E60: How Lightspeed Invests in AI and Beyond with Arif Janmohamed

0:04Welcome back to Turpentine VC, the podcast where we discuss the art and science of building successful venture firms, VC to VC. Today's episode is with Arif Janmohamed, partner at Lightspeed Ventures, who joins me to discuss Lightspeed's ambitious scope and scale, as well as their frameworks for investing in AI. We cover the evolution of foundation models to the future of cybersecurity, and discuss identifying challenges and opportunities for company building. Please enjoy the episode.

0:36Arif, welcome to Turpentine VC. Excited to have you on the podcast. Yeah, thank you. So this is the first time I've had a Lightspeed partner on the show. And so I'm excited to talk about the firm. And we've had Sequoia, we've had AC &Z, we had Thrive, we had Founders Fund. A lot of the major firms in the Valley and Lightspeed is in that conversation. And so I'm curious how you think about where Lightspeed fits into the ecosystem, where it fits into the market. How should we think about Lightspeed, the firm, and where it plays? Yeah, maybe I'll start with a little bit of history on the firm, and that'll give you sort of the arc of how we've come.

1:12We were founded about 25 years ago with the insight that enterprise investing was at that point in time, generally underserved market. At that point in time, it was the height of the dot-com, and not a lot of people were really investing in the picks and shovels of the primitives of enterprise. And what I mean by enterprise, it's everything from semiconductors to primitives like compute storage, networking, security. Today would be foundation models, infrastructure software, what people used to call middleware, and then all the way up through applications. We started out with a strong focus on enterprise early stage investing.

1:49At that point in time, the founders recognized that if you knew what you were doing in terms of how to build a company, And I will talk a little bit about that, how to be opinionated with founders and how to support founders through their journey that you could actually consistently build companies worth multi-billion dollars. So the firm was really built around that insight. And over the last 25 years, we've expanded our scope. So while we started out very much focused on enterprise and some consumer, since then, we've evolved the firm into a multi-sector, multi-stage, truly global firm. And I'll touch on each of those.

2:27So we started out by focusing on enterprise, as I mentioned. And over the last 20 years, we've been really fortunate to have partnered with some amazing founders at the earliest stages, Seed Series A, Series B. These could be the founders of App Dynamics, of MuleSoft, of Nutanix, of Rubrik, which just went public, and a number of others. We've partnered with those firms at the earliest stages and then supported them through their journeys. And in many cases, these are 10, 12-year journeys with those founders as they go from building the infrastructure and the product to building out the management team to really finding their first few customers to finding product market fit and then needing more capital over time.

3:06So what we've really leaned into is having that early stage ethos and partnering with those founders and then building companies side by side. As I mentioned, we're multi-stage and we're multi-sector and multi-geo. So today, the snapshot of the firm is we can support a founder and really get into business with a founder at any stage of their journey. But we bring that early stage ethos of really understanding what it takes to build a company and partner with a founder as not just a capital partner, but also as a thought partner to help them really scale their companies. We have$7 billion of capital out of our latest set of funds that was raised a few years ago.

3:45And that is really split up across early stage, what we call early inflecting companies, and then pre-IPO companies. So we can partner at Series Seed all the way through to pre-IPO. We partner with companies and we invest in enterprise, consumer, fintech, healthcare, crypto sectors. And then from a geographic perspective, even when the founders started the firm 25 years ago, they recognized that entrepreneurship is global. Innovation can happen anywhere. And so going back over 20 years, we've been investing in Israel, in India, in China, and more recently in Europe, as well as Southeast Asia and Latin America.

4:23So when we partner with a founder, we look for those founders that really want to go along, that want to build a global brand and want to build an enduring company. It seems that Lightspeed is as ambitious as any firm out there, the A6CZs or Thrives or whoever, in terms of scope and scale of all stages, a bunch of different geos, a bunch of different sectors,$7 billion. I'm curious, what does Lightspeed think differently about the world than the Sparks or the USVs or the benchmarks that have similarly had great success, but have chosen to just stay smaller or more narrow in their approaches? Is it a belief that venture is going to get bigger as a category and there's more dollars coming in and it makes sense to scale?

