Do Margins Matter in AI?

Harry, you big stud. It is great to be back. It is awesome to see

you. It has been too long, but it is amazing to see what this thing is built into. So congrats and thrilled

to be with you today, my man. Do you know what? I’m as surprised as everyone else, to be honest.

I mean, this AI stuff is just awesome. I’m a tech geek at heart. We all look a little taller and sound a little smarter when there’s an uptrend in the market, but this one’s different. Like, this is this is going to be the type of thing that we tell our grandkids about and that generations talk about this transitional moment. It’s absolutely awesome to be part of it. I think great businesses will be built and money will be made, but just from the technology side, what we’re going through is so damn cool to see, and I’m just loving it.

It’s neat to see mind blowing demos again, and to be part of discussions of what can be, and things that you couldn’t have conceived of a few years ago. And so it was tough. It felt like a steady gut punch coming out of the 2000s, and with the market pullback, and people questioning tech and so many board meetings doing layoffs and just having to survive. And it’s just awesome to be back on offense again.

So I would phrase it a little differently. I would just say that the pace of innovation is incredibly compressed right now. The best teams are using that for their advantage, iterating at mind boggling rates. The marginal companies are getting passed faster than ever. And we’re going to keep seeing that because the tech and the enabling tech is so damn good. I don’t worry about commoditization in the sense of price erosion, which is often implied. It’s often used as a derogatory term, but sure, you can think of perhaps foundation models as commodities in the way that hyperscalers are.

And by the way, the best business in the history of software is sitting there with AWS in what people refer to as a commodity. And so I think the same playbook is going to be run-in the foundation models. I think that the layers on top of those models are going to extract phenomenal value because they’re gonna deliver phenomenal value. And I think we’re gonna see great businesses built at multiple layers in the stack.

So it was interesting how you worded the question. How do you think about margins when investing today? And I would separate I would add some words in there, which is I care a lot about margins on investments we make today, but the margin profile of the future. And a lot of these businesses that are doing transformative things may have really crappy, certainly net margins, but more importantly, I think the question was going, gross margins, because of an investment profile that involves massive CapEx, etc. And so you can look at a business like a snowflake that had negative gross margins very late in their in their life cycle.

And that was a precursor to the LLM world, where these businesses have had very tough gross margins in the early days, and you’re now starting to see the leverage kick in. And so we are investing for the future. None of the investments we make are cash flow based in the short term, and in fact, very few of them are gross margin based in the short term. But a lot of them do require you to look over the horizon and see what can happen. And in a business like a Stripe or a Twilio or Shopify, they went through those journeys as well.

Many businesses have this really intense capital intensive investment horizon even outside of frontier tech. When we think about you said that the

We talk about this a lot, and I’d say, you know, we’re excited in many cases to be small investors in very large companies. You you look at an Anthropic or Perplexity or a Canva, you know, we have 9 figures into each of these companies. And yet we are still, you know, well below historical venture standards, where you aspire to own 20% or something in these businesses. It’s a long journey where billions more will probably be raised by at least a couple of those. And so it is a different venture.

If you believe that’s the end state, sure. But current reports suggest Anthropic may be raising at $1.70, and people are buyers at that number believing that they could be one of the next hyperscalers in a trillion dollar business. So 3.8x will keep you in business for a long time, but the reason to do it is because you believe it could be a 30x. That’s the basis of our Anthropic investment is we believe that it is a generational company. Now there aren’t going be many of those.

And so you have to be right. And that’s the scary thing right now is that the stakes are way higher than they’ve ever been. These businesses in some cases could still go to zero. And so you’ve got this these hyper power law outcomes that are scary. It is changing the nature of the game. I do think that scale matters for venture firms to be able to play over this arc of private life. On the flip side, the outcomes are going to be bigger than we’ve ever conceived of.

I sold my company years ago for hundreds of millions of dollars, and that felt like all the money in the world and was the top outcome for our software cohort in that vintage back in 2005. You know, now that’s a a seed round for some of these businesses.

Yeah. It’s a big landscape. And so there’s a skew with the dollars raised stats because of that concentration. I think the numbers right now, the top three LLM’s, if you include Anthropic, OpenAI and x in there, are going to raise $100,000,000,000 in this six month period, which is just an inconceivable number by any historical standards. Obviously, people are betting that there’ll be a venture return there. However, there are hundreds of other really compelling venture businesses that will be created in and around those ecosystems. And I do believe there will be great venture outcomes from a number of companies in and around those businesses.

And so the power law will play into the premium outcomes of those returns. But I also think there’s a lot of, you know, 10 Xs and many 100 Xs that are going to exist in and around those businesses. And so I do think that the economy is still vibrant and healthy. It’s just skewed much more than we’ve ever seen or or literally could have believed just a few years back.

I think it’s a it’s a legitimate question. Our view is no, but it’s a it’s an area of debate. And frankly, alpha comes from not only being right, but contrarian ideally, because you’re going to get some counter cycles in there. And our belief is that it’s going through another cycle. And AI is a foundational part of what vertical SaaS is ahead. Data models matter much more than they historically did. Connectivity and collaboration up and down the supply chain matter much more, marketplace capabilities are defensible moat.

