Moderna’s Founder on Innovation That Breaks Through

ALISON BEARD: I’m Alison Beard, and this is the HBR IdeaCast.

Today, we’re bringing you the third in our series of interviews with CEOs of major companies that we recorded during our recent Future of Business event. Noubar Afeyan is the founder and CEO of Flagship Pioneering and co-founder and chair of Moderna, the mRNA vaccine company. He thinks deeply about how man and machine can work together to solve some of the world’s largest problems, whether they’re in the world of science, business, or society.

Now, we’ve spoken to him on IdeaCast before, back in 2021, to talk specifically about leadership and innovation lessons learned in the development of Moderna’s COVID vaccine, but things move fast in today’s business world, and his thinking has evolved accordingly. He’s a trained biochemical engineer, entrepreneur, and executive who has helped spawn more than 100 life science and technology startups. He has a unique vantage point as to what really works when it comes to effective innovation.

I want to ask first, how do you define breakthrough innovation, not just in science, but across industries?

NOUBAR AFEYAN: I think that most innovation focuses on adjacencies and is viewed to be continuous innovation, which means that what you’re working on is an iteration over what’s been worked on before. And once in a while, you have discontinuous innovation where it’s not easy to connect what’s being proposed with what was done before. And that’s much harder or much less predictable, but often is categorized as a breakthrough because people neither expected it, nor could they in the beginning assess its value.

So I think this is a concept that applies to really any industry. The question is, can you predict breakthrough innovations or can you follow a process to make them happen?

ALISON BEARD: And given how fast the business environment is changing nowadays, do you think that all companies need to be in search of breakthroughs or is incremental innovation okay for many of them?

NOUBAR AFEYAN: I think you need to think about both and how they compliment each other. I actually think about breakthrough innovation as something that you really want to think from the future back to the present, whereas the more continuous innovation, adjacency oriented innovation should be thought of as from the present going forward. And I think you want to attack the space from both directions, both from the needs that you’re aware of today and ways you could become more efficient, more impactful, and deliver better value to your customers. But at the same time, you want to envision future states without necessarily knowing how we get there from here.

And if you find something compelling, come backwards and say, “What would have to be true and what can I do today to make that future possible sooner?” And the reason you want to do that, and you asked, is this something people really should be focused on? Is that I think that the disadvantage of adjacency oriented innovation only is that it can quickly become commoditized. Everybody spends money looking for opportunities in close by places, and it has the advantage of being reasonably easy to assess risk reward. You can get validation from experts who can tell you which ones are more likely than not to work.

All of those are reasons to be doing things close by, the problem is everybody else is doing the same thing. And so how you can create a sustainable advantage in those close by places is hard to fathom. Whereas if you’re willing to leap to new possibilities, new realities, at least you’ll have your initial loneliness rewarded potentially with owning the space.

ALISON BEARD: What advice would you give other organizational leaders, even outside of tech or life sciences, about how to structure for this kind of innovation that you’re talking about? Do you set up a separate division for it? Do you assign a key person in every area to think about? Is it about creating a culture where everyone’s on the lookout for breakthroughs?

NOUBAR AFEYAN: Yeah. I wouldn’t have everyone look out for breakthroughs. One, it’s not sustainable and possible, and two, it’ll be pretty disruptive to how they do their jobs on a day-to-day basis. The question is what ratio of your attention and future expected value do you see coming from the kind of continuous improvement, if you can call it that, versus the discontinuous, disruptive leaps?

And I think that there are ways one could think about it. We of course do it for a living so it’s a little bit less of a structure issue. That’s all we do. If I was a company that was leading in a space, but also didn’t want to get surprised with unexpected entrants into a space, I would likely have a group that is somewhat differently motivated, composed, rewarded than is the traditional R&D organization. That’s by the way, sometimes hard because everybody in R&D wants to believe that they’re doing far out things, but in reality, the definition of a far out thing is something that’s unpredictable and it’s very hard to get money in R&D budgets for truly unpredictable things.

