Pod Insights a16z Podcast a16z Podcast | On Corporate Venturing & Setting Up 'Innovation Outposts' About this episode Every big technological shift (per Carlota Perez) brings with a structural shift too — an “institutional adjustment” in how companies innovated and build new products, according to Steve Blank and Evangelos Simoudis. Large organizations used to (and continue to) set up remote R&D labs in places like Silicon Valley. But now, those companies are also investing more energy and resources in setting up corporate venturing arms and/or “innovation outposts” in such startup ecosystems — especially as they believe that startup-driven innovation is one of the best ways to keep up with and address disruption in their industries.
But… it’s not enough to simply establish a presence in these places; how do you also “sense” and respond to the right opportunities? Are they in the right places? Does beginning with corporate venturing really work for such outposts? And finally, how can these orgs avoid just acting out “innovation theater”?
Simoudis — who has also written about whether “the elephant can dance again” using the case of IBM and Watson/ AI — offers his views on how big companies can and should use the Valley (and other innovation clusters) in this episode of the a16z Podcast.
Transcript Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:00 Sonal? Hi, everyone. Welcome to the A6 and Z podcast. I'm Sonal and today, Michael and I are speaking to Evangelos Simoudis, who is currently a VC, but started his career at Caltech, where he was one of the original researchers of machine learning and has some of the very early patents on machine learning. Then he became a software engineer and he went to Digital Equipment Corporation and then to IBM. And basically, the reason we're having him on the podcast today, though, is that he's recently co-authored a series of posts that focus on the topic of corporate innovation, which we think is really interesting, given that so much innovation, especially lately, is somewhat startup-led.
0:37 Michael Copeland? Yeah, those two words don't necessarily always go together.
0:40 Speaker 02 They don't and they've tried. This is not the first time that this is being tried, but I want to say that there is a lot more seriousness and introspection this time around.
0:51 Sonal? Oh, really? Seriousness and introspection this time around? What do you mean?
0:54 Speaker 02 So I think that corporates, both based on the lessons they've learned in the late 90s, which was their last foray into the valley and other startup ecosystems, as well as the disruption that they're starting to feel today from startups, they are trying to, they're thinking a lot more carefully, or they want to think more carefully of what is the right way to approach ecosystems such as the valleys. And of course, this time, the world is a lot more complex. We have more than one ecosystem that matters. We have many more technologies that matter. But again, they're coming into it, it's the ones that are serious about it
1:38 Speaker 02 with a lot more, more serious intentions.
1:41 Sonal? Right. Well, one of the trends that you and Steve Blank point out is that there's been a recent wave of companies setting up, quote, "innovation outposts" in Silicon Valley. Amazon has a lab in Silicon Valley. The DARPA now has a thing in, I mean, they've always partnered Silicon Valley and actually funded much of the early Silicon Valley, for sure. But there's now a formal organization. It seems like every major company in the world has an innovation outpost in Silicon Valley and elsewhere. One thing that's interesting is you guys actually cautioned that some of this can become sort of innovation theater. And I'm curious to hear the difference between. How do you know it's just not innovation theater? And it's going to lead to something fruitful? It's not just like another failed siloed function?
2:23 Speaker 02 That's a very good question. And what we are seeing is that a lot more companies want to establish a presence. And most of them interpret this presence with the establishment of a venture effort. But that does not make them to be truly embeddable into the into the Valley's ecosystem and put them in a position to really take advantage of everything that this area has to offer and by extension, other ecosystems, startup ecosystems around the world. So what we try to caution is just because you're sending some executives here as a
3:08 Speaker 02 corporation, it doesn't mean that you're automatically positioning yourself to to take advantage of the of our of Silicon Valley. And maybe even more importantly, just because you're establishing a fund, it doesn't mean that now you're going to start seeing a tremendous deal flow and the right deal flow and participate actively and obviously benefit the the the corporation that you represent.