There’s a lot of attributes that matter a lot. And as I think back on our vertical SaaS investments, like a Shopify or ServiceTitan or a Toast, when they added payments, it became that big next horizon unlock for them and really doubled the TAM and the market caps for these businesses. I think AI is going to do the same thing. That what it can do with ServiceTitan, as they talk about automating the technician experience and the ability to go out there and have a co pilot alongside of you.

I was at the MaintainX board meeting yesterday. They’re doing the same thing on the factory floor. When you look at luxury presence in real estate, what they can do for the real estate professional to interact with their clients. And so the competitive landscape is heating up in the sense that more entrants are able to come over from horizontals and come up from infrastructure layers to try to make a run at these spaces. But I do believe great vertical software will still win. These are big markets.

These practitioners deserve great tech, and they will get it. And so in many ways, I love that it’s not as sexy right now, and people are distracted because we’re going to stick you know, our core and work with great founders in great markets, and I think that those will be rewarded over time.

Absolutely.

Okay. So that’s a great question, I will confess. We’re in the challenger business and AI gives the incumbent some advantages that didn’t happen in Cloud one. So in Cloud one, you had a business model dislocation going from license to subscription, and you had a delivery model dislocation going from on prem to single instance, multi tenant, cloud delivered. In this AI wave, it’s really the next horizon of cloud. And so you’re layering intelligence on top of cloud delivery and business models. You’re moving maybe to a token model or some other monetization of value, but essentially it’s an extension of cloud.

And the incumbents have platform advantage, data advantage, massive distribution advantages. And so the fast moving incumbents are absolutely going to make a run at being the leaders in the next cycle, which hurts the challengers. And that is a reason to be scared. I still believe that the high execution challengers will beat them over time, and they also have some inherent advantages and innovators and elements. Some of these things still exist.

But when I look at our own portfolio, I look at a company like a Canva or I look at a company like Intercom that’s at a scale where in some ways they’re already becoming an incumbent in those markets, and yet they’re disrupting themselves at awesome rates and have AI products that are already deep into the hundreds of millions in revenue. And it’s just fantastic to see what they’re doing, which then I think will be an indication of what the public incumbents may be able to do if they’re nimble and act fast.

what what Owen and the team have done with this fin product, and it’s a great use case for AI to be clear, like, where you have the customer data, that interaction. But I was at one of our portfolio company board meetings the other day, they mentioned that they had switched from human based interventions to the fin product from Intercom. And they showed the stats of deflections went up, I forget, to 90% automated now, and their NPS went up. I said, Okay, I get that the deflections and costs are going down, but it makes no sense to me.

Why is a robot better than a human in interactions? And they said, Look, they’re giving faster and more comprehensive answers. And so the recipient, the customer who has the question, is getting, they’re getting links and references and more information back than our humans were providing. So it’s a better experience. And that was an unlock for me where I realized like, actually, this can be a win win win on so many levels, and it’s starting to happen. And so, customer support and service and messaging and and help desk and ticketing and these things, you know, is one of these killer use cases that’s just starting, but it’s gonna roll through so many other areas.

Oh, that question’s already being answered. It’s not even a debate anymore, Harry. It’s over. These tech solutions are absolutely addressing software, hardware, and services budgets comprehensively, and they’re doing it in a very successful way. And if you look at early adoption in categories that skew this way, and so you asked, are we still doing early stage vertical SaaS? We’re going down accounting and legal and medical. We’re going through these sectors where there’s a lot of frontage humans doing busy work and paperwork, and we’re supercharging them.

We’re taking away a lot of the manual transcription and summarization and error prone laborious processes, and we’re freeing them up. You know, a bridge is freeing doctors up and their patient interactions to actually interact and talk with a patient, than having to turn around and type things into the computer for most of the meeting.

The game is on. And I think that, you know, Epic has had this this wonderful state endorsed monopoly for a long time. I hope that they’re gonna continue to be forced to be open as a system. And I think you’re gonna to see a thousand flowers bloom in the medical ecosystem because that is one of the most important areas for AI to address. If you read Dario’s essay from Anthropic Machines of Love and Grace, I highly recommend everyone reads it, but it’s a tech optimistic outlook of what AI can do.

One of the great statements he has is that one hundred years of medical research is about to be pulled forward in the next decade. And so this certainly goes into diagnostics and treatments, but it also goes into patient care interactions. AI can be so damn powerful when you use it to help patients at the point of treatment, at the point of care, for follow ups, for preventative medicine, those sorts of things. And Epic holds the key in terms of patient data that we need unlocked.

We need that treasure trove to be accessible for these apps and for innovation to happen, and I think it’s going to happen, and I’m very bullish on the potential for AI in medical use cases and healthcare more broadly to be utterly transformative into quality of life and the treatment processes for patients.