So you do need to think about it as having some constraints. You want to make sure that you have enough different avenues being pursued and options being pursued because this is a space that I would propose is more about uncertainty than risk. Uncertainty being about unknown probabilities, risk being about known probabilities, whether high or low. This is the domain of uncertainty, and it’s really hard to hedge mitigate uncertainty because there’s no signal there. What you need to do is to take multiple approaches and persist long enough to see if you can make them real, and if they can, to protect yourself so you can create a new value pool. That notion, which we can all theorize about, can be practically set up and industrialized. And that’s kind of what we have experienced doing over the last 25 years.

ALISON BEARD: With AI advancing so rapidly, are you finding new uses for it to augment human intelligence in your innovation process at Flagship? I’ve heard you talk about something called polyintelligence.

NOUBAR AFEYAN: Yeah. I mean, so we started applying AI in its most recent forms back in 2018 to what we do. You would expect that since we’re fairly fearless about trying things that are not yet ready and mature, et cetera, et cetera, it would make sense that we would attempt with the earliest of even notions of being able to generate, for example, novelty that given that it’s in our DNA, I guess, we’ve been doing that for quite a while. And we’ve learned both the possibilities and the limitations, but also become quite humble about the unpredictability of the major leaps that have happened even since we got deeply engaged in this.

I’ve called it oftentimes artificial intelligence as augmented intelligence. I’ve called it augmented imagination, which it has the ability to do. One person’s hallucination is another person’s kind of leap. And so applying it into novel spaces is a very interesting way to prospect for things that a given human may not be able to think of. I mean, you can ask for, and we do quite routinely hundreds and hundreds of new ideas in a space.

And then the question becomes, how do you analyze for them and see what might or might not make sense to try out? I’ve even called it sometimes alien intelligence. It’s really something that we have never had available to us. It feels unusual. It feels other than from this world because we’ve never had the ability to simultaneously come up with thousands. In fact, I’ll tell you millions of new ideas.

I’ll give you an example. Some 25 years ago, the very first company Flagship formed, unbeknownst to most to think we’re only in biotech is actually a company that used evolutionary algorithms online to create new products and then used panels also online to evaluate them in a way creating an evolutionary loop to create novelty and optimize for better solutions. This was 2001, 2002.

Roll the clock forward. That company went on for a decade. It was eventually merged with Nielsen. Today, we’ve actually gone back to explore the space, now armed with generative AI, armed with agent-based systems. And just recently we announced the formation of a project called Extuitive, which is essentially trying to do the same except with massive advances in capability. And literally watching 150,000 agents representing humans interacting with millions of objects is quite a scene to behold and you can do that with AI. So I actually think it’s almost unimaginable that you could compete without deploying these significant augmentation capabilities in doing innovation and breakthroughs.

ALISON BEARD: It sounds like in that description, you talked about both of the steps in breakthrough innovation that you wrote about in your HBR article with Gary Pisano. The first is variance generation that is considering different areas of research, asking what if questions based on what you learn. Taking a step back though, in that first step, how do you decide where to look when these are things that even experts in the field haven’t really considered before as with the mRNA technology?

NOUBAR AFEYAN: Experts in the field are very, very good at understanding the limitations of variation and asserting what exists today and what’s likely to be possible next, but they’re particularly not advantaged at speculating on what will be available next, next, next, because they will likely not be the experts in that. What makes somebody an expert is a ownership of the best ideas today, and those are not usually the next ones to come along.

It’s a rare expert who is humble enough to allow for the possibility that they’re wrong, which is the essence of an unexpected disruptive innovation. So I think that what you’ve got to do is really think about what you’re doing as a combination of leaping. Leaping meaning, not limiting yourself to slight variations of where you are, and then realizing that it’s going to be hard to know what will and won’t create value, but you can try.

You can essentially engage with people who can represent feedback that represents value, whether that’s eventual customers, users, you name it, wherever you can get selection pressure. And then do the iteration cycles. I mean, it’s clear in nature that variation selection and iteration creates unexpected novelty, namely life and all the living things that we have. And likewise, in markets, we see that all the time.

I mean, the influence of one mobile phone onto another, one running shoe onto 100,000 others, cars, and all of those markets have been evolving much like it happens in nature, except in this case, the purchasing kind of power and that feedback and markets is what’s driving what the generators and the designers are doing.

So I think all of that can be represented in digital markets for these ideas and we are very much leading in trying to demonstrate that in lots of different fields, taking our knowledge from ecosystems and ecology and nature into the realm of inanimate objects, designed objects, and that’s an interesting leap.

