Transcript: Mamoon Hamid, Kleiner Perkins

The transcript from this week’s, MiB: Mamoon Hamid, Kleiner Perkins on AI Investing, is below.

You can stream and download our full conversation, including any podcast extras, on Apple Podcasts, Spotify, YouTube (video), YouTube (audio), and Bloomberg. All of our earlier podcasts on your favorite pod hosts can be found here.

~~~

Masters in Business: Mamoon Hamid, Kleiner Perkins

Host: Barry Ritholtz  ·  Guest: Mamoon Hamid, Partner, Kleiner Perkins  ·  Bloomberg Radio

Announcer (00:00:02): Bloomberg Audio Studios. Podcasts, radio, news.

Barry Ritholtz (00:00:08): This week on the podcast, another banger. Mamoon Hamid is partner at Kleiner Perkins, where he’s been focusing on early-stage AI investments for nine years. He’s got a fascinating background — early investor in Slack, Figma, Glean, Box, et cetera. Previously, he co-founded Social Capital with Chamath and worked for a number of other venture firms, including US Venture Partners. I thought this conversation was fascinating and I think you will also. With no further ado, my conversation with Kleiner Perkins’ Mamoon Hamid. Mamoon Hamid, welcome to Bloomberg.

Mamoon Hamid (00:00:57): Thank you so much for having me, Barry.

Barry Ritholtz (00:01:00): So I’m fascinated by your background. You grow up in Frankfurt, Germany. You come to the US to go to college at Purdue — bachelor’s in electrical and computer engineering — a master’s at Stanford, an MBA from Harvard. What was the original career plan?

Mamoon Hamid (00:01:18): So let’s go back to, I think, 1986. Do you remember the Challenger explosion?

Barry Ritholtz (00:01:18): Sure. Every kid growing up remembers that.

Mamoon Hamid (00:01:28): And one of my teachers was actually supposed to go on the space shuttle, because there was a teacher—

Barry Ritholtz (00:01:36): Christa McAuliffe. That’s right.

Mamoon Hamid (00:01:36): Yeah. And every kid got fascinated, especially if you had a teacher going to space. So I followed the whole journey of the Challenger space shuttle and the teachers and all that. But with that also came this desire to learn more about space, and I instantly wanted to become an astronaut. Naturally — I think I was seven or eight years old. As I thought about high school, liking science and math, and thinking about where to go to college — and as you mentioned, I was growing up in Frankfurt, Germany — one of my uncles had given me this list of colleges, the top 10 engineering schools. I just applied to all 10. And one of them happened to be Purdue, where to this day the most astronauts have graduated from.

Barry Ritholtz (00:02:30): Really? That’s fascinating.

Mamoon Hamid (00:02:31): Yeah. So my path was aeronautical engineering and trying to figure out a way to get into space. I’ve yet to do that, but that is what led me down the path of applying to Purdue in the first place and the association with space and NASA — and then actually going from something so massive and big, space, to something so small, chips and semiconductors and transistors.

Barry Ritholtz (00:02:55): Which are enabling space, so there’s definitely a connection. Is it true that when you went to business school, you were already thinking about being a venture capitalist?

Mamoon Hamid (00:03:06): When I applied to business school — so I’d worked for a good six years after undergrad. I studied electrical and computer engineering, and that naturally made me think about a career in Silicon Valley designing chips, which is what I did for the first six years of my career, working in the semiconductor industry. But what became really interesting for me was the notion of startups and founding companies, and how these so-called venture capitalists were behind some of the most iconic companies I was coming across. So I actually wanted to get into venture capital, and that’s why I applied to business school — and specifically only applied to one business school, Harvard, because I naively thought that if you wanted to get into venture capital, you had to go to Harvard or Stanford. And I’d already gone to Stanford for grad school. So it’d be nice to get a change of scenery—

Barry Ritholtz (00:03:56): Just round it out a little bit.

Mamoon Hamid (00:03:57): And move to Boston.

Barry Ritholtz (00:03:58): Yeah — give up the nice weather. So in between college and grad school, you spent how many years at Xilinx?

Mamoon Hamid (00:04:09): I was there six years. So the story actually goes: I was 19 when I graduated from college, from Purdue, and I thought, okay, the best thing for a young kid is to continue on to grad school. So I applied and ended up getting in at Stanford. But I also got a number of job offers. This is 1997, the dot-com boom — and I’m thinking it’s a bit like this time. Should you opt out of the job market and get extremely valuable experience, or continue on with grad school? So I did the best of both worlds: I went to Stanford, took a few classes every quarter, and worked full time at Xilinx. And this is back in 1997, the middle of the dot-com boom.

Barry Ritholtz (00:04:57): Really interesting. You are known today as someone who thinks about software generally, and enterprise software in particular. That seems like an unusual transition from semiconductors. What led to that shift? What changed your thinking?

Mamoon Hamid (00:05:16): Great question, Barry. So in 2005, when I got into venture capital, my full intent was to learn how to invest in great semiconductor companies, or in founders who build the semiconductor companies of the future. It turns out that after the dot-com bust, there was not a lot of investment in infrastructure — data centers and networking and switching and semiconductors broadly. So I realized pretty early on that if I wanted to build a career in investing, you have to go where the puck is going — skate to where the puck is going. And I skated toward Web 2.0 and software. In my own firm, US Venture Partners, where I started my venture capital career as an associate, I was hired to help the partners evaluate semiconductor opportunities. That’s actually why I went there — there were some legendary semiconductor investors there, and some of my mentors even today were the folks running the firm. But I realized that all my friends in 2005 were moving to Web 2.0 and the internet. This is the beginning of Facebook, which happened to be started at Harvard when I was there. You were seeing all these people in my cohort’s age group moving into software and web, and I felt like I had to move along with that. My day job was evaluating semiconductor businesses, but in the evenings I was in San Francisco, going to the Web 2.0 parties and meeting all the founders starting software businesses. So I slowly started, as a side project — and the side project became the main project — to move from semis to software and the internet. But in the back of my mind, I always remained a semiconductor guy. And semis are back now, as you know.