We are. I just would push back on the on the ruthless leadership point in the sense of, I I love the statements these executives are making, which is we’re going to give you all the tools in the world to supercharge your daily job so that you’re doing the cool stuff again. You’re doing the strategic, the architecture, the direction. Tech is going to work for you. You’re not going to be a slave to tech. We’re going to grow the business, but we don’t need to grow the workforce to do it.

We’re going to supercharge what everyone’s doing. And so I think we’re going to see the era of the micro business. I think that we’re going to have 10 person companies that are crossing billion dollar valuations, and kids in schools are going to be able to launch businesses in real time in ways that haven’t been possible before. I think that is great for the economy. I’ve got three kids that are at various stages of entering the workforce, so there’s going be disruption, and that’s scary. And I absolutely admit that we’re going to go through this cycle that we all need to understand in terms of what entry level jobs mean, the training, the enablement, those sorts of things.

But society’s been through this many, many times before, and I believe that we will work through this cycle quickly and positively.

There’s this awesome history going back to Bessemer Venture Partners namesake, the Bessemer Steel process, which many people don’t know. You look at newspaper clippings from a hundred years ago, and there’s these great headlines and articles about the coming workforce dislocation and factory workers, you know, being displaced because the Bessemer steel process is so much more efficient and the struggles the economy is going to face and society is going to face. And yet literally fast forward, you know, a few years later, and buildings are built into the sky because skyscrapers are not possible with stronger steel, and railroads are built across The US, and transportation and connectivity and commerce unlocks.

You read the articles about, you know, the phone operators, and I think it was 4% of the female workforce was doing manual switchboards, and this idea of this huge dislocation of the workforce when that was automated. There’s hundreds of these microcycles that have gone through with different tech disruptions and things. It’s coming, definitely, And at the same time, opportunity is going to be created as a result. And the potential for these new workers to leverage technology to do amazing creative things. The micro film producer that can now, you know, release a movie that they can create on a laptop.

The ability to do apps, the ability to do fundamental research with agentic PhD level supporters in new areas of biology and physics and chemistry. We’re looking at fusion investments now that I think will be supported and accelerated by AI. There’s just this whole different wave of innovative unlock that will be possible, that will favor the the nimble and the reactive and the dynamic, but society at large will benefit.

Don’t get me wrong. It’s still a pretty damn good business if you ride that arc and scale it. But, yes. We just released a state of the AI report that broke this down and quantified it. We referred to these AI ecosystems as galaxies and talked about some of these supernovas and shooting stars that are emerging, where we’re seeing businesses go from zero to 100,000,000 in one point five years. That’s the supernova profile. Dario at Anthropic has now been open with it. They’re well past it, so I think he’s more comfortable sharing the numbers.

But zero, ten million, over 100,000,000, over 1,000,000,000 the next year. He’s openly said there’s a chance to cross 10,000,000,000 in the next year. It’s a curve that goes like this. You know, we used to have this chart in our state of the cloud report that had a seven year journey, and and those were centaurs to a 100,000,000. That has I mean, I

Yeah. That that’s cute now, but it’s off by an order of magnitude. This is a rare class of company and a rare breed. A small subset of even the companies we back, nonetheless all the companies that are started, meet that profile. But we thought it was important to document and share it, and say, this is actually possible now. Consumer like growth for enterprise businesses is happening, where adoption curves can pull through great products in inconceivable rates of adoption and speeds.

And the supernovas then yield to the shooting stars, which is kind of a four year profile, and I would say that’s the fatter part of the curve for businesses we’re fortunate to work with, where you see a pretty good number of companies going from zero to 100,000,000 in a four year arc. And again, that gets to this kind of quadruple, quadruple type cycle where the businesses are just scaling really steep curves. You had asked earlier about margin profiles. Some of these early on are needing to invest at heavy rates, but far from all of them.

You see a large number of these businesses that are doing it in pretty capital efficient ways, and that gets incredibly exciting. You know that you’re a data nerd like me, we talk about the rule of x and these trade offs between growth and efficiency and all that. And ultimately, we do still believe that all businesses should be valued as a sum of their future free cash flows, and that ultimately is the mark of a good business. And the incredible thing is that these businesses still have those fundamental economic profiles in most cases where they can throw off real cash flow at scale.

So actually, no. And that’s why I wanted to make that second point that efficiency still matters. We do think that trade off comes into play and we quantify it. The rule of x mathematically shows it’s about a two to two and a half x multiplier value of growth over efficiency at mid stage scale, call it 50,000,000 ARR or so. Early on, the math doesn’t matter much. It’s hopes and dreams, you’re just trying to get in market. So it’s an infinite multiple of anything. But when you actually get the engine going, there should be math underlying the fundamental assumptions.

And I assure you, one of the coolest things about the profile of Anthropic and Perplexity and Canva and these businesses is the math actually pencils out. You’ll hear these founders talk about like in the foundation model phase, it’s a bit misleading because the P and L doesn’t match beautifully, but you should think of these model releases as a product in and of themselves. And there’s a healthy life cycle to that product. And so, you’re monetizing last year’s training in this year’s revenue line while you’re investing in next year’s model, which is going to be monetized in next year’s release.

