a16z Podcast - Marc Andreessen on Startup Timing

Episode Date: July 2, 2025

What if now is the best time in decades to start a company?In this episode, taken from Speedrun, a16z’s accelerator for early-stage founders, Marc Andreessen joins games General Partner Jonathan Lai to make the case that we’re entering a once-in-a-generation window for innovation. From the rise of AI to the cultural and policy shifts reshaping the global economy, Marc explains why the next four years present a rare opportunity for builders to seize the moment.Along the way, they discuss market timing, platform shifts, and what sets successful founders apart - including lessons from Steve Jobs, insights into AI’s impact on storytelling and games, and why being “too early” can feel just like being wrong.Timecodes0:00 Lessons from Steve Jobs on Leadership & Innovation2:27 The AI Boom: How It’s Changing Everything5:52 Market Timing: The #1 Factor in Startup Success8:13 Why the Next 4 Years Are Critical for Tech11:30 AI & The Future of Gaming, Storytelling & Virtual Worlds14:28  Why Some Startups Fail While Others Explode17:11  The Role of Founders in the AI EraResources: Find Marc on X: https://x.com/pmarcaFind Jonathan on X: https://x.com/TocelotStay Updated: Let us know what you think: https://ratethispodcast.com/a16zFind a16z on Twitter: https://twitter.com/a16zFind a16z on LinkedIn: https://www.linkedin.com/company/a16zSubscribe on your favorite podcast app: https://a16z.simplecast.com/Follow our host: https://x.com/eriktorenbergPlease note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures.

This is not saying that every startup should be an AI startup or whatever, but there is this incredible new tool that can be used, and it's just sort of general fact that big companies, big incumbents, have a very hard time reacting to these platform changes. Sometimes they pull it off, but a lot of times they really struggle. And so it's sort of the best possible opportunity for a startup going up against an incumbent is when there's this kind of shift happening. What if right now is the best time in decades to start a company?

In this episode, taken from Speed Run, a three-month accelerator powered by A16Z designed to help founders move fast, Mark Andresen joins Games general partner Jonathan Lye to make the case that we're entering a once-in-a-generation window for builders. From the explosive rise of AI to shifting global policies, Mark breaks down why the next four years presents a rare and urgent opportunity for founders and what separates the entrepreneurs who sees the moment from those who miss it. Let's get into it.

As a reminder, the content here is for informational purposes only. Should not be taken as legal business, tax, or investment advice. or be used to evaluate any investment or security and is not directed at any investors or potential investors in any A16Z fund. Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast. For more details, including a link to our investments, please see A16Z.com forward slash disclosures. How much do you think disagreeability or just the ability to sort of fight and sort of disagree publicly as a necessary trait? founders, as exemplified by Steve Jobs. Yeah, so start with Steve, who I knew.

So Steve is one of the most disagreeable people in the history of humankind. You know, Steve would disagree with you over the shape of the glass on the table in front of you, like it was going to argue about everything. So it's where a lot of the genius came from, which is he was just not going to take the status quo for granted under any circumstances. Elon uses a term first principles thinking, and Steve had a lot of that as well. So there's a lot of genius in there. But if you read the books on Steve or you hear the stories, there's basically two stories about

Steve you hear. One is that he was a saint and he was perfect in all regards, which. which was somewhat true. But the other story that you hear is basically he was like a screaming lunatic and he would just run around and yellow people in elevators and fire people in meetings and just he was like, you know, often people have all these kind of awful horrible things. It's like angel devil kind of thing. And I think on this, I think the reality was somewhere in the middle, at least what I saw

and what I heard from people who worked with him for a very long time, was basically he was just like absolutely intolerant of anything less than first class work. If you brought him first class work and if you were top in your field and super diligent and on top of everything and had all the details figured out and knew what you were doing and were really good. He was like the best manager

you were ever going to work with and the best CEO you were ever going to work with. And the thing that comes up when people talk about who worked with them closely is I did the best work of my life working for him, right? And part of that is because

he really appreciated understood the quality of great work and he didn't tolerate anything less than that which meant that everybody around you also hit that bar. Sort of his approach for performance management is everybody's going to be doing top-end work if not they're not going to be here.