Barry Ritholtz (00:07:17): And AI seems to be the application of both semis and software, so you’re well prepared. We’ll talk about AI in a bit — I want to stay in the two thousands. When you were at USVP, you had early exposure to companies like Box and Yammer — I don’t really remember Yammer, I remember Box — big enterprise software deals. What did you learn from that experience? What have you brought forward with you from that era?

Mamoon Hamid (00:07:47): So Box happened to be my first investment at USVP, where I joined the board. It was an early-stage company — a few hundred K of revenue, two very young founders, Dylan and Aaron, 20 and 21 years old, dropped out of college. Sort of the prototypical founder, right? The archetype of a young founder. And they were going after storing your files in the cloud and sharing them inside your company.

Barry Ritholtz (00:08:16): Let me stop you for a second, because I think anybody under 40 is perplexed by what you just said. I recall in the late nineties and early two thousands, whether I was at home or at work or on a laptop or at the beach house, whatever I needed was always somewhere else. And the beauty of early blogging software was that I could upload files, charts, images — that was the closest thing to the cloud. It just didn’t exist then. If you wanted something you could access anywhere you had an internet connection, it literally did not exist.

Mamoon Hamid (00:08:56): Yeah. So maybe I’ll go back to exactly the point I made in my head, which was: if there is one application that moves into the cloud first, it’s going to be file sharing. I remember this from when I was on my Windows computer in the eighties and nineties — what’s one of the applications we all used a lot? Do you remember the Windows File Explorer? We were constantly clicking in and trying to find the file—

Barry Ritholtz (00:09:23): Finding something — or searching for it.

Mamoon Hamid (00:09:24): Searching for it, or placing it in a folder very nicely.

Barry Ritholtz (00:09:28): The name you used to put on a file was important, because if you couldn’t remember the name, you couldn’t find it. It wasn’t like, here’s a phrase that’s somewhere in this document, go find it. If you didn’t remember exactly where that was nested or what name you put on it — good luck.

Mamoon Hamid (00:09:43): Good luck, right? And so the world moved from the desktop to the browser — by 2006, 2007 we’re all using Firefox, Mozilla; Chrome’s not even existent—

Barry Ritholtz (00:09:56): It was Internet Explorer until Chrome came along.

Mamoon Hamid (00:09:59): Exactly right. And so my thesis was: one of the business applications that will move into the browser — software as a service — will be file sharing and collaboration. Because, precisely to your point, the file that you always needed was somewhere else. This made so much sense to me. At the time, in 2007, when I invested in Box, there were many of these companies doing file sharing, but it was mostly for consumers.

Barry Ritholtz (00:10:32): Dropbox.

Mamoon Hamid (00:10:33): Dropbox was in that same era, but there was Xdrive and Elephant Drive. As an associate, when you’re suggesting an investment, you’re going to do a lot of diligence — I remember the laundry list of companies I looked at. There were probably 40 companies doing something similar, but most of them were dedicated toward consumer use cases — photos, music, stuff like that.

Barry Ritholtz (00:10:56): The Napster era was right around then.

Mamoon Hamid (00:10:58): Exactly. And so the hypothesis was: this stuff will be relevant to large companies, who will want file sharing and collaboration for their companies. And Box actually pivoted from being a consumer company to being an enterprise company. That’s when I got pretty excited, because it lined up with this view I had that large companies would move from file servers in their data centers, or wherever in their buildings, to files that reside in the cloud.

Barry Ritholtz (00:11:28): And they’re willing to pay for it.

Mamoon Hamid (00:11:30): They’re willing to pay for it.

Barry Ritholtz (00:11:31): Unlike back then, when consumers were so reluctant to pay for anything. So it’s interesting, because you’ve had a lot of early investment success with a variety of companies. Is it easy or difficult to learn from past winners? Is every startup different, or do you develop a little pattern recognition that gives you some clues — hey, these guys are onto something?

Mamoon Hamid (00:11:58): Yeah, I think there’s definitely some compounding of learning from early wins and losses. You brought up Box — my first investment. I got to spend a lot of time with the founders, got to learn the business with the founders, actually, because they were young and I was young — I was in my twenties when I joined the board. From that experience, I learned a lot about what it meant to sell software bottoms-up into large companies, which led me to the investment in Yammer — which many folks may not remember, but it was an enterprise social network circa 2010, kind of like Twitter meets Facebook, but for your company. Microsoft ended up acquiring it in 2012, and it’s part of the Microsoft suite now, part of Teams and all that. But the experience of that bottoms-up adoption — you looked at Yammer, and a lot of large companies wanted an enterprise social network, kind of like a town hall, a messaging platform. You share a file, people comment on it, like people do on Twitter or Facebook. That was taking some of the learnings from the web era and the social era and applying them to the business world. So at Yammer I learned a lot. It was a quick journey, but there was a level of engagement and monetization that was good. And a year or so later, I came across another company: Slack.