And so there may be these order of magnitude step functions where each atomic unit of product is highly successful and profitable, and yet the P and L looks upside down because of this hyperscale, no pun intended, growth rate that they’re enjoying in that forward investment cycle. And so even though it seems crazy, I believe it would be economically imprudent not to forward invest when you have that market demand there and when you can show economics working at each fundamental level.

be fits and spurts. And I think that was part of the people being underwhelmed with the five o release and discussions. But, you know, we’ve had these cycles before, and there will be, you know, breakthroughs. But fundamentally, I do believe Does

It does. And there have been some hard miles here. We’ve been through a lot, Harry. But the curve is still up into the right. Without a doubt. And I do believe that we will cross over this term. You know, people use various different terms about levels of reasoning and awareness and AGI and the like. I have no doubt if we’re not there, we’re going to blow past it very soon, and that we will get to this notion of higher level reasoning that does mirror the world’s smartest scientists.

And I think that’s coming in the next eighteen months, and that these curves are going to continue. How we harness that, how it instantiates itself will be the opportunity for us all to figure out and monetize, but I don’t think it’s slowing down. I do think that we’re also getting many more hardware approaches and solutions out there so that it’s not as wonderful as NVIDIA is, it’s not just an NVIDIA world anymore. The chipsets from Amazon, Google, and AMD, and others are becoming quite capable. You’re also going to see different approaches, different optimization paths, innovations in technology that unlock leap aheads in terms of training capabilities and cost to deliver inference.

I do think that we’re going to continue to see innovation there and the scaling laws continue to hold.

Yeah. I think that works in some spaces. I think that’s counter to my earlier point, though, which was the great products are being pulled through. People are finding them. You know, ChatGPT didn’t scream from mountaintops. They they delivered world changing user experience, and people showed their friends. Perplexity is doing that in search and answer engine world, and these models at the API layer are doing this for business users who are looking to connect them. And so I think the capital is important in terms of building and forward investing, as we talked about, for the business model.

But I actually think marketing and sales have less of a role in this new economy than they did before. And that these products in many ways are selling themselves, and product led growth and innovation is the unlock for this supernova and shooting star aspirational growth profile, because you can’t put human sales reps against these things. Just the sales learning curve that Mark Leslie talks about when we backed him decades ago in Veritas is no longer applicable because you just can’t possibly throw the bodies at a zero to 100 growth curve inside of two years.

Just the, you know, the sales models don’t support that.

So thank you for the I I think what it was meant as a compliment, but I’ll also say that we spend a lot of time I

No, it’s fair. And what I would say is, I mean, we’re certainly not value investors. We pay market clearing prices. We lean in where we believe it’s there and we’re buyers again at some of these top pegged rounds deep into the hundreds of millions. I think it’s very clear that we play to win. I think, though, the distinction there is that we do fundamentally want to understand how the businesses become self sustaining and scaled. We will walk from a lot of things that we don’t see the unit economics penciling.

One of the most famous and most painful for me was Tesla early on. It’s on my anti portfolio. If you go to the Bessemer website, we have a page dedicated to our screw ups, that’s one of mine. And it was because I couldn’t fundamentally see how the unit economics of the Roadster were ever going to work. And to be clear, they didn’t. And without the, you know, the DOE bailout loan and things, Tesla wouldn’t have existed. But what I missed was that Elon’s a force of nature, a generational entrepreneur, and he put that company on his back and powered through so many subsequent layers that the next arcs of the model worked and pulled everything else through it.

And that’s one of my big regrets, is that our job is to see that potential in entrepreneurs, to create those unlocks. And that’s the challenge I put back on myself, how to break this notion of short term discipline for the long term horizon of what’s possible. What prices do you do Perplexity at then? I mean, we’ve done the last several rounds. So I I forget the valuation of the first round. I give Pete Sensini, by the way, a lot of credit, who’s a good friend and a great investor.

He also did Databricks, but he was very early, with Arvind. As that business was being formed, we did that that, I believe, what was technically the second round. So I I think that was branded a b. It may have been technically an a and then and then subsequent rounds. But, again, I wish we were earlier and larger shareholders and still regret that we didn’t see what what Pete saw early in that business.

that question, and I hope you continue to ask that question to to LPs and later stage investors as well because it’s looked down upon right now. It’s sort of a dirty word. If we went and sold part of our position in some of these companies, people might think there’s signal risk there. There’s issues. And to be fair, Bessemer has this awesome history and we’ve generated billions for our LPs. So we don’t have DPI pressure that some emerging funds might or whatever. But I think that stigma is wrong.