5:05It's a better product to founders. How do you think about that? Yeah. At our core, we're conviction-based investors. And so So at our core, we're not the type of firm that just goes very wide like a seed firm and just tries to capture everything. But we do recognize that companies can be anywhere on a global basis, that they can that we might not hit those founders or might not find those founders at the earliest stages. And we have the capacity to partner with founders over a very long period of time. And so, as I mentioned, we're we tend to be capital partners as well as stock partners. A lot of what we do is lean into helping those founders from either the earliest stages at Seed Series A or bring that ethos that we have of company building and thought partnership to any later stage partnership with those firms.

5:51We recognize that venture capital as an industry is evolving. And from our perspective, we wanted to evolve to not only continue to invest behind the very best founders that we started the journey with, but also invest with those founders that we may not have partnered with at the earliest stages and to do so on a global basis. That's a good overview. I want to talk about AI investing. You think a lot about something your firm thinks a lot about. You guys are making a bunch of bets, a bunch of exciting bets. Talk about how you see or sort of how you and the firm approach in terms of where you're looking at, where you're excited about, where you think, hey, this doesn't quite make sense, whether it's timing, whether it's because it's too crowded, whether the price is too high.

6:35Why don't you sort of share your position? Yeah, AI is a topic du jour, but it's not new to us at Lightspeed. And for us, we've been really investing and learning about AI for almost a decade. And you can look at some of our companies that we invested in, like Series C or Series A companies like Glean, for example, or like Moveworks. These are companies that are now scaling quite rapidly. For us, AI is not new. It's been an evolution of machine learning into deep learning into what now people call AI, but it's not a brand new sector. We fundamentally believe that AI is something that will permeate into every part of what we do, whether it's at work, whether it's at play or in other ways.

7:19It is a fundamental technology and a foundational technology. And I know that's a little bit of a playoff words in terms of foundations, but AI isn't new. If you sort of go back a decade ago and you typed in A-R-I-F space J into Google, it would automatically assume that you wanted to search for R.F. John Mohamed. If you went into LinkedIn and you searched for someone who would know a priori who you were looking for, and if you go to Netflix or even seven years ago, it sort of knew what you were going to watch, but it gave you the benefit of choice, right? AI isn't new, but it's finally getting to the point from a technology perspective where it can be embedded into every type of technology.

7:58And so for us, as we look at AI today, we're really seeing this as a massive platform shift. The last major platform shift was to cloud, and there were some smaller platform shifts, obviously, before that. But this is a major platform shift. And as we think about how that platform shift will evolve across the sectors that we look at, whether it's enterprise or consumer or fintech or healthcare, it will permeate into every single piece of the stack. And so the way that we're thinking about AI is really, one, breaking down what the stack will look like on a go-forward basis at the base primitives.

8:31And then those base primitives today are revolving into what people are calling foundation models. And then surrounding those base primitives, you've got to run them, you've got to access them, you've got to manage them. you've got to secure them. So there will be security paradigms around those foundational primitives. And then there are the applications and those applications could be in healthcare, they could be in A software or what people use, vertical software SaaS. They could be in consumer technologies and Lightspeed is very much aggressively looking at all pillars of the stack. What we did internally was because of our sector focuses across enterprise and consumer, Infintech and healthcare, we have expertise across each of those pieces, each of those areas.

9:15And we brought together a cross-functional team to look at and understand how AI will really permeate across each of these pieces of the stack. So for us, it's one of these things that's just the foundational technology. We're very excited about what it's going to do. It opens up a massive camp, right? It increases addressable markets. And if you sort of look at the way that we've been investing in AI over the last year, we've put about a billion dollars to work across 62 companies. And that is at all the layers. So if we take a step back, I talked about foundation layers, right? We're investors, seed investors in Mistral, we're investors in Anthropic, but we've also looked at other types of foundation models, whether it's a robotics foundation model in a company called Skilled, whether it's a video foundation model in a company like Pika, whether it's applying music foundational models to a consumer company like Suno, we've been very much thinking broadly across the stack.