3:37 Michael Copeland? You'll start seeing companies, but they may not be the right companies. And they may not be the companies that actually come around later and and take you on as a corporate entity, right?
3:46 Speaker 02 That's right. I mean, it is very interesting to me when I meet a corporate delegation that has recently established itself here. And they will tell me how they participate in a couple of meetings and they interpret that as now being part of the part of the network. And a little bit later, they will ask exactly the question that that you asked, why aren't we getting into the right deals? Why can't we talk and why can't we syndicate with the the right VCs that?
4:17 Sonal? And what's what's the answer to that? Why can't they?
4:20 Speaker 02 Well, first of all, in our business, there is the aspect of trust. You need to develop. But more importantly, I think you need to, as a corporation, you need to demonstrate that you're, what you're trying to accomplish. Just because you have a venture fund and just because you're willing to spend some money doesn't mean that you become a value-add partner of a startup or of a group of startups. So what we have found is that the ones that have a particular reason, they're trying to solve a particular problem, they have engagement from the highest levels of the corporation, and it's not only an ointment, but there is actually engagement
5:03 Speaker 02 on a continuous basis. They're the ones who are putting the right foundation to be able to truly take advantage of a startup ecosystem.
5:16 Sonal? So that's some of the things that reveal what would make an innovation outpost work. A broader question I have is just how big companies in general operate compared to startups and VCs as well, if we're talking about this innovation ecosystem. And one thing that I found recently, which I included in our newsletter, is a finding that decision-making in companies has actually significantly slowed down. I thought that was fascinating because one of the observations that you guys have shared is that sometimes the speed of response is even more important than the response itself. And I was wondering if you can share some more insight about why that is.
5:51 Speaker 02 Well, one of the observations we made is that you need to understand what, the corporation needs to understand why they want to have an outpost. Most of the times, there is some sense of sensing, being able to sense what is happening in a particular ecosystem so that they can respond.
6:12 Sonal? It's kind of like an earthquake, like, what is that?
6:14 Speaker 02 Right, a Richter scale, yeah. You have sensors out there, yeah. Very much so. I mean, Caltech, which is a world-class institute on earthquake analysis, geophysical analysis, they run an entire network of sensors up and down the West Coast, right? You should be thinking as your outpost is providing you with that sensing ability. An early warning system or something, yeah.
6:38 Sonal? Not the day before.
6:39 Speaker 02 Yeah, again, some corporations, or most corporations, feel that just because they establish a presence... In Silicon Valley, they're done.
6:50 Michael Copeland? Right.
6:51 Speaker 02 And one of the advice that we provide is that first determine which are the ecosystems that where relevant work for the problems that you're trying to address is being done. So if relevant work is being done in Silicon Valley and in Israel and Berlin, well, you should consider having presence in each one of these ecosystems. And maybe the presence is not of the same size, but nonetheless, every one of these presences provides you with signal. And then based on that sensing, you can determine what is going to be the nature of your response.
7:34 Speaker 02 So we have identified five different ways of responding. You can invest, you can invent, you can acquire, you can partner, and you can incubate. And determining which of these types of response is the right one to start with and the right one for the particular ecosystem that you've chosen to be present in is actually a hard problem. Most of the times when I'm asked to advise corporations, I come in the middle of the movie, as I call it, because they have already taken an action. Ninety percent of the times is to establish a venture group.
8:18 Speaker 02 And in retrospect, we determine that maybe establishing a venture group wouldn't have been the right first step. Maybe they should have established a business development capability before they even think about establishing a venture group.
8:34 Michael Copeland? There was this wave of corporate venture in the late 90s, and it ended badly for the most part. Now, we're back to 2016, or last year for that matter, and they understand differently about what technology means for their business. But it sounds like if they're just coming back with a similar approach of like, well, we'll just set up a venture arm, it's not going to work this time either. So what do they understand differently? How is it different? And are they here to stay?
9:04 Speaker 02 Well, so let's start with some numbers. So back in the late 90s, roughly between 98 and 2000.