If these businesses went public as they used to, mean, my very first IPO, Cornerstone On Demand, went public with $50,000,000 of ARR and I think with $700,000,000 market cap, and they traded up to billions over time. But that used to be a really successful IPO back many years ago. And now you look at Canva at $40,000,000,000 plus and Anthropic at $170,000,000,000 plus and Perplexity, you know, deep into DecoCorn status and the like, these businesses, you know, aren’t going public anytime soon. And yet, from an investor standpoint, there’s an argument that they should be handed off to later stage investors and hedge funds and things.

And so I do hope and believe for the industry that liquidity in the secondary markets is is viewed more favorably and more active. This is a change in my view. I’ll confess that I I was pretty hard lined against this, not for founders and teams. I always feel that taking some pressure off for them is good, but I didn’t love it when our co investors were looking for liquidity early in businesses. But I do think that in these mid stages when people are staying private so long that an outlet’s healthy.

And I do think that LPs deserve that. I think for emerging funds, it’s important that you have these options and that the world doesn’t judge you negatively for it, but actually understands that there’s some economic necessity in in a healthy ecosystem and capital flows both ways.

Oh, I hope so, Harry. God, I hope so. Definitely, the discussions are heating up again. I do think we’re gonna have a healthier IPO market at the end of this year and in particular going into next year, but in many ways, we have to. It has been record lows and record bad in the last several years, and that’s not sustainable for the capital flow reasons we just talked about. This liquidity discount hasn’t made sense. If you go back in the markets a decade plus, private markets traded at a discount because they were illiquid and there was uncertainty and less disclosures in those things, and logically they should.

Now it’s a growth adjusted discount. Of course, when private companies are growing faster, you have to normalize the multiples accordingly. But rationally, there should be a discount for the lockup characteristics and the information flows and those things, we haven’t seen that in many years. I hope and would love it if the public markets return to premium multiples and the private markets price off of those. I’m not convinced we will, but I am optimistic that we are finally going to see more IPO activity. When you look at our Cloud 100 list, which will be released here shortly, I’ll give you a little bit of a spoiler piece of news there.

We’re over a trillion dollars in private market cap now among just the top 100 cloud AI companies right now. Just a astronomical number to consider. And so there’s a trillion dollars How

much

of that is

I think this is entirely legitimate, Harry. And that’s the that’s the crazy thing. Now, of course, it’s skewed towards the top. I think OpenAI, x, Anthropic, Canva, Databricks, Stripe, go down the list. I think those are incredibly high quality companies that essentially are tradable public entities and a private wrapper today. So I think that those are entirely accurate marks and very real. The quality level of this list has never been higher. I think there’s buyers and sellers at the marks all the way down the top 100.

Now you can make a case that 101 to 300 may have some walking wounded, some last round prices that are artificial and the like, but I think we’ve cleared most of that out of the system. When you look at Mr. Irrelevant, if you use the NFL draft analogy, number 100, it’s an awesome company on the list. It’s a great business that certainly people would be buyers at or above the last round marks, I and think you’re gonna see that across the list. And so, again, there’s a trillion dollars of enterprise value sitting there that’s not yet in the public markets and and should be soon.

Yeah. They should, and I and I think they will, so I’m not not scooping any news. And I put them in a in a general basket of very short term, meaning, you know, in the in the q three window or the like. But clearly, they could have been public long ago, and they’re in no rush. And they, of all founders out there, are thinking incredibly long term. The ultimate giving pledge that they did, giving away 30% of their economics for public good, including, you know, a lot of initiatives in in Africa, I think shows

You you would have absorbed the, the 30,000,000,000 happily. See, Harry, the power line. I

deploy more than 29,000,000,000 Harry. Don’t get greedy, come on. Well, I could. Sam, do you want it? Yes, please. It’s amazing how many SPVs have popped up with that exact value proposition. Oh my God. It’s like there’s a

doubt. And

that is probably another sign that things may be a little heated. But I do think that great companies ultimately like Canva, deserve to be public, will be public. They’ve hired a great CFO in Kelly. They’re certainly giving indications that they’re headed that direction. But they continue to think long term. And they’ve made it clear to investors that they do not want a short term mindset, and we bought in. It’s one of our largest investments in our firm history. We’re hundreds of millions in, and we’re thrilled to be part of it.

And so they keep building value, and we believe that it it will be a great public company when they choose to go public. But, you know, the the urgency level there is moderate.

That’s awesome. So it used to be Twilio, then probably Stebbhub, Canva, Anthropic, probably that bucket. We’re very comfortably go deep into the hundreds of millions. We understand that, you know, these these rounds have gone grown to a point where scale does matter, and we need to be able to support our companies all the way up. And so we’ve we’ve added, you know, growth capabilities to be able to do that.

is a very different muscle, and I’ll tell you also, as you’re alluding to, it goes against your instincts when when you have a business that’s cranking and you own a lot and it’s marked up, you know, 10 x, and another round comes up that’s also at big forward multiple. This mental disconnect of, hey, let’s let someone else now come and mark it up and price it and let them run. I’m already sitting at a 10x, that’s great, Versus this, I want to be a buyer again and reset everything.