10:11And then I mentioned a few enterprise companies, but there's obviously been a number of others. Glean being one of the ones that we're very, very excited about, or Moveworks. These are touching every part of the enterprise knowledge worker, for example, and bringing co-pilots into the enterprise. Or a bridge in the healthcare space, which is very much helping doctors and medical practitioners just ease the way that they work and do their work in their daily lives. Say more about the opportunity for foundation models that are more specific in terms of what separates one that you think could be an independent company versus one that will get swallowed by some of the bigger players.

10:48I think until now at the foundation layer, it's been a combination of access to data, access to compute, and access to capital. And that is why we're seeing sort of a concentration, at least in the LLM space, around companies like OpenAI, like Anthropic, and like Mistral. In other areas, it's a combination of architectural models, software technology that glues together those models, and then vertical integration that builds applications. And so it's not just about, from our perspective, not just about access to data and compute, but as we go forward, it's also understanding verticalized domain-specific use cases, workflows, and automations.

11:32Talk about what meta recent moves mean for some of these independent players. Maybe we'll focus with the ones you're focused on, like Anthropic or like Mistraw, or more just the foundation model sort of space in general. They're pushing the envelopes, and it's fantastic to see how they're pushing the envelopes and democratizing access to a foundation model. I think you'll continue to see additional innovation come out of Anthropic and Mistral. We can't talk about that yet, but I think you'll continue to see those two companies continue to innovate. And I think Matt is really pushing the industry forward, and they'll push Anthropic and Mistral to continue to innovate.

12:10Hey, we'll continue our interview in a moment after a word from our sponsors. How deep do you go to seek out an answer to a question? Maybe you've spent hours clicking the source links on an obscure Wikipedia page, or maybe you're even the type of person who checked out the entire shelf on the topic at your library. If you're nodding along, then check out GiveWell, an organization that researches questions about global health and philanthropy, even if a satisfying answer might require years of reviewing studies, talking to experts, and over 300 footnotes. GiveWell has now spent over 17 years researching charitable organizations and only directs funding to a few of the highest impact opportunities they've found.

12:44Over 125 ,000 donors have used GiveWell to donate more than$2 billion. Rigorous evidence suggests that these donations will save over 200 ,000 lives and improve the lives of millions more. GiveWell wants as many donors as possible to make informed decisions about high-impact giving. You can find all of their research and recommendations on their site for free. You can make tax-deductible donations to their recommended funds or charities. And GiveWell doesn't take a cut. If you've never used GiveWell to donate, you can have your donation matched up to$100 before the end of the year, or as long as matching funds last.

13:17To claim your match, go to givewell.org and pick podcast and enter Econ 102 with Noah Smith and Eric Torenberg at checkout. Make sure they know that you heard about GiveWell from Econ 102 with Noah Smith and Eric Torenberg to get your donation matched. Again, that's givewell.org to donate or find out more. I'm curious how you think about the other sort of potential investment types, whether it's the application side or other sort of infrastructure, the other levels of the stack, so to speak. How do you think about that? Yeah, it's probably the most exciting time. It's a massive architectural shift.

13:54And we saw this last when companies started to move from on-premise software to the cloud. And what that meant was at that point in time, spend on a global basis was probably about one to two trillion dollars. And we're still probably four or five innings into that transformational shift. But what happened at that point in time when companies moved from on premise to cloud was you saw not only the base compute network and storage primitives start to shift over to the hyperscalers today. but you saw new foundational security models pop up and security companies pop up. Today, you can call them CrowdStrike, Palo Alto Networks.

14:33There's Zscaler, Netscope. These new companies were built for the cloud era. There was also a number of new platforms that popped up, whether it was Snowflake or Databricks or Nutanix that were compute and data-centric platforms. And then, of course, the wave of application software towards SaaS. As we step back and think about how companies are going to adopt AI, we think that not only is the entire stack, which is today about four and a half trillion dollars, so enterprise spend on a global basis, four and a half trillion dollars is going to shift to AI first workloads that have foundational models at the core of them.