Barry Ritholtz (00:13:36): Slack, yeah.

Mamoon Hamid (00:13:37): And it was sort of similar — now it was true messaging. Actually, on my way over here, I was walking through and saw a bunch of colleagues on Slack, which makes me really happy. Slack is used broadly across the globe to this day in 2026. The lessons learned in that 2010-11 era led to the investment in Slack in 2014, for me, when it was a 10-person company. So back to the point about compounding of learning: Box led to Yammer, Yammer led to Slack, and since Slack there have been others. But there certainly is some pattern recognition around products that are working.

Barry Ritholtz (00:14:17): And then in 2011, you co-founded Social Capital with Chamath — a very famous model, which was: how can we address many of the social ills that are hurting the country through the intelligent use of startups, technology, et cetera. Was this a reinvention of venture capital, or just a new set of tools within a partnership? It was kind of novel for its era.

Mamoon Hamid (00:14:50): It was novel for its era, because we decided we would go after education, healthcare and finance — three of the largest parts of society — and address inequities in those areas with our investments, believing that technology has the ability to democratize access to healthcare, education and financial services. And that’s largely played out over the last 15 years. We just thought there would be a ton of opportunity — as venture capitalists, we’re seeking out opportunity, and we thought that by going after large pockets of GDP, you’d identify really exciting opportunities. It turns out my interests remained in enterprise software, and I spent a lot of my time in enterprise software even when we started Social Capital.

Barry Ritholtz (00:15:46): So in 2017, you leave Social Capital for Kleiner Perkins, a firm that has long been iconic. The laundry list of companies Kleiner has backed: Google, Cisco — were they early in Apple also?

Mamoon Hamid (00:16:00): I think they were. Google, Amazon, Sun Microsystems, Genentech, Intuit — unbelievable. Tandem, if you remember. The list of greats is amazing.

Barry Ritholtz (00:16:12): So they’re an iconic company, but they’re not the dominant force they once were. When you joined, what made that opportunity so attractive?

Mamoon Hamid (00:16:22): So look, I had long admired Kleiner Perkins — an extreme reverence, I would call it — going back to my days moving to Silicon Valley. I mentioned I moved to Silicon Valley in 1997 and worked for this company, Xilinx. Think about my first few weeks on the job: I’m in my cubicle, and I’ve got a Sun Microsystems workstation, which was actually a dream, because in college we had to share 20 of them among 2,000 of us. Now I have my own Sun SPARC — I think it was a SPARC 20. And guess what? There’s a Netscape browser, and I’m buying books for grad school on Amazon. And by the way, there are a couple of guys down the hallway at Stanford who are starting this company called Google, and I’m starting to use that search engine. The one commonality among Xilinx, Sun, Netscape, Google and Amazon is that Kleiner Perkins had led the Series A — was the first institutional investor — for all five of those companies. So as a young guy, I developed this extreme reverence for Kleiner Perkins because of the investments they had made in these history-making companies. Which is also what led me to think about venture capital as a career as a young engineer — I wanted to be like those guys; they were investing in all the coolest companies I was using as a 19-year-old. And one of the people behind many of those investments — Sun, Netscape, Google and Amazon — was my partner, John Doerr. So for me, there’s this extreme reverence for John Doerr and his career, and trying to emulate that. He was an electrical engineer from Rice, went to Harvard Business School, worked at Intel Corporation, then came to Kleiner Perkins out of business school. It had a deep meaning to me. And truth be told, in my business school essay — which I still have — I wrote that I wanted to go work at Kleiner Perkins. That was 2002, when I wrote the essay. Then, when I tried to apply for a job in 2005 coming out of business school, I didn’t get very far. But I did end up there in 2017.

Barry Ritholtz (00:18:45): So eventually, if you keep plugging away, you get to where you want to go. That’s great. Coming up, we continue our conversation with Mamoon Hamid, Kleiner Perkins managing member, talking about the reboot of the firm. I’m Barry Ritholtz. You’re listening to Masters in Business on Bloomberg Radio.

Barry Ritholtz (00:19:07): I’m Barry Ritholtz. You are listening to Masters in Business on Bloomberg Radio. My extra special guest this week is Mamoon Hamid. He is managing member and general partner at Kleiner Perkins, where he is pivoting the firm toward early investments in software, artificial intelligence and automation. So, you co-led the refounding of Kleiner Perkins in 2017-2018. The firm was refocused on early-stage, Series A investing. Tell us what was behind the thought process. What made you say: we don’t want to be bigger, we want to be smaller and more focused?

Mamoon Hamid (00:19:49): If I look back at the decades of Kleiner being probably the most successful venture capital firm throughout the seventies, eighties, nineties, and even the early two thousands, the one thing that defined Kleiner Perkins was that it was a small partnership of seven-ish or so partners who sat around a table in Menlo Park, meeting companies and having healthy discourse and debate about which companies to invest in and what the future of technology would bring to the world. It was defined by a small group of partners who were in many cases technical. They were operators. They had a passion for technology and its impact on humanity. That was what I kept coming back to — that was what defined Kleiner’s decades of success. And we went back to the future, in 2017 and ’18, to that model. Today our partnership is six partners, and we have three more investment professionals. We’re a very small, nimble team. We have two funds, and this team invests from both pools of capital.

Barry Ritholtz (00:21:11): So early stage — is that the seed round, or is it a little more developed?

Mamoon Hamid (00:21:17): The early-stage fund is seed and Series A mostly, and maybe some Bs. And the growth fund is Bs and Cs all the way to — we invested in the last Anthropic round at a $900 billion valuation, which is rare, for special companies. But it has the ability to invest across stages, even to the pre-IPO round.