Put in 200,000,000 that now I need to go back to work and prove that I can dig out and get a return on again, etcetera. It’s intimidating. And we’ve actually added people and processes to make sure that we don’t get subject to this kind of mental inertia. We invite in another partner to look at it.

And our best deals, we’re constantly saying, okay, we have the Century team, believe which will be the iconic companies in next Century, and it’s also a bit of a riff on our Cloud one hundred, is a team that’ll come and help and basically partner with you on a deal and say, Okay, let’s take a fresh look at this, let’s re underwrite it, and let’s make sure that we believe there’s a 10x here ahead that can be there, in which case, you know, let’s double down. We’re trying to break that mental trap of being comfortable with success and being afraid to really back up the truck.

And so you’ll see us doing that more and more, I think.

It’s incredibly hard, especially when, you know, we’re we’re all investing at big numbers to begin with. And and we have a lot of times where literally the day the round is announced, someone will offer them a two x step up. You know, that could still be a great investment.

I think there’s a real risk of overfunding businesses. And so there’s a good in there, which is I I do think that there’s a positive when you get, you know, great firms and partners in there. There’s a signaling benefit, there’s an impact. I do think that we actually can help move the slope of the line a little bit for these companies and add value. And so one of the ways that we add value is that it makes it easier to raise downstream capital, and I think that’s totally true and a good thing.

But there is an excess, and the people showing up the next day with huge markups pushing more capital can be seductive to founders. And so part of the discussion is, look, if you have that additional capital, the critical thing is not to spend it in a disruptive way, and that’s the foie gras analogy, where you choke on the capital and you don’t want that. And if you execute on the plan every day, you’re going be adding value. So let’s make sure we’re fully capitalized to play out this next horizon of risks and goals and investment we want.

And if we could take a little extra capital to do that, maybe we do. But what’s more likely is let’s actually go out and execute, and let’s put this first wave capital to use, and let’s build more value, and those investors are going to be there at even higher prices downstream. And the important thing, and we try to get alignment with our team members and things, is we want to build a ton of value, and that if the round gets so frothy and runs away that it’s even too highly priced for us to double down, that’s okay.

That’s a good thing for the company. And if your cost of capital goes down a ton, then we’re your partner. We’ll go out and raise at a very high number and take very little dilution, and everyone wins because we’re shareholders and we’re aligned.

Oh, I mean, not only do I have the anti portfolio of the misses that we didn’t do, you know, the Tesla and Atlassian and companies like that front and center, But, yeah, doubling down on on every one of our winners. I mean, I’m fortunate. I think ServiceTitan was my thirteenth IPO, you know, have have a couple dozen unicorn investments. And so mathematically, every one of those, I should have done every subsequent round and wish I did. But I would also Is there one that comes to mind more?

I would say on the positive, Twilio, we did exactly this. I would say with a company like Procore or ServiceTitan, we still were very large shareholders, but we had a lot of people come in and follow. And I think this was one of the things, actually, you talked about vertical SaaS, so there is a good lesson in this. We underestimated TAM and weren’t sure these could be $50,000,000,000 businesses because we didn’t yet unlock the payments expander. And so we misassessed the total TAM and therefore got weak need investing into the billions.

And we should have. But we left a lot of money on the table. We own 28% of Shopify and Twilio at IPO, and we owned you know, well less than that of of Procore and ServiceTitan just because we we included a lot of other investors downstream.

Not only do we do it, we require it. Every one of our IRs, investment recommendations, the is our memo terminology, has a scenario analysis at the end. And we also have actually published many of these on our website where we’ll go back and publish the memos. And it’s kind of embarrassing when we do, because you look at the just goes nuts upside scenario, and they’re embarrassingly small. And it’s not because, as investors at the time, we don’t believe that they could be much more, but we’re trying to be rational, and we’re trying to bracket it in a medium term horizon that our partners will understand.

And yet, the tiebreaker of these deals is always the one that you and your gut believe can just go nuts, and where the 100x scenario is there. It’s always this amusing back and forth where the vast majority of our deals that we put forward, you know, solve to a three x. And you’re sitting there and like, you know, you look at these scenario analyses and it’s like, why is it that every memo I’m reading solves to a three x? And it’s because, you know, people are trying to balance and be rational and talk about capital loss and all these things.

But at the end of the day, the deals that get done, it’s the ones where the the the partner is sitting there saying, I’m pounding the table that the the high end and more is possible. That really is the the qualitative overlay that has massive quantitative implications.

And, you know, go back go back to the way that, you know, the Facebook, if you did a TAM analysis on Harvard, you know, Hot or Not,

Yeah. And so I I don’t fully concede that. Maybe the tiebreaker here in my mind is I look for at least exciting adjacencies. So you’ve got to have some killer unfair advantage to get started. You’ve got to have this mindset of, Okay, they know what they’re going to go attack first. They’re going to build a killer product. They can get into some vortex of growth and launch. I may have a lot of questions about the TAM, but there’s enough adjacencies, enough things that could go right that they could layer things on.