15:08But now we're going to have to really reinvent all of the glue that will tie together how to run, how to monitor, how to do observability, how to make sure that these things are built reliably, how they're secured, right? There's an entire sandwich of infrastructure software and security software that needs to be built. On top of that, that will enable a new type of AI application. People are talking about work as a service, and that's going to expand the TAM, the available TAM, pretty significantly. And what I mean by that is, in that shift from on-premise to cloud, you were basically shifting workloads and you were changing business models, right?

15:46The The old way of doing things was you sort of bought enterprise software up front. You took a long time to get that enterprise software deployed. And software as a service came along. And then all of a sudden you could buy it on a monthly basis or a yearly basis. Today, it's not just about enabling users on a monthly basis with AI software. It's actually really replacing people and some of the work that we do. It's data-centric today. It tends to be pulling in a lot of data that happens to be dispersed. It's doing some manual processes. It's massaging those manual processes, and it's outputting a work.

16:24And that's just the first, the tip of the iceberg of what we're seeing. But the reality is that will then expand the addressable market for AI software significantly, probably by 10x. And there could be anywhere between$15 and$20 trillion of market cap available, which is why people are so excited about the applications of AI into first enterprise software, and then how we actually do things like searching online, watching things online, creating online, distributing content online. All of that, a lot of the manual processes around that, a lot of the generation of things can actually be replaced and can actually be automated quickly.

17:01We've started to see this in a few companies. We're lucky to have investors in a company called EvenUp, for example, which is in the legal space. And, you know, they're not selling on a per user per month basis, what they're actually selling is almost a virtual paralegal. And the value proposition to a small law firm is now you could prosecute more business on a monthly basis, you can actually increase your revenue pretty significantly, while adding these virtual AI bots, AI agents that can actually do a lot of the work of what a paralegal can do. And they do it 24 hours a day. They do it much faster.

17:38The accuracy is getting much better, et cetera, et cetera. We're investors in a company called Navan, which is in the corporate travel space. And they've infused AI into their platform. And now it can actually help people that are traveling on a global basis and do it automatically. In many cases, they don't even know that they're talking to an AI agent, but when they're stuck somewhere trying to get home, an AI agent can actually figure out the next best ticket, the next best route, do all of the manual processes that it takes to actually re-ticket and rebook and actually issue a new ticket and make that person feel like they've been treated with empathy, with care, right?

18:19Moveworks is building AI agents that help enterprise knowledge workers resolve their help desk and service desk issues. Instead of standing on the phone for 45 minutes with a human being to get their password resolved or trying to get access to enterprise software or trying to do multiple multi-step multi-process queries and actions that could be anything from hey can you go find me the the three people that i need to contact at customer x and draft me an email the moveworks bot will do that for for you in partnership and this is just scratching the surface of the way the world is going to go i'm curious there's this conversation about sort of the co-pilot world where these software will be tools to make people more efficient.

19:03And there's the other conversation around, no, actually it will just prevent the need to hire them in the first place or to have as many on staff. And that of course, those are different worlds. And so I'm curious, what is the right litmus test when we're thinking about the future for thinking about, Hey, is this going to enable someone to do their job better? and we'll have just as many of this industry or this type of job or maybe even more? Or, hey, this is actually mean that businesses can have less of those people and thus have less overhead and be much more efficient as a result. I'd say it's a little bit of both, right?

19:38It'll allow people within companies to focus on much higher value tasks. There's just a lot of people, unfortunately today, that are doing manual repetitive tasks that are prone to error, that they can only do eight hours a day and are very repetitive. AI can take over a lot of that data-centric stuff. It can generate outputs with much greater accuracy. And humans can actually come in and supervise that. It'll also, you talked about co-pilots, it'll almost give magical superpowers to humans in terms of stuff that would in many cases take 15, 20, 30 minutes to compose the right email with the right details or the right marketing posts or the mic or the right even video content, that can take 30, 60, 90 minutes.

20:27With AI, it can be output in two to three minutes, and then it's a little bit of wordsmithing. That is superpowers that you're giving to people to be much more productive and to focus on things that are much more rewarding. Every time there's been a technical inflection, I think people have been afraid of jobs being lost and people not having anything to do, whether that goes back to the automation of building cars, for example, when Henry Ford came about or even industrialization of farms. This is yet another architectural shift and technology shift for humankind that is going to allow people to focus on much higher value things.