Now you asked a little earlier about polyintelligence. So let me just say what we mean by that. There’s a fascination today, of course, even though it’s 75 years after AI was initially described and kind of envisioned with the notion of how humans are making room and making space for artificial intelligence. But what’s interesting to me is that that for the first time has forced humans to allow for the possibility that they’re not the only intelligent thing on the planet. So far, that’s how we defined it because we controlled language and we said intelligence means humans.

Now that machines can be intelligent, at least according to some, I would conjecture that the next simple step will be to say, why can’t nature be intelligent and why can’t we treat all living objects, whether it’s cells in your liver or plant cells or viruses or entire ecosystems as intelligence?

And by intelligence, I mean systems that are able to adapt very rapidly, able to anticipate, able to do a lot of the processing we associate with neural networks or computational neural networks today. And I believe, I wrote an essay on this early in this year, and it’s very much driving some of our thinking now, that when we start thinking of nature as a series of intelligences, if there is such a plural word, then we will start understanding what’s happening in biology very differently.

And the payoff for that is that I think what we’re going to find is what we’ve learned so far about biology in the reductionist, deconstructivist way is a tiny fraction of what we’re going to understand about disease, about where we get food and how these systems are affected by infections, et cetera. All of that rendered onto a platform of thinking of intelligence systems is what’s ahead in my view.

So the interesting dilemma, of course, will be that we will have gone as humans from having one system we don’t understand, nature, to now two systems we don’t understand, nature and AI models that involve a billion plus parameters. And it’ll be interesting to see how humans adapt to that reality. I don’t consider that a major threat because humans have been living without understanding the nature around us for a long, long time. And we’re adding a piece, frankly, to our nature, it’s just another system that is extremely complex and that we as humans can hopefully harness positively, not negatively in order to impact some of the biggest problems in the world.

ALISON BEARD: Yeah. Harness and learn from, because your piece with Gary Pisano basically argues that the approach to experimentation can mimic natural selection and evolution. But explain in more detail how you have learned from nature in that way to sort of shift Flagships process or create Flagships process to experimentation and that in such that it differs from other organizations, how they might approach it.

NOUBAR AFEYAN: There’s a bunch of ways in which we use the basic ideas of Darwinian evolution in how we operate. Let me just reiterate first that we think large value pools exist beyond adjacencies of the here and now. And we think those value pools are largely unattainable based on current financial or financing approaches because you can’t fit them onto a risk-reward paradigm because it’s very hard to assess the risk because you don’t know what the probability of success of something that’s never been thought of or done before is, and therefore it’s also hard to get meaningful feedback on what the reward would be.

But if you are comfortable prosecuting uncertainty, and I would argue that nature is essentially uncertainty in that whatever hasn’t evolved yet, hasn’t evolved yet, you could sit around and make bets on what’s likely going to be the condition that wins out in any given competitive situation in nature but I think that’s been proven a difficult thing to do.

So I would say that we are essentially replicating the notion that variation, selection and iteration where you also have this important component of descendancy. That is inheriting the solutions that prevail into the next generation and the next generation of what could be a set of game generations in the context of computational systems or in nature, it’s literally generations of cells that are able to pass on the positive solutions to the next.

In fact, I increasingly view all that nature is doing is learning from the surrounding of it each cell at a time, each species at a time, each ecosystem at a time. So I do think there’s a lot, not just metaphorically to learn, but also operationally.

I’ll give you an example. People may not realize that in a way our bodies experience, for example, RNA naturally coming from outside in the form of viruses. And so a lot of what a virus is doing in coming into the body, getting into cells and essentially replicating itself, but for the replication step, that’s what our vaccines do. We get them in the body, we get them into the right kind of cells, they get taken up, and then the RNA actually takes over the protein production to make a particular protein. In our case, it’s a protein that’s immunogenic and therefore causes an immune response. But much of that intricate molecular pathway is also used by viruses to actually now in their case, replicate themselves.

So there’s a lot of things anywhere you look, I’m sure that a lot of people who deal with cybersecurity and how to protect computer systems from being infected by computer viruses, there’s a lot to be learned and has been learned from how our immune systems do this and everywhere you look, I think you’re going to find more and more examples of nature’s intelligence and how it can inspire and inform humans or machine intelligence.