And so I want to see that the three-dimensional cube of segments and products and users that that can flex over time, and we don’t have to have it figured out. We don’t have to know exactly what it’s gonna be, but I have to believe that they’re playing in a big enough pond where good things can happen. I think that’s the difference, and I will totally concede that there are times where we’re not imaginative enough to go after it, and great founders will break through at times.

But I think that combination is still powerful. The investments that we’re making today, and certainly that I’m making personally, tend to still overweight massively those two things, team and TAM, and at least our vision of the TAM horizons. But I would say, like the analogy we were talking about of vertical SaaS before, where you go through workflow automation, then you go into payments, now you go into AI, and you go into services, Even what seem like small markets can unlock massive dollars when you’re creative about the horizons, and that’s what great entrepreneurs will do.

They’ll go attack those markets, suck up the value, and really deliver awesome product, and and can build great businesses. And and as long as you, you know, you price things rationally at each step and walk it up, it probably back solves into Jason’s math also, but from a top down rather than bottoms up point of view.

distributed early. We didn’t necessarily sell. So what we did is we gave people the choice. And a lot of our LPs, mind you, do sell pretty quickly after getting stopped by mandate. That left a lot of money on the table for a lot of folks. We absolutely wish that we had held onto Shopify and not distributed when we did. Hopefully, some of our LPs and certainly some of my partners have held and been able to benefit from the run up. But at the end of the day, you know, it’s a fantastic company.

I think there’s still a long journey ahead, and that’s why you see people still buying even at these valuations.

Crushing. Absolutely. Is a one of those force of nature generational entrepreneurs as well. Do you agree with Sequoia’s evergreen fund structure?

I agree that there’s a lot of positives to it. This idea that you have incentives to manage all the way through. Bessemer actually has a heritage where many decades ago, we had some evergreen components to it. But I do think it’s hard. I think that public management’s a different beast. I do think that the economics should be different. And the end of the day, especially in a DPI world, LPs get paid to manage capital and do the allocation, a lot of them want the capital back.

And so the merits of consistency and fund flows and those things have some trade offs with hold periods and public duration. And I love innovation in capital markets. We’re seeing other firms adding asset management businesses and debt products and roll ups and doing all sorts of things. You know, there’s there’s some kernels in there that we agree with. There’s a lot in there we we probably aren’t gonna pursue. But as a fan of capital markets and innovation, just I I applaud creativity and pushing bounds.

Do I agree with the Walmart analogy? No. But do I agree that scale is important? Yes.

Well, actually, the direct analogy of the low cost provider, lower cost of capital, you know, sort of push to the bottom, I think you’re seeing a maturation of the asset class. And I think the analogy might be the investment banking world. And that you have platforms like the Goldman’s and Morgan’s and JP Morgan’s that are providers of broad, global, multi asset, multi stage, multi sector that can be full service shops. That’s very much the path that Bessemer’s on. We have nine offices around the world. We manage tens of billions in assets.

We’re multi stage. We want to be able to support our companies all the way through. That’s very much the mindset we’re in. I do also believe that there are specialists, maybe, you know, the Tiffany’s analogy is maybe, I don’t in the banking world, maybe that’s the catalyst or what have you, where very good at very specific things. That’s very much the strategy you’re running. That’s the strategy that Benchmark continues to run, and I think that there’s a lot of opportunity there. And so there can be a bimodal curve in terms of approaches.

In particular, I think geographic firms or very specific sector firms. Healthcare has been an example where sector expertise has been really an advantage, but you need scale. And so I think healthcare specific firms have had success for that reason. This maturation is gonna make it tough in the middle.

Yeah. I actually agree with you for the most part. I would say that you will get alpha from some of those funds, but the important thing is not to get ossified in an approach. And this is very much why we don’t hire sector specific investors, and we don’t give you air cover if your sector goes out of favor. Meaning, our job is to make our LPs money. And if you’re in a sector that’s cooling off, you better get the hell out and go somewhere else that’s gonna make money or, like, you should stop investing.

The risk of having a semiconductor fund or a semiconductor team is that, you know, you carve out, pick a number, 500,000,000 to invest there. Like, you bet your ass they’re gonna invest 500,000,000 in semi, whether the right answer was 2,000,000,000 or 0. And that’s the risk. And so we have a very different approach, which is we are constantly optimizing the incremental dollar across sector, but also stage and geography. And we compete for dollars. And that is a mindset that we love, and it requires constant reinvention.

The term we use internally is roadmaps. At every one of our off sites, partners are presenting new roadmaps, and they’re talking about theme, sectors, sub sectors, investment hypotheses that they have that they’re going after. They’re getting feedback. They’re sharing it. They’re iterating. And if you don’t constantly reinvent yourself, you don’t have a future investment. Because the markets are moving fast, it’s our imperative to to get ahead of the next trend, not sit comfortably in a sector that has had a good run and believe that you’re entitled to another good ten years ahead.

that is why I’m laughing. It’s not a silly question. It’s a painful question.