9:15 Speaker 02 There were roughly 500 corporations that had established venture groups in our area. And of those 500 by 2003, 50% had closed down. Now, by, I will call them unofficial statistics, there are roughly a little over 1,000 corporations, which in a sense makes sense because many more geographies have opened up in the intervening, you know, 10, 15 years. And I wouldn't be surprised if we have the same drop-off rate when the next recession arrives and when the next crisis or whatever arrives. Now, there is the one important difference.
9:58 Speaker 02 Back in 98, 99, I feel that corporations had approached the internet broadly as a communication medium, communication collaboration. This time around, not only they are seeing the broader impact of digital technologies, but more importantly, there are several fundamental technologies beyond information technology that are hitting them at the same time. So I was working earlier in 2015 with a foreign corporation, very large corporation,
10:42 Speaker 02 and their space is materials, but they were here to understand software and big data. Then by extension, Internet of Things, and by extension, software as a service models.
10:55 Michael Copeland? Is it because they wanted to become a software company? I mean, they're a material science company? Or is it because they feel like software might really start to get into their side of the business somehow?
11:05 Speaker 02 No, because actually they were seeing that software and data and 3D printing had already gotten into their business. And so they were not starting being ahead of the ball. They were, they knew that they were starting from behind. But my point is that they, they had already realized that they're, they're being hit.
11:28 Speaker 02 Or they're being threatened, maybe, from a variety of perspectives. And this is not about just learning about one technology, checking it and moving on with life in a business as usual way.
11:39 Michael Copeland? Right.
11:40 Speaker 02 So now they're seeing that there are many more things that are, many more technologies, many more business models.
11:47 Sonal? Right, like software as a service is actually a completely different mindset.
11:50 Speaker 02 Precisely. And by the way, their customers now are asking them to start using this type of models as opposed to them thinking, oh, a startup is using it now. Let me think about whether it applies to my business as well. I mean, sometimes their customers are ahead of them and they have to react.
12:11 Sonal? That's actually kind of a sad place to be, though, when you're having to do that direction versus leading your customers where you want them to go.
12:17 Speaker 02 I've been in this space and this industry long enough to have seen all of these ups and downs. And what is fascinating to me today, without trying to create any hype, is that what we are facing, what we're going through right now, is truly a technological revolution, not unlike what happened back in the Steam age.
12:44 Sonal? Right, or the industrial revolution.
12:46 Speaker 02 Or the industrial revolution, that magnitude. You know, the first internet wave was also very important. It was mostly a single thing here. We're being impacted by many more technologies, many of which are acting as building blocks, many of which weave together with business models. So the level of complexity that corporate executives have to contend with right now is unprecedented. And their pace of work. Is does not prepare them very well for having to make this multidimensional decision.
13:24 Sonal? Right, nor does the structure, because first of all, all these things are coming together as building blocks that might interact in completely unexpected ways. So there's that interaction effect. But there's also this thing where one of the theses that we've put forth is that it's actually very difficult for a non-software as a service company to actually become a software as a service company, because it requires completely re-architecting everything from the ground up, from your sales organization to the way you're departmentalized, to just about everything, customer service, like, sales, the mindsets.
13:53 Michael Copeland? Compensation, yeah.
13:54 Sonal? Right, exactly. And so.
13:55 Speaker 02 Incentives.
13:56 Sonal? Right, and so the question that I have is when you describe this as a big structural shift, what are the ways that over the past hundred years to now, that companies have had to change their organizational structure? And where might we go next? Like, how is the nature of the firm being redefined? I always love this topic.
14:13 Speaker 02 So we have some hypotheses, and one of the major hypotheses that we have is that the innovation outpost, if done correctly, can become a major change agent for the corporate structure. How this type of organizations can start impacting the way you do research, the way that you are changing from being an Horizon One corporation to being an Horizon Two and Horizon Three corporation, how you need to to change your core without throwing away all your DNA.