ALISON BEARD: For leaders and organizations that want to pursue this kind of breakthrough innovation that you’re talking about, how long can they expect the process to take? And then how do you justify the time and cost when it is an experiment? There isn’t guarantee that the work will even yield a breakthrough or be able to move from that sort of experimental phase to successful execution.

NOUBAR AFEYAN: That’s a very important question and it’s very, very tricky and it’s very case specific industry specific. For example, I’ll say as a simplification that if you have a very high profit margin industry, you can afford to do a lot of this because the consequence of being right is a massive blockbuster. That’s why it’s so prevalent in the pharmaceutical industry.

But in a cutthroat industry where you have very low margins, the added uncertainty, let me use the word differently, of relying on breakthrough innovation makes it almost impossible. And so you have to kind of surrender to the unpredictable external attack that might completely destroy your franchise, but nevertheless, there’s not much you could do to protect yourself. I’m reminded of the Innovator’s Dilemma that Clay Christensen needs to talk about in the sense that sometimes people who are highly cognizant of the fact that they may be under threat, nevertheless, don’t actually act on it because it comes across as a very different kind of threat that it ultimately evolves into.

It’s very hard to predict. I also would say that you need to be somewhat opportunistic in doing this in not falling into the trap of saying every innovation has to be a solution to a problem because often what we find is that if you search for the problem and the solution at the same time, you might be able to matchmake where there’s new value pools in a way that if you insisted on owning the solution space or the problem space to be your point of departure, you will miss out on all things that essentially would be matched in the journey.

So it’s the highest form of kind of a scary proposition to say, “I’m going to look for value in a particular market that is far from what’s being done today. I’m going to be guided by what exists, but also a systematic way of exploring beyond adjacencies. I’m going to let large crowds of customers or experts, non-experts guide the way I evaluate the terrain, but then I’m going to iterate, iterate until I get some hits and when I get some hits, I’m going to try it out. I’m going to prototype it. I’m not going to assume that I could deductively reason my way into a breakthrough.”

That’s a fool’s errand. I think that most people who make breakthroughs describe them as though they were the result of genius. And my experience, and maybe I’m just particularly not that, is that’s got very little too genius. It’s got to do with opportunism, value orientation and the propensity to iterate.

And sometimes, by the way, recognizing the thing right in front of your nose that’s been there for a while, but you choose to ignore it because it doesn’t fit the current paradigm. All of that is what leads to success versus this kind of hard work and reasoning and experts telling you what to do.

ALISON BEARD: It sounds like open-mindedness is a key element there too, sort of knowing that you’re trying to tackle one problem, but then being open to solving another, if that’s where your research leads you.

NOUBAR AFEYAN: Open-minded and open-ended.

ALISON BEARD: Yeah.

NOUBAR AFEYAN: That’s another thing that people have called it. Evolution is entirely an open-ended innovation paradigm. And if you’re open-ended about it and open-minded to see value, I think you’ve got the right combination.

ALISON BEARD: Great. Well, I want to get to some of our audience questions. Chris Bremman says that sometimes he or she feels like there’s a significant disconnect in organizations between driving breakthrough innovation and the commercialization of that innovation. Chris has multiple entrepreneurship programs in the company that’s intrapreneurship and they’re trying to bridge that gap, but is that really the best way to drive the kind of innovation that you’re talking about to get to that commercialization of it, to get to the end consumer, or are there better ways to organize for success?

NOUBAR AFEYAN: I mean, look, so I’ve spent the last 38 years being involved in what traditionally would be called entrepreneurship, but the last 25 of it, I’ve created a company whose whole life is to conceive and create companies. And I’ve increasingly viewed it as intrapreneurship because within Flagship, we launch six to eight companies a year.

And unlike a dominant player in a space where intrapreneurship usually means a small little piece of the business where you’re trying to create new products and new markets, for us, it’s the whole business, but it’s still intra in the sense that we’re not individual people, small teams raising money, writing business plans. We don’t do any of that.