Barry Ritholtz (00:21:41): And is it a coincidence that the growth fund is two and a half times the size of the seed fund? At that point these companies are bigger and require a bigger check — or is that just happenstance?

Mamoon Hamid (00:21:54): I think our funds are sized based on the opportunity set in front of us. Our early-stage funds have been almost exactly 35 companies for the last 15 years. So 35 companies per fund, which we think of as the right number of shots on goal for an early-stage fund to return multiples.

Barry Ritholtz (00:22:17): Like 25 to 30 million per?

Mamoon Hamid (00:22:20): Exactly. So it starts out with maybe, in some cases, a $5 million check, and the subsequent checks are another 15 or 20. Or the first check could be 30, and then with pro rata you’re investing, let’s say, up to $40 million per company. And then your growth fund is doubling down, investing a lot more in those companies.

Barry Ritholtz (00:22:39): Really kind of interesting. So the focus is artificial intelligence startups across the software, healthcare, transportation and autonomy industries. Let’s unpack that, because I’m hearing a little overlap with each of your prior venture experiences. Tell us why those four areas are so attractive.

Mamoon Hamid (00:23:04): This is a truly once-in-a-lifetime revolution that we’re going through with AI, and the number of exciting companies and people that we’re seeing right now is at an all-time high. The whole world, in some ways, is being refactored with AI — and this is just the very beginning. So I would say all parts of the economy, even beyond those four areas. Like I mentioned earlier: healthcare, financial services, all sorts of knowledge work. It’s going to be all sorts of physical automation in terms of robotics. Even space, even defense, drug discovery, materials discovery. I think it is all fair game at this point in terms of where the exciting pockets of innovation are, because there has never been a tailwind like this — one that allows all parts of the world to be refactored based on the biggest technological revolution ever.

Barry Ritholtz (00:24:30): I’m glad you described it that way, because I keep hearing people compare AI to the internet, and that seems too contained, too timid. I wonder if you agree with the thought that the only thing remotely comparable to this is the industrial revolution — which, centuries later, we are still dealing with the impact of.

Mamoon Hamid (00:24:54): I absolutely agree with you, Barry. It is like the industrial revolution. It’s like the railroads. It’s like the printing press. It is that. It’s not the internet.

Barry Ritholtz (00:25:04): That’s really interesting, because when I think internet, the first thing you think of is: oh, this is a bubble and this is going to collapse. But you mentioned you’re an investor in Anthropic — these are, forget not-profitable companies, these are companies with giant revenue streams already, and they’re barely a few years old. How big can this sector get? Is this going to take over every corner of the economy?

Mamoon Hamid (00:25:31): Let’s talk about that. I think that’s the real conversation — the very exciting conversation one can have about this topic. I’ll start at a very high level. The GDP of the world today is about $120 trillion, and about half of that is labor — the labor component. So roughly $60 trillion. And of that $60 trillion, roughly 60 percent or so is white collar — maybe 50 to 60 percent, somewhere in there. So that’s anywhere from $30 to 35 trillion. And if you look at tokens, and what the frontier model companies provide, it is units of labor. We’re already seeing how those units of labor are being utilized in computer science — software development — in law, in medicine, in drug discovery. These are little agents and buddies that we as humans now have to help us do more with our intellect. The way I see it, we’re talking about trillions of dollars opening up for these companies. Exhibit A is a company like Anthropic, which, as it has publicly stated, has gone from zero to a $45 billion revenue run rate — that’s unbelievable — in a matter of less than three years. And that is likely going to double. The company started this year, I believe, at 20, and has already more than doubled in the short year we’ve been in so far. These numbers are astounding, not only because these companies are selling software or technology — they’re selling units of labor. And the labor markets, as we all know, are the biggest component of the world’s GDP. We’re talking about trillions of dollars of opportunity. That’s what excites us so much about this time: it’s not just about selling tools and software that we’ve been accustomed to selling to IT departments. It’s selling actual labor — to companies, to corporations, even to consumers who are using AI in their personal lives.

Barry Ritholtz (00:28:01): So let’s talk a little bit about that. The fear I keep hearing is that everybody’s going to lose their job. It’s a very Malthusian argument — that this technology is going to replace labor the way the steam engine did. I’m getting a sense from the data, and from analysts like Torsten Slok, that this isn’t a replacement for white-collar labor, it’s an enhancement — or at least that’s the argument. Give us your perspective on that.

Mamoon Hamid (00:28:33): I actually fully agree with the point of view that you have. It’s like getting email. When we got the computer, the people who were using the typewriter started using the computer and started doing other types of jobs. Even in the steam era, the industrial revolution, we found ways to repurpose jobs and people and their skills. I don’t think humans are going out of style. I don’t think the world is going to be largely unemployed and on UBI because we’ve displaced all this work and all these people with AI. That’s the extreme where the mind goes, but it’s just not the reality. And that bears itself out in the numbers we see — record low unemployment rates. Actually, we need more labor, more people, than we ever have.

Barry Ritholtz (00:29:30): More skilled labor. We are seeing a decrease in job availability for kids right out of college, for unskilled labor. If anything, is this likely to force more people to get more technical, to up their skill set?