As a consequence, the best people in the world are going to love being here because they're surrounded by the best people in the world. And by the way, that's also something, of course, that Elon shares. And so the stories that you hear about the screaming or firing people in meetings or whatever, it was always a side effect of it's not first-class work. And it's either somebody who really should have been able to do it, who didn't,

who didn't work hard enough, or it's somebody who was just not capable of doing it, right, who needed to go find something else. But the result, the core thing always, and this is why there was so much love for him, right, is the core thing was, I do the best work of my life working for him. Having said that also, Steve himself matured, right? And his career kind of got, time dilation kind of happens. And so you kind of think, well, Steve built Apple, the company that we all know and love today. And it's like, well, yes, but it took a very long time and there were twists and turns along the way.

And the big twist and turn, of course, is he got fired, right? So he started the company in 76 with Steve Wozniak. You know, Apple 1, Apple 2. They built Elisa, actually, which failed. That was the other thing, by the way, is not every Steve Jobs project succeeded, right? And he learned a lot from the failures. And now everybody's forgotten the failures, but you can read about them on Wikipedia. and the Lisa was before the Macintosh, and it was a complete failure.

But then the Macintosh started to work, but before the McIntosh really took off, he brought in the wrong CEO, and then there was a revolt inside the company. And he got fired. And then he spent 10 years out of the company, right? Which is actually when I got to know him is when he was off doing Next. Next was like a complete wipeout. Like it was years of real pain and agony. And they had this little side project nobody understood,

which is this little graphics company, Pixar turned out. Anyway, the point is by the time he came back into Apple in like 97, I think maybe been out for like 12 years. and again what people who knew him better than I did said he learned how to actually be a great CEO not at Apple but at next because he spent 12 years actually doing it the hard way where he wasn't being shower with praise he didn't have the magic touch the product fundamentally didn't take he had to like as we say now pivot because we have this great new term pivot you should just say fuck up you know pivoted next by the way when I was in college in 92 not 89 when I got 89 I got 89 I got Illinois 89 we were one of the early adopters of the next computer N-ExtT was the name of the company.

It was called The NextCube, was the computer. And it was sort of the post-Macintosh. And so it was like, literally, it was like a $15,000 macintosh. And it was like fully modern, not only graphical, but it was like full Unix at a time when that was really unusual for a desktop computer. It had this incredible state of the art. It had the first CD-ROM drive. And then one of the great stories is it was a perfect cube. It was a perfect 12-inch cube, 12 by 12 by 12.

And there's a famous story of his designers at Next when they were designing it. They came in with a sort of optimal, faredo optimal hardware design, built for manufacturing, cost optimizations. which was 12 inches by 12 inches by 13 inches. And he said, fuck you, go back and make it a perfect cube. And they're like, Steve, that's going to double the cost. And he's like, I don't fucking care. Make it a cube. And so it was like a perfect cube.

It was slowish. Like, it took forever. It was like walking through molasses to use the thing. Completely flopped. Nobody wanted it. Pivoted the company to software. Nobody wanted the software. Anyway, the point is, like, that was really hard.

