is actively in the labs with the scientists provide a tangible competitive advantage?
A: Without a doubt. In a startup environment, you are building the plane as you fly it. A startup CEO doesn't have the luxury of sitting up in the executive suite, acting as a mere process leader. You need to be a content leader.
Whether you are on the cutting edge of a new platform or forging a new medical strategy to treat disease, you have to be on the field scrumming with your team, rather than looking down from the owner's box thinking other people are going to do the work for you. That simply does not work in this industry.