E438: Will Robbins on Elon Musk, AI & the Future of Silicon Valley

How I Invest with David Weisburd

Will Robbins joins David Weisburd to discuss how artificial intelligence is fundamentally altering the venture capital landscape by drastically reducing startup formation costs and diluting the traditional value of top-tier VC

Key takeaways

• AI coding tools have collapsed the traditional startup funding ladder into an "L-shape," allowing companies to validate ideas with minimal capital before raising massive seed rounds.
• Traditional VC brand prestige (e.g., Sequoia, Kleiner Perkins) has diluted due to market saturation and information overload, making founder reputation and mission storytelling more critical for talent acquisition.

Main topics

• AI impact on startup costs and funding rounds
• Founder identification and the "X factor"

Notable quotes

"In a world where it's so easy and cheap to go build software... the only thing standing between you and a fully featured product is no longer a year of engineer time, it's 50 billion tokens."

Conclusion

The episode provides a pragmatic blueprint for modern venture investing,

Transcript preview

Speaker 2 (0:00) In a world where it's so easy and cheap to go build software and get to product market fit, where you need almost no money to spend on tokens to write your code for you, the only thing standing between you and a fully featured product is no longer a year of engineer time, it's 50 billion tokens. If you look at the average unicorn startup, the profile looks something like some 40-something-year-old executive spends out, builds a company in the domain they know about. But if you weight these companies by market cap, it's a very different story. If you look at the... college and grad school dropouts. You get Facebook, you get Google, you get Airbnb, you get Stripe, you get so many others. You're Speaker 1 (0:33) not looking for somebody with the right product or right market. You're looking for something else. What exactly are you looking for? So before we started recording, you told me that VC brands no longer matter. What did you mean by that? There's a great irony today that I see, which Speaker 2 (0:47) within the Valley, people obsess over the relative hierarchies of different venture brands, right? How does Sequoia compare to Kleiner Perkins? How does Index compare to Bessemer? Speaker 2 (0:56) To me, I find that funny because if you pull a person off the street here in New York City, they've never heard of any venture fund and they don't really care. And so in a world where talent and capital access is the most important thing for a startup, then maybe which venture firm you work with matters more. But in a world where software is largely a solved problem with AI and the costs of company building have fallen so precipitously, and if distribution and real world trust network are the main constraints to AI and software percolating into the real economy, it's unclear to me if there's any venture brand that has a real edge there. You see a lot of firms building. executive education centers where you can go, you know, pull people into a conference room and, you know, tell them all the cool things that their competitors and peers are doing with software. And of course, the portfolio companies of these firms. And I think that works well, but the trust and the reputation doesn't transfer over. I'll give an example. We had a portfolio company. Did the first check at my last firm, Contra, about five or six years ago. They raised money from a bunch of investors. They scaled up and they sell to insurance companies. Speaker 2 (1:55) Yeah, there's one advisor who has been, you know, sorry to throw the other VCs under the bus, but I'm in the same bucket. There's one advisor who was a formerly a senior executive at the U.S. Medicaid program, has been more helpful to this company than all the VCs combined, right? This is just one guy who had a small number of advisor shares. And why is that? It's because the government and because the insurance companies know who these Medicaid leaders are and Medicare leaders are. They don't know who the VCs are, right? And to me, this is kind of a shocking. revelation in such a large part of our economy that the whole healthcare world can't really go pick and choose what tech solutions to use because there's just no legibility to it all. Me Speaker 1 (2:34) and my business partner, Curtis, were talking about this curse of knowledge where you forget how much you know and how insidery your world perception is. Most people in the United States not only don't even know who Scott Besson is, but don't even know the position of Treasury Secretary, what it does, and all these things that, to me, and you seem very obvious. But just to steel man your argument, arguably the number one thing that a startup needs to get right is talent and being this talent magnet. Elon Musk famously is the best talent magnet of our generation. That's in many ways a secret sauce of how he's able to build these companies. Don't you need the VC brands to build out the right talent early on, especially before the company itself has a brand? I Speaker 2 (3:16) think the frontier for these companies is elsewhere now. There's a couple reasons for that. One is that there are so many rounds being done, right? It used to be that five years ago, if you had a tier one brand go lead your round, you'd have a TechCrunch article, it'd have your three hours of dominance on the Twitter timeline. Everybody would read the article and talk about it and say, did we see this company that this firm led a round into? But nowadays, again, the industry is so mature, right? It's a good jungle where you see a different high profile Series A announced twice a day, seemingly, right? And so there's an inherent dilution effect that's through nobody's fault, but it still creates a fuzziness to the signal if there's 300 companies that all have the same signals that you do. And so you can debate whether