Mamoon Hamid (00:29:49): I actually do believe that. It is not: oh, you don’t need to be a software developer and study CS anymore because these software jobs are going away. It’s that now the job of a software engineer is to manage a whole host of agents, make them do work for them, and be the brains behind the operation. Think of it as having all these little agents — little employees — working on your behalf. That is the higher-level thinking. When we go to school, we learn how to problem-solve. If we’re solving a math problem and it’s hard, we think about many different ways to solve it. The same applies here: how am I going to use AI to help me solve problems? We have four kids, believe it or not, and I’m telling them: go into math, science, and actually art — have spectral diversity in your learning — because the skills that mattered in the past, solving math problems with pen and paper, will really matter in the future.

Barry Ritholtz (00:30:55): So let’s bring this back to how you think about the opportunity set that’s out there at Kleiner Perkins. You do structured reviews of every interesting deal that was passed on. I’m kind of fascinated by that. I know a lot of VCs hold their misses as a badge of honor — some firms post them on their website. What’s driving the thought process around revisiting missed deals or mistakes? What does the process teach you?

Mamoon Hamid (00:31:35): One of the things I did when I got to Kleiner Perkins in 2017 was that we would look every week at that week’s Series As that got done by our peer firms — about 30 or 40 firms — and whether we had seen the company that was invested in or not. A simple heuristic: are we seeing the things that matter — and they seem to matter because our peer firms invested in those companies. We’ve been doing this now for the last nine years. Initially our goal was that we should see 60 percent — and seeing means you met the company. For us, that now hovers around 70 percent or so. You don’t want it to be a hundred percent, because then that’s just the game you’re playing — we just see everything. And you don’t want to be at 20 percent, because you’re not seeing enough. We think 70 percent is a good number. And then we look at it: if we saw 70 percent of the good stuff, was the better stuff in the 30 percent we didn’t see? Or, of the 70 percent we saw, did we pass on the good stuff and do the bad stuff? We go through that exercise quite frequently. We just had an offsite a few weeks ago, and we went through it again — we pour salt on the wounds. Okay, we saw these companies and we passed. Why did we pass at that early-stage round? Just to remind ourselves why we need to adjust the way we do things. I’ll give you an example: Anthropic, which is an incredible company. The Series A was not a traditional one — SBF from FTX led the Series A, famously. We know how that worked out.

Barry Ritholtz (00:33:24): We do know how that worked out.

Mamoon Hamid (00:33:24): It was an amazing investment and very prescient on his behalf. I forget how much it would be worth today, but it would be worth a lot. But the Series B was a sort of non-consensus, non-obvious round. We actually met with the founders — but we met them over Zoom, and we played with the product. You don’t get the same visceral feeling about a company and the founders and their ambitions and aspirations and what they’re trying to do with their company. So — what a miss, right? I think that round got done at a $4 billion valuation or something, which is not a small valuation, but still—

Barry Ritholtz (00:34:08): Compared to today.

Mamoon Hamid (00:34:09): Compared to today. And we looked at a lot of our passes that were good companies, and many times we just didn’t meet them in person. The pandemic era bred some really bad habits — you did a first meeting over Zoom. I looked at my own investments, and for the last 20 investments I’ve done, the first meeting was always in person. So I’ve driven my whole calendar to: I just don’t want to do any meetings on Zoom anymore. I want to meet people in person, and if it’s worth a 30-minute Zoom, it should be worth a 30-minute in-person meeting. And I’m glad we get to do this in person — big difference. It wouldn’t be the same thing if I were on a Zoom screen doing this with you.

Barry Ritholtz (00:34:59): That’s exactly right. What about the reverse of that? If you’re analyzing your misses, do you ever review your hits, your wins, and ask: why did we get this right? What can we take forward from this?

Mamoon Hamid (00:35:14): That’s a great question. What we do look at is: is the portfolio that we built better than the portfolio that we missed? And it’s actually a toss-up.

Barry Ritholtz (00:35:32): Huh — that’s really interesting.

Mamoon Hamid (00:35:34): In the sense that it would be bad if we didn’t see companies and that basket of companies was way better than the companies we saw. We try to be very intellectually honest about whether we’re seeing the right stuff or the wrong stuff — and it turns out it’s a toss-up.

Barry Ritholtz (00:35:56): The reason I ask that question is that in the public markets, you learn more from your misses than your wins, because it’s very hard to tell the difference between skill and luck in the public markets. I’m curious if the same sort of thing applies to venture.

Mamoon Hamid (00:36:15): I think it does. You really beat yourself up on the misses, and on the ones you did do, you’re just like: okay, check, it happened — and maybe you don’t think about them as much as the ones you missed.

Barry Ritholtz (00:36:29): Really, really interesting. Coming up, we continue our conversation with Mamoon Hamid, partner at Kleiner Perkins, discussing the state of venture investing today. I’m Barry Ritholtz. You are listening to Masters in Business on Bloomberg Radio.

Barry Ritholtz (00:36:46): I’m Barry Ritholtz. You are listening to Masters in Business on Bloomberg Radio. Mamoon Hamid is my extra special guest. He is a partner at Kleiner Perkins, where he’s driving the firm’s focus on early-stage investments in artificial intelligence and related technology. So let’s talk a little bit about the state of venture investing today. When you meet a founder for the first time — preferably in person — what are you looking for? What are you trying to spot that isn’t in the pitch deck they sent earlier? How do you separate intensity from delusion? What are you trying to identify in that first meeting?