14:56 Speaker 02 I mean, so these are very, very important and very critical problems for corporations to start thinking about. We do not believe that everybody will be able to make that leap. And frankly, some of the statistics show that the life of the corporation is shrinking, right? You have companies that are that remain independent or for shorter time.
15:23 Michael Copeland? Yeah, the span of the the number of companies in the S and P 500, or their longevity is shrinking quickly.
15:29 Sonal? I mean, the company you worked at was is probably one of the very few hundred-year-old companies left, IBM, and there might be very few of them left, it's true. And actually, it begs the question, because it's almost like this assumption that a company should have a long lifespan, and it's actually like, well, do we really want that to be the case? If companies are no longer the containers of security and people can actually move around.
15:48 Speaker 02 And you see it, I mean, first of all, you see it in the tenure of CEOs, which is shrinking significantly. But just what you what you mentioned, I mean, you see it in the attitudes of employees where the aspiration of working for the same corporation for 30 years is is no longer valid. You, I won't call it, we have moved to the gig economy, but we're moving in that direction, sometimes by choice, sometimes by necessity. But but I think that these are all trends that that are going to to impact how how the corporation looks and which corporation survives and the like.
16:34 Michael Copeland? Does the outpost ever have to become the center?
16:37 Sonal? Can it even become the center? Because I just don't think there's ever been a case where that's happened.
16:42 Speaker 02 So I don't think it becomes a center, but but it becomes the transformation agent. The whole idea is that corporations do not experiment enough, in our view. What can a corporation take away from Silicon Valley or from any other innovation ecosystem? It's not the fact that you come into touch with two or three or 10 startups. It's the fact that there is a set of now best practices and one of these that you can take away, best practices of how to create startups, how to finance startups, how to manage a portfolio of startups. And what is important about this
17:25 Speaker 02 is that the corporations need to find ways to experiment rapidly and inexpensively. And and what we have learned here to do very well. Maybe better than any other ecosystem around the world is how to quickly start companies, to get them financed, to allow them to pivot as necessary, how to back our winners and let our not so strong players move on. And I think the corporation needs that type of a playbook. So the innovation outpost becomes that change agent for launching these experiments.
18:09 Speaker 02 I think a number of these experiments become very successful. So essentially, you need to find ways and what we are advocating is how do you change the core?
18:20 Michael Copeland? Are there specific examples of that that you can point to that you've either observed or if you can, worked with?
18:27 Speaker 02 For me, a very interesting example is what Google is doing with Alphabet. And the reason I find this interesting is because you're seeing their venture capital groups working in tandem with the rest of the corporation. You see them sometimes acquiring a company and letting it develop before bringing it into the core. You're seeing them merge things, like you saw, for example, what they did with their AI efforts. You see how they deal with moonshots like the Google car. I'm not saying that they always get it right, but there are some interesting lessons there of how do you continue evolving the corporate
19:11 Speaker 02 structure. Another example is BMW's iBrand, an organization that has been set up very differently from the major business units of BMW, trying to look at not only the manufacturing aspect, the design and manufacturing aspect of the car, which is what BMW and other automakers are known for. But the the whole aspect of mobility services, right? How do I? How do I service a consumer beyond just owning a car? And how do I service that consumer throughout the the life cycle of the of a transportation experience?
19:55 Speaker 02 Not only when that person goes to a to a dealer. So and there are other there are other examples. But these, to me, are some more recent experiments that point in that direction.
20:07 Sonal? You have an article out with another co author about evaluating moonshots in Harvard Business Review. Can you tell us what that framework looks like?
20:14 Speaker 02 The moonshot is a is a major effort by a corporation with the idea of of addressing a big or a big problem. So in the case of of IBM, which was the company that we evaluated through our framework, they they are developing Watson as a as a new type of software platform which uses artificial intelligence and can be used as the as the building block for building, for creating the next generation of applications, whether they are enterprise applications or or consumer applications. And they're investing significant sums of money in order to see them through.