I think that kind of corporate venture creation as a way to innovate where you align the practitioner’s rewards with the kind of output they generate, which is typically not the case in a traditional R&D organization, where you protect people from working on unreasonable things instead of insisting that they describe why, what they’re doing is reasonable, and therefore, in my view, quite uninteresting often because the reasonableness filter forces people to say things that are likely going to work and therefore aren’t going to be disruptive.

So there are ways to create the right motivation, but entrepreneurship, in my view, allows for that, provided you don’t kind of force it to be entirely commensurate with everything else in the R&D organization, which in my view is prospecting for different value pools in different ways and therefore should be rewarded in a very different way.

I ran companies in the past where the sales folks who worked in my companies often made a lot more money than I did. And as a CEO, and I had no problem with that. Other people think that there should be some hierarchy of pay. I think compensation should be variable depending on what kind of risk people are taking or what kind of value they’re building.

ALISON BEARD: I do think it’s going to transform education. Let me just end with a question about the current political environment, particularly in the United States. We’re seeing extreme anti-science rhetoric and cutbacks in research funding. You are also an immigrant to the United States and we’ve seen anti-immigrant stances coming out of our policy, current policy environment. So I’d love to hear just your thoughts on how that’s affecting scientific and technological progress in the country, sort of those two twin trends that we’re seeing.

NOUBAR AFEYAN: Well, that’s above my pay grade but I will venture to say as a personal level that both science and immigration are ways that clearly threaten the status quo. It’s about change, it’s about progress. I think that the United States, I immigrated here because I consider it a country of immigrants and I think I’m right. There is nobody who’s listening to this from the United States who isn’t the descendant of an immigrant, just a question of when you cut it off.

And so I think the major advantage of this country is its regenerative nature, much like nature that we’ve been talking about all along, which is progressing and learning from the iteration loops that it goes through, this country has been a hyper evolutionary, hyper adaptive place because of the constant influx of new thoughts, new risk taking appetite, new levels of desperation of people coming here and trying to make it work, which I think has made for a great society.

And I think it will continue that way because that advantage is hard to replace by any other natural resource, in my opinion. There’s other countries that have more natural resources than us, there’s other countries that have more people than us, but I think this special place we found ourselves where we welcome uncertainty in science and knowledge and we embrace it and we recognize that at any given time we may be wrong about what we thought we were certain about, but ultimately over time we’ll get it right.

The thing about going after science today, which is troublesome, is that these are things that were fought over in various centuries old debates around what constitutes truth, what constitutes fact, and what the scientific method can do to essentially excavate the unknown in a responsible way. But if you can essentially replace scientific fact with opinion or offer opinion as though it was scientific fact, that really blurs the lines.

And I think that we need to, as a society, kind of recognize that the consequence of that are worse medicines, more dangers that we face in the decisions we’re making. If you’re going to design flying objects and you can’t really rely on established science, I don’t see this stopping. I’ve told people pretty openly, I don’t see this stopping with mRNA or vaccines. It’ll affect all the medicine and all of anything new that raises a question about, “Well, how sure are we that the science is real?”

And the answer is as sure as this large scientific establishment using a longstanding process of hypothesis falsification as the arbiter of truth at any given moment. And I’m pretty sure that ultimately the benefits of the centuries of experience we have in doing that, coupled with the centuries this country’s had of immigrants flowing into this soup, if you will, of this melting pot of capability and creativity, I think those two things are worth fighting for.

ALISON BEARD: Noubar, thank you so much. I really appreciate you taking the time today.

NOUBAR AFEYAN: Thank you for having me.

ALISON BEARD: That’s Noubar Afeyan, founder and CEO of Flagship Pioneering and co-founder and chair of Moderna. Check in next Thursday for another future of business episode.

If you found this episode helpful, share it with a colleague, and be sure to subscribe and rate IdeaCast in Apple Podcasts, Spotify, or wherever you listen. If you want to help leaders move the world forward, please consider subscribing to Harvard Business Review. You’ll get access to the HBR mobile app, the weekly exclusive insider newsletter, and unlimited access to HBR online. Just head to hbr.org/subscribe.

Thanks to our team, senior producer Mary Dooe, audio product manager, Ian Fox, and senior production specialist, Rob Eckhardt. And thanks to you for listening to the HBR IdeaCast. We’ll be back with a new episode on Tuesday. I’m Alison Beard.