or not the brands have diluted. You can debate whether or not it's now table stakes for founders trying to go attract talent and capital and so on. But regardless, I think we can agree that's just not enough and that the real frontier has to be either a highly differentiated, unique brand. And by the way, this is why a lot of the solo capitalists have been so prominent recently in that, you know, Dylan Field is like a more unique name as a kind of, you know, Figma CEO founder type brand than a firm. It's also why you see the mission and the storytelling and the going direct to your customers takes such a precedence in that people want to join inspiring missions. And so the wordsmithing around your mission and why people should join you is almost more important now than capital. Speaker 1 (4:32) Is this why you've seen founders and CEOs being so obsessed about their public perception and their personal brands? Speaker 2 (4:37) There's markets in everything, right? There's markets in capital, there's markets in talent, there's markets in information. And maybe the interesting way to frame this environment we're in now is that each market has its own ups and downs. And sometimes these are correlated, right? Sometimes they're not. And maybe a tough economy, maybe capital is more scarce, but then talent is also more hungry for... quality roles. And so, you know, maybe an unfavorable capital market can actually mean a favorable talent market. And this is the way that most businesses in the economy probably operate. A funny quirk about the tech world right now is that things are just fully on all the way, both regards to capital markets and the competition for talent. And so the argument against people raising less money now in venture, because you can go build companies and build software so much more cheaply, is probably that you're going to have to spend twice as much on half the number of seats and it's going to balance out. Capital will be consumed just as it was before, rather than people choosing not to raise early on when starting companies and just building software with AI. Is Speaker 1 (5:35) that how you explain this paradox where it's become increasingly cheaper to start businesses and to scale businesses, but yet they're raising $50 million seed rounds, for example? Speaker 2 (5:45) So this is a huge topic of debate, actually, with a lot of real estate friends. Me and others have hypothesized on this idea of the token round, right? What is a token round? It used to be that in the old days, you'd have a sort of de-risking ladder, right, where you raise a seed round, you hire some engineers, you incrementally de-risk your business, and then you raise a larger round, and then you get to some sort of product market fit, whether it be a million dollars of revenue or some other signal. And you go raise $20 million, and you kind of ladder up the capital curve as you go prove out each incremental element of your business. In a world where it's so easy and cheap to go build software and get to product market fit, that ladder just can collapse and become more of an L shape where you need almost no money to spend on tokens and to kind of write your code for you. And then once you can go test it in the market and get feedback, you can immediately go validate and go straight to a $20, $30, $50 million round with that validation. And so there's a bit of an amorphous or maybe more rough capital requirement curve these days. One consequence of this might be that You just raise less money and you raise half a million dollars or a million dollars to spend on tokens with the belief that the only thing standing between you and a fully featured product is no longer a year of engineer time. It's 50 billion tokens, in which case you can go raise very little and dilute very little and then scale up and raise your scaling around from a multi-stage firm when you have that validation. The other argument is that there's markets in all things. And so even though the cost of company building is falling so much, First of all, you can go take on other types of businesses that are more operationally intensive and expensive. And you see this in like the deep tech companies, you see this in sort of like the neobanks, right, which obviously have a lot of ops and other requirements. And then you compete for talent. And that's what makes the $5 million seed round kind of such a persistent thing. So those are two sides of the coin. I think there's truth to both. And we've seen both. I've seen a lot of companies that raise $5 or $10 million out of the gate to go hire a bunch of people and to do a lot of stuff that you can't just do with software. I've also seen people raise very little and spend even less. I'll give you an example. I have one friend who previously took a company public. He then started a second company. And of course, having taken a company public before, had a lot of offers for a lot of capital from a lot of different places. Being the disciplined, humble operator he is, he thought, I don't need that much money. I'm going to raise only $4 million. And I'll build a team, I'll build a product, and I'll go from there. And if there's any gap, I'll just, you know, I'll self-fund it. But I don't want to overextend and build a big preference deck. I thought that was very prudent. And six months in, he has a fully featured product ready to launch. How much do you think he spent out of that $4 million he raised to go build a fully featured product? Small amount. $800,000. He didn't realize that coding models would get so good. And of course, this is a pure software product, right? And very few companies have only software as their main product building constraints. But I see this and it happens. And I wonder, man, play this forward another few years. If people start to see these examples, what happens to the seed round? I think these days people should be building funds and sizing their funds with that in mind. So you can play, you know, both within