20:56 Speaker 02 They need to consider a variety of of dimensions. And one of the dimension that that we talked about is experimenting with business models before they they identify the the right one. Through this experimentation, analyze their failures as much as they analyze their successes, because they can learn a lot from these failures. And one of the issues that many corporations have and their employees is that they become risk averse because they do not want to fail. And when you're, our point is that when you are involved in moonshots, not everything is going to go as planned.
21:40 Speaker 02 And if it goes as planned, it means that you didn't take enough risks. So they really need to impart that type of approach to their employees.
21:49 Michael Copeland? Is the moonshot, is it a quantitative thing or a qualitative thing? Which is to say that, you know, what's a moonshot for Google may not be a moonshot for a century old tire company or something like that.
22:00 Speaker 02 Well, I mean, for us, the notion of the moonshot was more defined, an effort to address something new by making a very significant investment. So just to put into perspective, if I'm a large corporation and through my venture arm, I invest in a, in a small startup, maybe I'll invest five, 10, 15 million dollars. Publicly, IBM has said that their Watson moonshot is over a billion dollars. I would imagine the Google car is a similar, if not larger, investment. We talked about the BMW i brand, that's a multi-billion dollar investment.
22:43 Speaker 02 So the way you differentiate is not only by the fact that you're trying to do something new, but also because you're investing a significant sum of money. And you're solving a very complex problem. Again, what IBM is trying to do is trying to make artificial intelligence extremely easy by encapsulating it in this platform so that any application developer through the, through a series of APIs now can get access to all of that intelligence. I mean, and, and they're, one of the major application areas they have is in the healthcare sector.
23:26 Speaker 02 So you can imagine, or what they want to imagine is doctors, you know, naturally interacting with applications that are based on Watson in order to provide advice, in order to make difficult decisions. So again, this is, this is a very hard problem. And that's why I said it needs the, the proper experimentation, the proper patience, the proper personnel, frankly, I mean, hiring is, is another big dimension. Who do you hire?
23:55 Sonal? So evaluating or defining the moonshot, it's something new, it's a hard problem. It's got significant investment behind it, the right kind of hiring. Astro Teller made an argument in an op ed for me a few years ago that it has to be 10x better. It can't be just, you know, 10, like, it has to be not just incrementally better, but like orders of magnitude better. He also made the argument, interestingly, that there has to be some sign today that you can actually get there. So it's not just so pie in the sky, like, human beings need to fly without wings, but more that there is some indicator, whether it's Moore's Law is on a trajectory or a curve, that there is some indication that we can actually get there, even if it's not available right now. A question I have for you, though, Evangelos, is Watson is an interesting example on multiple fronts,
24:39 Sonal? but I can't help but wonder, how can it compete with what a new class of startups can do? And when you have other companies that have a whole different type of data that IBM does not have access to, like Google and Facebook, for example, and what they're doing. And that's just the example of AI. This plays out in countless domains where can a big company actually compete with startup-driven innovation?
25:04 Speaker 02 So I would say the short answer is yes, because when we talk about startup-driven innovation, we, we think of it as the corporation adopting many of the practices that make startups successful. One of these practices, significant partnering and knowing who to partner with and how to develop these partnerships efficiently. So to us, I mean, to Brad Power, my co-author in the HBR article, and me, this was one of the, the most successful aspects that IBM has, has done. And we encourage them to, to continue, to continue doing it.
25:46 Speaker 02 Because I think corporations are starting to understand that in this environment, regardless of how much capital they want to throw at the problem and how big and how good they are, they cannot solve things on their own. And they, they need to co-innovate. Steve and I are finding that corporations that have adopted
26:09 Speaker 02 Open innovation principles where they are co-innovating with other large corporations are becoming more successful at adopting startup-driven innovation.