Mamoon Hamid (00:37:32): It is the most interesting time in my venture career, for obvious reasons — this AI revolution. The tailwinds that AI brings to companies being started today are unlike anything we’ve seen before. So the quality of ideas is at an all-time high, and I would say the quality of the people is also at an all-time high. Those two forces combined mean there are a lot of really high-quality ideas and founders that we’re seeing — and the volume, because of that, is at an all-time high. To your question of what we’re looking for: obviously everything today is AI-enabled, with AI tailwinds — and venture capital isn’t venture capital unless there’s a strong tailwind of something. That makes this such an exciting time, because the winds are so strong — in some ways because the frontier model companies are providing better and better models, which allow companies that build on those models to provide a strong value proposition. Take, for example, a company like Harvey, which is AI for legal. It’s in most of the Am Law 100, sells to large enterprises and Fortune 100-type companies, and is becoming the de facto way that law firms and legal departments are using AI. And yes, they have a lot of secret sauce on top of what the model companies provide, but at the same time, they get better and better as these models get better and better. This is a category that just got created in the last few years, and there are numerous examples like it in every sector of society. Just take a step back — I don’t know what your reaction was to ChatGPT when you first saw it, but for me, it was the same reaction I had when I first got to use an internet browser. That Netscape moment was equivalent to the ChatGPT moment. And we started to think about the second-order effects — what are the things that come from here, from ChatGPT, from LLMs? We thought about the labor pyramid — think about knowledge work. We went straight to: what are the most highly skilled, highly paid jobs in the country? If you look at the top 20 list of jobs by dollars earned, it’s some form of engineer, doctor or lawyer. So what we did at Kleiner Perkins was invest in companies that do AI for legal, AI for software development, AI for medicine. We invested in companies like Harvey; Windsurf, which got acquired by Google; and companies called Ambience and OpenEvidence — to go after the top of that labor pyramid. And back to your point, it’s an enhancement for those people. It was supposed to be like a copilot — which it is—

Barry Ritholtz (00:41:04): Registered trademark, Microsoft Corporation.

Mamoon Hamid (00:41:07): Yeah, exactly — Copilot. And in some cases it’s become an autonomous agent for those people. Software developers run agents to do work for them now. And so we worked our way down the labor pyramid a bit — what’s the next level? If doctors, lawyers and engineers are in the $200K-plus-a-year pay range, what’s one level below that? You’ve got financial analysts, salespeople, nurses. So we’ve invested in that next class of companies: Rogo, which is AI for finance; Hippocratic, which is an agentic nursing platform; Nooks and Revo, which are helping salespeople — copilots for their work. So we’ve worked our way down the labor pyramid. And if you think about the pyramid, guess what’s at the bottom? Physical labor — the lowest paid, lowest skilled in some cases. Eventually robotics will get there, and hopefully do a lot of the backbreaking work that nobody should be doing. That goes along with the timeline of how we’re thinking about these things: you start with the highly skilled, highly paid, and over the next decade you get down to the lowest-skilled, lowest-paid work.

Barry Ritholtz (00:42:31): So you mentioned the tailwind behind AI. What does that do to valuations? How do you underwrite startups in a market where even the half-decent companies look kind of expensive?

Mamoon Hamid (00:42:46): Great question. When you have companies that have gone from zero to a trillion dollars in value in the last three to four years — when you’re a founder, what do you do? You point to that. That’s what I can be; I can get to a trillion dollars, because I’m going after really large problems. And we love that ambition. What do people do then? They say: well, if the probability of a $1 trillion outcome is 1 percent, and you add up the probabilities for all the other outcomes, you’ve got yourself an expected value of $10 billion or more — at least that’s the minimum. You do the 1 percent of $1 trillion, that’s $10 billion, and you add up all the other probabilities, and it’s probably like 30, 40, 50 billion. So if you’re looking at a Series A company where the ambition is tremendous, the valuation may just be — you’re raising a hundred million at a billion-dollar valuation for two guys out of the gate. And those are real examples. I think we’ve got a bit of this pointing at the large outcome — and if it’s paired with smart people and an ambitious idea, it has that kind of potential. But the reality is there are usually only one or two of those real outliers that are Mag Seven scale. Gravity in itself doesn’t allow for the world to have a hundred trillion-dollar companies.

Barry Ritholtz (00:44:28): Plus, if it’s a trillion-dollar total addressable market, it’s going to attract a lot of competition, a lot of other startups — perhaps more than if you’re focusing on a smaller niche. Although even in cloud storage, you mentioned 40 companies in the early days. There have got to be tens of thousands of companies going after each of these segments in AI.

Mamoon Hamid (00:44:53): Yeah, there are. And the hard part of the work we have is identifying what we think will be the leading company, because in technology, generally speaking, the leading company gets most of the market cap — which, if you look at the Mag Seven—

Barry Ritholtz (00:45:10): Winner take all.

Mamoon Hamid (00:45:11): Winner take most, right? They take 90 percent of the market. Nvidia — 95 percent of all GPU spend. Take Google, Alphabet — multiple businesses. Or take a Tesla; take a Meta — owns social. So winner takes most in technology.

Barry Ritholtz (00:45:30): Really interesting. You’ve described the AI moment as one of the most important company-building opportunities in our lifetime. What’s the risk of venture funding too many of these startups? Or is this just a fat head of winners and a long tail of well-we-gave-it-a-shot? Is this simply the nature of this business, where a couple of winners drive all the returns for your funds?

Mamoon Hamid (00:45:59): Venture is a power-law business — there’s no question about it. There’s this article I read the other day where 90 percent of the AI revenue is in the hands of two companies.

Barry Ritholtz (00:46:13): That’s crazy.