the small token round where capital requirements are lessened, but also the larger capital intensive physical world or ops heavy round. Speaker 1 (8:49) This episode is presented by Juniper Square, the operations partner for private markets. Speaker 1 (8:56) In venture capital, it's reverse. It's an access class, not an asset class. So your customer is really the entrepreneur. So if you're not building around the entrepreneur's capital needs, what exactly are you doing? Speaker 2 (9:05) Right. And that's a funny question. Who is the customer in venture? Is Speaker 1 (9:08) it the LP? Is it the founder? Speaker 2 (9:10) Is it the Twitter followers who promote you and give you the attention you want so desperately? It's always hard to say. Obviously, it's that famous Buffett quote that's in the short run. It's a mongerism, I'm not sure. In the short run, markets are voting machines, right? In the long term, they're weighing machines. And I think kind of similarly, you can probably say that in the short run on a fund by fund basis, you live or die by your LPs. I think in the long run as a firm, right? If you want to have a 20 year career, you have to live or die by your founders. Because if you don't have the deep founder relationships, the flywheel is not going to persist. I Speaker 1 (9:40) want to go back to your friend who raised $4 million, spent $800,000, $3.2 million in the bank. A lot of VCs will say raising too much capital is a problem. Because it creates poorly run companies. Is that just VC rationalization into why founders should raise that at lower valuation or is there something there? Speaker 2 (9:59) Everybody has different incentives and people's viewpoints are largely downstream of those. Maybe the classic example of VC versus founder incentives is that of acquisition offers, right? Where a VC only cares if you can go 100x the investment and return the whole fund on that one deal. And so anytime you have a 10x acquisition offer that may be life-changing for the founder, or maybe even in the best interest of the shareholders, given that you may not see a long-term 100x outcome, the VC is going to want to push for that, and Speaker 1 (10:25) the Speaker 2 (10:25) founder is going to want to go. take liquidity or Speaker 1 (10:27) shut down that business, sell it to somebody and then go start a new one that does have the venture outcome possibility. Speaker 2 (10:32) And I think every business is so unique and I almost reject all forms of, speaking in blanket terms, there are companies that should go raise a 500K and stay very lean and just buy 10 different Claude Code max accounts. and just write a ton of software and build a lot of product, there are teams that should go take on more ambitious and operationally intensive experiments and treat their time and treat the learning as the more valuable asset than the capital. One Speaker 1 (10:57) of the parts of this decision making is the dynamics of the business, how it scales, whether it's win or take all, whether it's network effects. My personal view is that Almost all startups, one way or another, are network effect businesses. What are your views on this? That's a great Speaker 2 (11:13) question. I think the essence of all great software companies and technology companies must be some accumulating advantage. A network effect is maybe one category of those. Maybe the economies of scale that made the retailers like the great outcomes of the 20th century. So powerful. Can you double click on that? Sure. Maybe I'll tell this lens. This is a good way to go talk about Zepto. This Indian retailer has been a really interesting technology company I've been involved in. We forget that retail is just like the greatest business model of all time, right? If you go look at Amazon and Walmart and Target and Home Depot, right, there's trillions of dollars of market cap between these companies and savings to the consumer that have been provided are unbelievable, right? The access and the savings as well. We live in a fully kind of economized retail landscape managed by complex software systems and ERPs and systems of record and obviously decades of experience. And so we forget the water that we're swimming in. But, you know, Without these large retailers, goods could be 50 % or 100 % more expensive than they are and so much less accessible. And so I think the economy of scale is one of the great moats of all time. And it's one that is great, hopefully, both for consumers and for businesses. Speaker 2 (12:18) There's a company called Zepto in India. It's a very unique company. There were two 17-year-olds I met who wanted to go to college at Stanford. COVID happened, and they thought that going to school and doing Zoom, Zoom College was a bit of a fool's errand. And so they came to me when I was at my last firm, Contra, and said, we would love to go start a company and skip college and build a grocer in India. It's so hard to get groceries. The retail infrastructure is disorganized. It's unreliable. Our neighbors need food. So we offered to invest. We made an investment. They incorporated. They didn't go to college. They stayed in India. And at the time, it was just a WhatsApp chat. Organizing some grocery deliveries between neighbors became a little software app where you can go purchase groceries online, connect to the farmers or whoever was producing it. Speaker 2 (13:06) and order food by the bunch. Fast forward about five years, they built a network of dark stores that cover a mile long or kind of mile radius. The core insight behind Zepto was that the Indian retail landscape was largely made up of mom and pops, right? This very diverse, very ingrained kind of ocean of people. It's almost culturally embedded. Exactly, right? And so the team realized that building software for these mom and pops was too hard to go. do. It didn't make sense to go sell software to these businesses with their amazing variety and diversity and custom workflows. They decided that they should