And he had to make every part of the company work. He had to try to figure out how to optimize it the hard way. retain a team through 12 years of basically failure. Basically, people say, as he had this incredible growth in innovation skill set from Apple phase one, and then he had an incredible management skill set from the sort of wilderness years. So by the time he came back to Apple 97, he was both. And he was also at that point a great CEO, but he maybe not would have ever become the

Steve Jobs that we know had he not gone through the heart period. So a big undercurrent to what you've been discussed in just the concept of like market timing, right? Like what's the right time to sort of enter a market? It's the technology truly ready for sort of the consumer or the enterprise version of that technology. And so thinking back to like the 90s, you wanted to be not pets.com, but Chewy eventually did it wrong, right? Like some of these companies that came 10, 15 years later, if you're a builder right now and you're selecting like markets right now to

build in, how would you think about what the right time is? Is it too early for certain markets and so on so forth? Yeah, so one of my observations, it sounds crazy, but I think might be true is I think actually all the ideas might be good. Qualified founders who know what they're doing, I think generally are right about the opportunity that they're going after and that it's just okay, if you're like a qualified founder doing what you're doing, you should be so deep in the domain that when you think about, okay, this technology now makes this use case possible and we can build a product that does that.

People are going to want that. I think the accuracy rate on that is almost 100%. The problem is the timing. And early is the same as being wrong, right, in practice, right? And you see this in the pattern, which is basically is anytime there's an overnight success, like a technology company that just like lights the world on fire is doing great.

What you find is basically there were a set of companies that try to do that same thing five years earlier and failed, five years before that and failed, five years before that and failed. where that failed, often goes back. By the way, I mentioned the e-commerce and the phone thing. The first actual smartphone came out in 1987. Right?

And so there were literally 20 years of failed attempts to do smartphones until the iPhone, you know, 20 years, right? And so I knew many of the founders trying to do smartphones before the iPhone, and they were very, very smart, capable people. Many of them very successful and in other domains are very successful later on. It was just too really. The technology wasn't ready yet. And then why is this hard?

It's hard for basically two reasons. One is because to be a successful founder by definition, you have to live in the future, right? So you have to envision a world that doesn't exist yet in which the thing that you're building is something that everybody's going to use. And like people around you can't see that yet yet. And so like you see a vision of the world that doesn't yet exist. And that may be right and may not be right. Or it may happen now and it may not happen now. And then the other thing about it is the world gets a vote. And, you know, I think one of the really interesting kind of conceptual questions is there's like the push side of you trying to make something happen in the world and trying to get the world to understand something and want to do something and buy something. But then there's like some sociological. thing on the other side, which is, okay, when is the world actually ready for the new idea? Here's another just, like, really amazing thing. Somebody got one of the small versions of Lometer run on Windows 98. They booted up a, literally, a Dell desktop computer from, I think, 1998, with a fresh copy of Windows 98, and they got Lama running on it.

And so, like, all of those old PCs literally could have been smart this whole time. They really could have been. And, like, we had neural networks, right? And there were lots of people working on AI. Yeah. But, like, we could have been talking to our computers in English for the last basically almost 30 years. We now know.

It is the alternate history. Crazy alternate history. And like, we didn't. There was an AI startup boom in the 80s. There were a lot of really smart people in the 80s who thought this was all going to happen then. There were people in the 50s who thought it was going to happen then.

And so this timing thing is just like incredibly difficult. And then just practically speaking, as a startup, you only have so long to either prove it, right, or not. And basically, you know, give or take five years total, right? If you don't get traction in the startup in the first five years, your odds of success are very low just because it's very hard to hold the team together.

You're going to race against time. Yeah, you're in a race against time. And then you end up in this frustrating, you know, I have a lot of friends who've ended up in this frustrating circumstance where you started a company in 2010 you thought now was the time your company failed in 2014

and then a company started in 2015 that just hit it and went. And you're just, fuck, right? And so the timing part of it is very hard. I don't think there's an actual answer of the timing thing other than I think that's a really key part of they call entrepreneurial judgment.

Like, I think that's a big thing that the founder has to really work hard about and think hard about. And again, it's this thing of like reconciling. I live in the future and I can see it happening but like, can I actually deliver it in the form of a product that actually works the way people expect?