Mamoon Hamid (00:46:14): OpenAI and Anthropic. The remaining 10 percent is in a bunch of companies that are excellent companies, but the scale of those two companies is so massive that it dwarfs all the other great work happening in tons of other companies that we’ve backed. And by the way, those will be great outcomes and great companies — there’s no question about it. It’s just that the power law really shows itself when you look at the numbers. Yes, our job is to be in the companies that make history and follow the power law. What we find is that you would think there are dozens of companies — there are actually tens, maybe fewer than 10, and there are two to three that are converging to become the winners in a category. Everyone’s trying to identify those two to three companies and be in them. One of the challenges we face is that we try to invest early in what we think is the winner in the category. Sometimes you don’t know early that this is the winner, and if you invest too early, you conflict yourself out of the eventual winner.

Barry Ritholtz (00:47:28): You only invest in one company per silo, so to speak? Just to avoid those sorts of conflicts.

Mamoon Hamid (00:47:35): Yes. Typically we’re joining the boards of these companies, and if you have confidential board-level information and you then invest in a competitor, all of a sudden there’s a chance of a conflict of interest. So we definitely try to avoid that.

Barry Ritholtz (00:47:51): So I’m kind of fascinated by this: as venture investors, you obviously see the promise of AI across all these different economic sectors. I’m curious how you are using AI internally at Kleiner Perkins. Are you using it to source deals, or do due diligence, or predict specific outcomes? How does AI fit into your operations?

Mamoon Hamid (00:48:17): We have definitely been maxing out on AI internally — not only because we invest in these companies. Glean is a company we incubated inside of Kleiner Perkins, actually — it’s in year seven now, a pre-AI company started by an incredible engineer, Arvind Jain. That’s our knowledge management. Every single piece of knowledge inside Kleiner Perkins resides in Glean, and you can go to it, query it, chat with it. If I want to find your phone number and email — say I’ve never met you before, but I know someone at Kleiner probably knows you — I’ll go to Glean. If I want to ask about an HR policy, I’ll go to Glean and quickly ask, because it just knows everything about Kleiner Perkins. It knows investment memos, it knows cap tables — it knows the really confidential stuff. It’s permissioned in a way where the people who are supposed to know can know. That’s part of the magic: it’s very safe and secure, and you trust it to know the things it’s supposed to know and not know the things it’s not supposed to know. So that’s one example. But we also get so much information — board decks, long board memos, financials. I’m going to a board meeting after this, and I got the board memo. The first thing I do is send it to an email alias that runs it through AI, and it produces a summary of the board meeting and questions I should be thinking about — an instant step I take once I get the board materials, so that I start thinking about it before I actually go read the board memo. I sort of have a preview of it in my mind. We do a portfolio review every four months or so, and we have a couple hundred companies. It used to be a very manual process — we had a dedicated person working on it. Now our technology team has built a system where we take these summaries and they get piped into this portfolio book that we create, with all the financials and all the metrics. We’re heavily leveraging AI there — in this case it’s Glean and Claude, the underlying models, obviously. And we’ve done a bunch of other things. I actually love to rate my meetings, just so I remember the tens and the nines that I should have paid attention to but forgot. I want to have an exhaust of all the things I’m encountering in my real life, and AI is an amazing capture of that exhaust — providing intelligence and signals to our team, piping it into our CRM. There are all these cool things we’ve done. We have an internal tech team — an amazing team of four folks who build a lot of these tools — and we are definitely maxing out on using everything that’s out there.

Barry Ritholtz (00:51:20): Really quite fascinating. So, final question before we get to our favorites that we ask all of our guests: what are investors not thinking about when it comes to AI — or anything else — that perhaps they should be? What sort of topics — policy, data, geography — what’s getting overlooked but shouldn’t be?

Mamoon Hamid (00:51:42): I think right now we’re going through a time where software is considered to be dead — they call it the SaaS apocalypse, right?

Barry Ritholtz (00:51:53): Although they’re just coming off their lows.

Mamoon Hamid (00:51:55): Yeah. And there’s always an overreaction: oh my God, it’s going to be an AI capex world, and only chips will matter — only fiber and data centers and power will matter. The reality is that the way we as humans interact with technology is through software — through the things we have known as software and tools. And by the way, CIOs in large companies buy from companies that sell to them; they don’t just buy a smart agent. So I think the pendulum has swung a little too far, and we underappreciate what software still does and will continue to do forever — for our enterprises, for governments, et cetera.

Barry Ritholtz (00:52:41): Software: not going away. Really interesting. All right, let’s jump to our favorite questions, starting with: who are your mentors who helped shape your career?

Mamoon Hamid (00:52:51): One of my mentors is Irwin Federman, whom I dearly love. He’s 90 years old now. He’s a New Yorker — he sold peanuts, I believe, at Dodger Stadium when the Dodgers were still there in the fifties. He was my mentor at USVP, and he’s a legendary semiconductor investor, believe it or not. He was one of the co-founders of SanDisk Corporation, and before that he was CEO of Monolithic Memories.

Barry Ritholtz (00:53:22): I hope he still has a few shares of those.

Mamoon Hamid (00:53:24): He probably does. And SanDisk is in this memory hype cycle we’re going through — hype or not, there’s a real need for memory. I believe it’s now maybe a half-a-trillion-dollar market cap company. Something crazy — don’t quote me on that, but memory is having its day right now. In any case, he was a mentor because not only did I work for him, but I saw through his lens how to be a great board member, how to make investments, how to back people, how to build relationships with people. He also gave me feedback that no one else in life would, because I think he loved me — I really felt the love — because the kind of feedback he gave me was feedback that I don’t think people would generally have the courage to give you. It’s pretty direct — tough love. And I love that about him. It reminds me that I need to go pay him a visit.