just go build the Walmart of the Indian 21st century. They built dark stores that cover about a mile radius. They have a system of delivery drivers that will go and 10 to 15 minutes ago, pick up an order that's packed, deliver it to you, and then go back to the store. And they do that over and over and over, dozens of orders per day. They built a network of these dark stores, which were each supported by one like big mother hub warehouse. Imagine a giant, you know, million square foot warehouse in each city that then serves these, you know, dozens of kind of smaller. spoke dark stores that are hyperlocal and Speaker 2 (14:11) It's been remarkable growth. I mean, they started about five, six years ago. They're now doing about 2 million orders a day, about $4 billion in sales annualized, more or less. And the founders are still, I think the CEO just turned 24 two weeks ago. And so it's a remarkable kind of Indian story, both in terms of the team and a very kind of young, precocious Zuckerberg-style Cinderella story. I think it's also a story of the kind of development of nations and how Zepto is not only able to go deliver food in 15 minutes on demand in a very convenient way, it's able to do so with better selection and better prices as well, because the economy is of scale, right, and buying things in bulk. and organizing them into warehouses and building software tools to go plan your inventory and understand the lead times and purchasing amounts leads to great efficiencies. That said, it's a tough business, right? There's no true low-hanging fruits in retail. I remember the Zepto team would do things like redesign their packaging to save a single rupee on an order, right? And rupee is about a penny, more or less, right? And so to redesign your entire packaging flow to save kind of one penny per order, right, shows the kind of level of operational rigor and discipline you need to have to win in the space. But once you build a retailer, then you have the kind of scale, you have the selection, you have all these other benefits of the loyalty and the infrastructure and the cross-selling, right? On that point, we think of companies like Instacart that are marketplaces, right? I think it was 2019 that Instacart started doing advertising. Now ads are about a third of Instacart's revenue and almost all of their profit margin, right? It's almost remarkable that these companies we think are operational businesses or marketplaces are really sort of like economies of scale with some sort of cash flow machine strapped onto them. The Zepto team understood this early, and they layered in ads to their product to help people find things they're looking for. They layered in private labels to go kind of verticalize certain SKUs that were higher margin and more reliable and repeatable to sell and build. And they also tried to verticalize relationships with the farmers, actually grow the food and have their own needs on capital and logistics and so on. If you do it from the very start, you can build a cash flow. flywheel, but then it gives you the right to go to exercise your economy of scale and become a really big outcome. And so we forget this in the West because there hasn't really been a truly great retailer since Amazon, probably. Speaker 1 (16:27) I want to get back to the Zepto founders and when the world would compel to 17-year-olds that are going to Stanford to quit and start a grocery delivery service in India. But before we get into that, you mentioned this Mark Zuckerberg type story, this whole Stanford dropout or Harvard dropout. Is that overdone? I think no. Speaker 2 (16:48) And I'll tell you why. So if you look at the average unicorn startup, right, the profile looks something like, you know, some 40-something-year-old executive spends out, builds a company in the domain they know about, right? But if you weight these companies by market cap, it's a very different story, right? If you look at the college and grad school dropouts, you get Facebook, you get Google, you get Airbnb, you get Stripe, you get so many others. I think it tells a very different story. With the market cap in mind, which is really what we should all care about as investors and VCs, is just the magnitude of outcome that fits such a power law. That's all that matters. And so the more interesting question might be, why is that? Is it something innate about technology or power law talent, you could say? How identifiable is it? And what are the things you go do to go find these people early and track them into the valley? And will that continue? There was a bit of a dry spell and call it the 2019 to 2022 era where I think companies like Zepto did emerge from that. But there was no stripe that came out from that era. But then you look at the kind of 2021 to now vintages and you see companies like Cursor, one of the greatest seed investments of all time. Right. Mercore, a $30 billion company, etched $20 billion company. Right. That's very remarkable. amount of value creation started by, you know, what amounts to kind of, you know, young college dropouts. And you could say that these things are random and unpredictable. You could say that new wave of AI and software created kind of new opportunity sets for people. Or this is probably my favorite explanation. It's that COVID created this wave of people not wanting to go work a job or go to school. And it turns out that the talent was there all along. It was very tracked towards these institutions that took up a lot of your time and energy and focus. And so if people realize that you know, going to remote college was a waste of time, as did the Zepto guys, as did a lot of these other founders I mentioned. And you just say, hey, why don't I go raise a little bit of money and try to build the company? And then, you know, worst case, I go back to school when COVID is done and I can enjoy my college experience. You'll realize that a handful of these people end up a couple of years later saying, oh my God, I have a company