26:22 Sonal? I think that's actually a key point because of the common problem with absorbing any outside innovation, whether it's from an outside R and D lab, an internal captive R and D lab that's just headquartered somewhere else, a startup or partnering with another company. The main problem is NIH, the non-invented here syndrome, and how do you then get over that? So there is certainly an advantage in shifting the mindset of the company and its practices so it can actually absorb what's new when the time comes. It just still seems, though, that there's no way a company can overcome its core incentive to make money from its core business.
26:59 Michael Copeland? It sounds like you might be arguing for, you know, the sort of the natural atomic unit of innovation is the startup. And once it reaches a certain age or a certain size, it's over, right? So, I mean, honestly, Google struggles with this, which is why we see Alphabet now. But how do you, I mean, in some ways, it makes me think of like this sort of what we're desiring youth, like, you know,
27:19 Sonal? Right.
27:20 Michael Copeland? big companies just want the youth that these startups have. But is there something more sort of engineered and cultural that they can wrap their arms around, that they can bring in? Does Facebook fail at some point for lack of being a startup anymore? Google, same thing?
27:35 Sonal? Facebook's also experimenting the way Google is in the sense that when they acquired WhatsApp, they made sure that Jan Coombe had a seat on the board.
27:42 Michael Copeland? Do you think it's impossible for companies who didn't come up like Google and Facebook and others did?
27:48 Speaker 02 The short answer is I do not know. What we are saying is that in this environment that is being set up with these forces that are at play right now, the more traditional ways that corporations have had to innovate need to be rethought.
28:06 Michael Copeland? Right.
28:06 Speaker 02 Okay. We believe that the startup playbooks or best practices that we've learned from startups and their entire environment Have an important role to play in this rethinking. Will everybody get it the same way, right? No. Will everybody do it the same way? No. But there's still something too significant to consider, especially for corporations that are thinking how to deal with their, you know, Horizon 2 and Horizon 3. I think corporations, because of their scale issue, because of the, sometimes because of the short-termism
28:50 Speaker 02 that they operate in, whether it is because of activist investors, whether it is because of their own doing, they have become very good and have optimized how to deal with their operations in the short term. And because of that, more so than in the past, they are facing risks on how to deal with their longer term. And this risk, as I said, it's not only because of what they're doing today, but also because of all of this technological revolution that is happening. So they're being hit, as I said, from a variety of directions. To address their problems,
29:33 Speaker 02 they cannot just do one thing, okay? They need to be doing several things.
29:38 Michael Copeland? We talk about what corporations can learn from Silicon Valley, but we run the risk of the outpost becoming an outpost, right? So what can Silicon Valley learn from Silicon Valley? And how do we ensure that what's here and what's so valuable continues?
29:54 Speaker 02 Many times, whether it is us as investors or entrepreneurs or startup executives, we do not give the proper credit to the corporation. We are not seeking to have a value exchange. It's this give and take. And one of the recent examples that brought this to bear for me is the automotive industry. I've been writing about, I have been working with automakers and have been writing about the disruption and the challenges that the incumbent automotive industry has.
30:35 Speaker 02 There is also a lot of enthusiasm in the valley from our own companies that we can completely overtake the automotive industry. And I think this is a fallacy. Completely obliterating it and making Silicon Valley the next Detroit is hard for me to imagine that. Maybe I'm too short term thinker myself. But I think this is, this provides a great opportunity where the local automotive ecosystem can really start collaborating with the incumbent automotive ecosystem and build something great, right? So that's how I believe Silicon Valley and any other ecosystem, startup ecosystem
31:20 Speaker 02 avoids being disrupted itself.
31:22 Sonal? That was all we have time for. Evangelos, thank you so much for joining the A6 and Z podcast.
31:27 Speaker 02 Thank you very much. We'll see you at the outpost.
Previous Company Building in Crypto: A Conversation between Brian Armstrong and Chris Dixon Next What's Next for Education Startups Page updated October 9, 2026 · transcript and summary are machine-generated