And then is there enough precondition in the world such that they're going to want to actually do this thing? And I just like, I mean, these guys have to tell you, like, I don't even try to forecast these things myself anymore. Like in meetings, I just, I just figured like, let's just assume the founder is more likely to be right than we are just because, like, this is the best, the founder's best guess is probably the best guess that exists

because the founders have the most information can kind of synthesize a view on this. But having said that, you know, we're going to back, you know, I don't know, like a third of the companies or whatever, and we're going to back in a given year. It's just going to turn out that they were too early. By the way, the other corollary to this is if it's going to work, it almost always feels like you're too late.

So as a founder, it almost always feels like, I wish, I should have started this company. If only I started this company two years earlier, right? Because it's like you're going up in this situation. You're like, oh, my God, like it's going to happen and the market's coming and it's going to happen. But like my product's not quite ready yet. and like, oh, shit, like, I'm running out of time, right? And, like, my experience at least is that that feels terrible in the moment because it feels like you're not rising the occasion,

but that's actually the positive sign. What excites you the most when you think about AI and sort of what it can do to create storytelling and just the entertainment industry at large? So aspirationally, you know, there's opportunity for like a complete reinvention of the entire process of storytelling and games and immersive worlds, right? And, you know, this is the sort of thing. Instead of us making them, maybe they're going to get,

hallucinated. They're, you know, they're going to get dreamed. You know, you're playing a game and it's going to basically, right, it's going to be an AI sort of in symbiosis with you as the player and it's going to, you know, adapt, it's going to learn from you, adapt to you, and make you the best possible experience, you know, with the best possible, you know, the tradeoff between, you know, ease and difficulty learning curve and the most possible entertaining scenario for you. And, you know, the best games, people will play for, you know, many thousands of hours. And then by the way, the businesses can be really amazing underneath this because they don't, you don't bear all the cost of manual content creating.

And so, you know, you build an engine like this and it runs, you know, it runs forever. And so, like, that's super exciting. And then, you know, being able to have virtual worlds with like, you know, super smart, you know, lots of, you know, populated by, you know, these incredible personalities. You know, it's just a whole new, amazing kind of storytelling. So, like, I know, and we're big believers and all this. And like, I think that's all going to happen. It's going to be, it's going to be fantastic.

I will say I am also seeing this through the eyes of my nine-year-old. So I'm re-going through the, you know, the process. of learning about gaming and coming up to speed as a gamer through my nine-year-old. And so a couple of key lessons. One is, well, Warcraft apparently is, or not work. Minecraft is immortal. So, like, he loves Minecraft more than life itself. And so that, like, that, whatever, whatever those guys did is just, has worked.

And by the way, really interesting thing, like the nine-year-olds, at least the nine-year-olds, they don't care at all about visual fidelity, visual quality. Like, he could not give it. And so Minecraft and Roblox and games that look like that. that are fantastic. And he just discovered his first, I just learned about I.O. games. People have heard of those. And so he just discovered, he just discovered his first person shooter, which just thrills me to no end. That's, that's called shell shockers. And it's basically it's counterstrike, but with anthropomorphic eggs. Oh, wow. So it's like full military

hardware. And then when you shoot another player, he explodes in a shower of yoke. And it's a completely browser-based game, right? It's multiplayer, but it's all browser-based, and so it punches through all the school firewalls. And so the, and it runs great on Chromebooks. He's very frustrated. He can't get it. When we take away his laptop, he's currently on a laptop break. When we take away his laptop, he tries very hard to get it working on the web browser on his Kindle. That's his main form of frustration is life right now is that, and we tell it's good that Shell Shockers doesn't run on the web browser on the Kindle, because if it did, we'd have to take away the Kindle. So, but anyway, but,