Barry Ritholtz (00:54:25): Well, you’re in New York — you might as well.

Mamoon Hamid (00:54:27): Oh — no, he actually lives in the Bay Area now. He’s a New Yorker who relocated to the Bay Area, I think, 50 years ago.

Barry Ritholtz (00:54:34): So let’s talk about books. What are some of your favorites, and what are you reading currently?

Mamoon Hamid (00:54:38): I’m just starting on this book called Believe — why you should believe, especially in this era of AI. I’m on the board of a company called Thrive Global, whose CEO and founder is Arianna Huffington. Arianna gave me the book. She and I have very aligned views on faith and spirituality — Arianna, I believe, is Greek Orthodox, and I’m Muslim — and we talk about how faith guides our lives, and about this age of AI. That’s actually the conversation she and I have. She said: I have the perfect book for you. It’s about how we should believe even more in this age of AI, because it helps us understand the world — we’re trying to make sense of it all the time. Religion can give us a bit more of a constrained view of what the world actually is, because otherwise it’s just a black box; you won’t be able to comprehend the vastness of what we’re trying to comprehend as human beings. And especially with AI, we’re pushing the boundaries of what’s possible. I think there’s actually more in the scriptures than you’d think — that’s the view she and I share, and this book hits home with it. It’s by the New York Times columnist Ross Douthat. Believe — check it out.

Barry Ritholtz (00:56:16): On my list now. What about streaming? I know you host a podcast. What do you either watch or listen to these days?

Mamoon Hamid (00:56:25): My wife and I love to watch Dateline, 48 Hours — these crime shows.

Barry Ritholtz (00:56:31): All the crime shows.

Mamoon Hamid (00:56:31): All the crime shows. In some ways it just takes things down a notch, but it’s also so instructive about human psychology — what motivates people to do not-so-great things. And there’s generally a theme to it at this point; it’s a very repetitive theme.

Barry Ritholtz (00:56:52): It’s Dunning-Kruger. They have no idea about the trail of DNA evidence they leave everywhere. Anytime I’ve watched that show, it’s like — what are you doing?

Mamoon Hamid (00:57:01): And it should be that it’s harder and harder to commit crimes.

Barry Ritholtz (00:57:07): And yet—

Mamoon Hamid (00:57:08): There are still crimes.

Barry Ritholtz (00:57:10): Still — and just as many as ever. Only people are getting caught more easily.

Mamoon Hamid (00:57:13): Yeah. And the cell phones — they’re pinging those towers.

Barry Ritholtz (00:57:18): What do you mean you weren’t in the house? We can tell you were within a hundred feet of this person at that time.

Mamoon Hamid (00:57:23): Exactly. So it’s probably not a very full answer — you’re probably looking for some cool show that I watched.

Barry Ritholtz (00:57:30): No, not at all — I’m fascinated by that. It’s funny, because there used to be this giant gap between the CSIs and what was actually going on. But if you watch the two of them, it’s really closed. Maybe there’s a little selection bias here, because all those shows are about the people who got caught — so you’re seeing where the technology worked, where the forensic science got that guy. It’s really very funny. All right, our final two questions. What sort of advice would you give to a recent college grad interested in a career in either venture investing or technology?

Mamoon Hamid (00:58:12): It’s probably the same advice I would have given 20 years ago, or given myself coming out of college: go work at a fast-growing company, where you can learn from the growth it’s encountering, but also from the people — and build the network that you’ll have for the rest of your life. It’s probably the best time in your life, coming out of college, to learn from others around you and to experience high growth, because from high growth, lots of lessons are learned. If you can find a way to get into a high-growth technology startup — an AI startup — it is the best way to develop the skills, but also the empathy of what it means to have carried a bag and sold something, and built something, and shipped something. I always tell folks there shouldn’t be a direct path into venture capital. It’s a second thing — it’s not the first thing you do coming out of college or grad school. You have to have built and shipped and sold, and developed that empathy for high growth and the lessons learned, before you get into our career.

Barry Ritholtz (00:59:26): Really, really interesting answer. And our final question: what do you know about the world of venture investing or technology today that might have been useful 25 or 30 years ago, when you were first ramping up?

Mamoon Hamid (00:59:40): It’s all about the people. It sounds so trite, but it is. I say: ordinary-looking people doing extraordinary things. And how do you assess those ordinary people who are doing extraordinary things? By really understanding the people — their intentionality, their desires, their ambition, what drives them, their motivation. Which goes back to why you meet them in person: you’re trying to figure out why they are doing this. Building a startup, a company, is hard work. It’s a sacrifice on life. So there had better be a good reason why you’re doing it.

Barry Ritholtz (01:00:18): Really interesting answer. Thank you, Mamoon, for being so generous with your time. We have been speaking with Mamoon Hamid. He is partner at Kleiner Perkins. If you enjoyed this conversation, well, check out any of the 647 we’ve done over the past 12 years. You can find those at Apple, Spotify, YouTube, Bloomberg — wherever you get your favorite podcasts. I would be remiss if I did not thank the crack staff that helps put these conversations together each week: Alexis Noriega is my video producer, Sean Russo is my researcher, Anna Luke is my podcast producer. I’m Barry Ritholtz. You’ve been listening to Masters in Business on Bloomberg Radio.

~~~

The post Transcript: Mamoon Hamid, Kleiner Perkins appeared first on The Big Picture.