21:00Hey, we'll continue our interview in a moment after a word from our sponsors. Hey, everyone, Eric here. In this environment, founders need to become profitable faster and do more with smaller teams, especially when it comes to engineering. That's why Sean Lanahan started Squad, a specialized global talent firm for top engineers that will seamlessly integrate with your org. Squad offers rigorously vetted top 1 % talent that will actually work hard for you every day. Their engineers work in your time zone, follow your processes, and use your tools. Squad has front-end engineers excelling in TypeScript, React, and Next.js ready to onboard to your team today.

21:37For backend, Squad engineers are experts at Node.js, Python, Java, and a range of other languages and frameworks. While it may cost more than the freelancer on Upwork billing you for 40 hours, but working only two, Squad offers premium quality at a fraction of the typical cost without the headache of assessing for skills and culture fit. Squad takes care of sourcing, legal compliance, and local HR for global talent. Increase your velocity without amping up burn. Visit choosequad.com and mention turpentine to skip the wait list. And what about the, zooming out a little bit, in terms of the boom bubble conversation, are we sort of, is it 99?

22:16Is it like, what's the right way to think about where we are in the market on the timing side. Yeah, I've got enough gray hair to actually have lived through 99. And, you know, I first came to the Valley in the 90s. And so I remember that period. I also lived through the dot com bust as well as the global financial crisis. I think that anytime that there's new technology, there is absolutely exuberance and there's a period of overinvestment. What is different today is if you sort of step back and think about the 90s, there were a lot of companies that didn't really have business models, did not necessarily have sustainable technology advantages.

22:50Today, it's different in that the technology has come a very long way. The business models are actually generating revenue, right? Some of these AI bots are just in a situation where they're actually delivering tangible value that people will pay for. And the markets here are massive, right? As I said, you're not just replacing technology with a business model shift or not just replatforming technology from one side to another. you're really changing the way that companies can work. You're changing the way that people can work and you're making us much more productive. Now, that's not to say that there'll be investment in some sectors and there'll be some level of over enthusiasm in some sectors.

23:29Ultimately, the very best companies will win, the very best founders that are focused on building the best technology, the best go-to-market engines that can very much differentiate themselves from others. Those ones will win. At the end of the day, you probably will have some money lost, but you'll also have, because of the size of the TAM and the expansion of TAM, as we've talked about, we will have some companies that will be built that are orders of magnitude bigger than the last generation of companies that are built. It's just inevitable, and that's why you see capital coming into this market, and that's why you see capital really getting behind some incredible founders.

24:06And if you sort of look back, when that last platform change that I talked about to cloud happened, you saw a number of lookalike companies pop up at the same time, right? At the same time, people were saying, hey, this is over-invested, this is too busy, but some major winners came out of those. And Lightspeed was very lucky to partner with some of those, right? But you just, you will have competition, you will have money lost, but you will have a lot of money made in this cycle. And do you think there's something fundamentally different about this wave that we can't really pattern match to previous waves?

24:38Or Or do you think the trends are helpful to pattern match? You know, in venture, you almost have to be a historian. It very much helps to understand how things played out the last time. But also, patterns are just that. Patterns sort of repeat, but they don't always repeat. And so you can't be rigid and saying, because it happened this way, it will happen this way in the future. We have to not only be historians, but we have to be constant students. And so what's really interesting in this cycle is the pace of change and the pace of innovation and the rapid, rapid innovation and competition that's happened.

25:15It's incredible. It's exciting. And it just keeps us on our toes. Yeah. In terms of opportunities, do you have a request for startups that you're excited for people to go pursue or things that you're looking to back or just maybe a sort of a type of idea or type of company that you feel is underrated or a trend that we're not paying sufficient attention to? There's been a lot of ink spilled around how AI can actually change every part of our world. I'd encourage founders to not just look at the traditional stovepipes in software, enterprise software, and to think really beyond sort of CRM, human capital management, ERP, and supply chain, etc.