Anyway, I watch him doing this, and then, you know, and then I had this moment when Chad GPT came out where I was like, wow, this is amazing. And, you know, I can't wait to show this my kid. And I'm going to, you know, I felt like, that felt like Prometheus bringing fire down from the heavens to my, to my kid. It's like, I'm bringing him, you know, AI, and it's going to be this tutor and coach, and it's going to be with him forever, and it's going to talk to him,

and he can ask him questions, and he'll teach him about all these things he's interested in. And I, I install it on his laptop and show him, you type in a question, it'll answer the question. And, you know, and I'm like, this is like the best thing I could ever possibly do for my kid and, you know, he looks at me and he's like, so? And I was like, no, it's like, it took 80 years with the neural networks and the this and the training data and the whole thing

and it like does the thing and it makes it da-da-da-da. And he's like, that is a computer. Like, if it doesn't answer your question, it's like, what would it do? Well, against this backdrop, right, like we've got new administration, new policies, new technology, sort of new sort of cultural movement potentially. It feels like we're very much at a moment in time. I've heard you describe it before as like we could be in the precipice of literally

a golden age, right, for America and for the world. What are some ways that startups today, like in 2025 could be thinking about capitalizing just on this moment, right? Yeah, so look, I think for at least the next four years, I think it's basically blue skies. And so I think that, you know, now is definitely the time to build. You know, as I said earlier, like our whole theory of all this is that prime time for startups

is when there's a platform change. and I'm not, by the way, this is not saying that every startup should be an AI startup or whatever, but like there is this incredible new tool, you know, that can be used. And it's just, again, it's just sort of general fact that big companies, big incumbents, have a very hard time reacting to these platform changes. Sometimes they pull it off, but a lot of times they really struggle. And so it's sort of the best possible opportunity for a startup is going up against an incumbent is when there's this kind of shift happening.

So I think that's happening. Yeah, there's going to be four years of clear, clear sailing, I think, in the U.S. You know, the rest of the world's in a funny state. You know, the EU has all but made AI illegal at this point. They literally, this is the amazing stuff, they literally say publicly, they're senior officials, literally said publicly a year ago.

They're like, well, we've realized that you cannot be the world's leader in AI innovation. And so therefore, we will be the world's leader in AI regulation, which is one of those things that I think you only say if you're an unelected bureaucrat, completely lost in your own thoughts. But, you know, they passed this very draconian AI law that as a consequence, even the big AI labs now are not launching new AI products in the EU Because they can't figure out how to do it legally The UK actually backslid on this. They they should UK should be a haven for for openness and instead they they basically have matched the EU on bad regulations and seem like they're going sideways on that so so and this is important because like the the main US trading partners are actually our main our main trading partners at China our main trading partners Europe

like the the the the globalization is sort of fundamentally, you know, in the West is based on the economic relationship between the U.S. and Europe. And then, yeah, I mean, look, you know, the entire world is opening up. I mean, you know, kids everywhere are now online. They have access to all this. They want to use all this technology. They want to build it. You know, they want to be part of these companies.

Some of that obviously happens through, you know, migration, but a lot of that happens through just like everybody's online now. And so there's, you know, talent all over the world. And, you know, hopefully the Internet itself, you know, and general economic arrangements stay free enough where people all over the world can come together and build things. Sure. That's very exciting. Definitely feels like we're a unique moment in time to be building. I think it's a great note there to wrap it on. So thank you, Mike and Jason. Let's give him a hand.

Good, good. Thanks, everybody. Like, it just drives you fucking nuts, right? And by the way, is it okay? This is an adult audience? Can I swear, is this everybody over the age of 18? It's hard to tell. If you're below the age of 15,

It's just whatever the new thing is, it's just the way the world has always worked. It's totally normal. If you're between the ages of 15 to 35, you know, the new thing is very exciting and you might be able to make a career out of it. If you're above the age of 35, the new thing is unholy and against the social order

and it's going to destroy civilization. Thanks for listening to the A16Z podcast. If you enjoyed the episode, let us know by leaving a review at rate thispodcast.com slash A16Z. We've got more great conversations coming your way. See you next time. Now, now, now, now.