Oh, Harry, I’ve had so many. I mean, in our industry, you just wake up and you read TechCrunch or or listen to some of your podcasts, you’re reminded how bad we are in the at this job because there’s so many cool things happening that we’ve missed. My first roadmap at Bessemer was RFID, Radio Frequency Identification, was and is a $0 market. It was a total dud. Thankfully, I only made one investment there and it also had a SaaS underpinning, we ended up pivoting and making a few bucks.

It was a stupid idea. What I credit my partners with is one, patience, but two, direct feedback. And so we iterated and pivoted. Because I wasn’t hired as a radio frequency identification investor, I was given the opportunity to pivot. My secondary roadmap was cloud, by the way, which ended up being a pretty nice second act, but it was a really bad idea. And so I had my first three investments were all very bad. My next two ended up being, you know, billion dollar

Three bad is Patients and support from the partnership to to make enough shots on goal to to get some statistical relevance out of the sample size. I remember one of my great senior partners, Hardiman, who was a professor at HBS for years, and I would go sit in one of his classes and have a long dinner or lunch with him. And he drew on the board my career, which is basically the straight you know, you know, with some bumps and then like a little bit of a tick up with Cornerstone On Demand and Eloquence of my early things.

But he’s like, he’s like, just give it time, dude. Like, you’re, you know, you’re wandering in the desert a bit. I know you’re anxious. You’re type A. You want success. But like, this business is all about building a portfolio and putting yourself in position to be successful. Don’t shoot out of the gun crazy big checks out of the gate so if you go zero for three, you’re Ease into it. And that was hard to take at the time because we’re all aggressive and enthusiastic in this industry, but it was the right feedback.

Resetting and learning and trying to get better allowed me to be in a good headspace for subsequent investments and to keep going and to have some confidence. And I do remember back when I was an entrepreneur, one of my board members was Robin Vasant. He was at Mayfield at the time. And I went to see him when I was going back into venture and joined Bessemer and getting some advice. He said, Your first investment’s going to suck. It always does. I was talking to about a deal specifically, he basically said, Well, so don’t do it.

And I’m like, But Robin, by definition then, I’ll never do an investment because you have to get over that first one. And he’s like, Yeah, but this one really Unfortunately, he was right on both. My first investment did suck, and I shouldn’t have done that one. But I think the point kind of applies, which is it’s a learning game, so just you need to be in it to have enough shots on goal to score some. And cycles matter a lot. I really feel bad for great potential investors who joined our industry with checkbooks in 2020, 2019.

And they did great deals, but at market prices, which were way too high. Many of those people got washed out of the industry and never got a chance to make their fourth and fifth investment like I did. And we may never know if they could have been great investors. And so cycles matter a lot, and having enough bullets in the gun to hit a target matters a lot.

Very much so. I mean, time diversification is one of the few things we can manage in our industry. You mentioned it in terms of exits with the LP comment where concentration matters. That tends to be true, but it’s not something we can control because there tend to be IPO windows, M and A windows, and our job is to try to take advantage of them and they’ll cluster, but we can’t usually create those when markets are closed. However, entry diversification we can create. Few of us are good enough to actually play the markets counter to trends and be most aggressive when things are cool and to really pull back when they’re hot, which is probably what you should do.

But at least some notion of smooth and consistent to weather those storms so that you can take advantage of the cluster of exits, it tends to be about the most successful model for time diversification and venture.

The latter. I do think that this was a point in time where people were looking for that reset, where for positive or negative reasons, they they wanted, you know, a fresh start. Because in venture, I do think partnerships and platform matter. Ultimately, a lot of those goals are to then go out and build up another firm. And so you’re either running from something or to something, but at the end of the day, I think the best firms are pretty flat at the top. And so you’re not seeking better economics, You’re really seeking a better environment or better structure.

And so I do think that a lot of great partners were able to launch out and get funded and kick off. And we’ll see platforms then built out of some of those new funds and probably more reinvention in the industry than we’d seen before. But it is one of the few asset cycles where past performance is an indicator of future success. Private markets are very much networking ecosystem based, and you see very analytically, I referenced my partner, De Hardiman, who taught at Harvard before. He would quantify this, and he wrote the private equity and venture capital textbook, and we do the data.

But I forget the exact numbers, but if something like eight out of the top 10 firms in one cycle would repeat in the next, because there was this virtuous cycle. And so the challenge is which of the new funds are gonna break in and be those those next two that disrupt, and there’s opportunity there, and maybe it goes up to three or four. But I also think that there’s a benefit to being a consistent, stable platform, and many of those will persist. A final one

I think the next wave of liquidity is gonna come from a combination of several buckets. I think PE will be one of them. It’s gonna be a pretty fun time for PE as these companies are long in the tooth, vast majority are not going to get public, consolidation will make sense. Because at the core, lot of these businesses are very high gross margin and are run pretty inefficiently because we’ve got a growth mindset, and we’re certainly not optimizing for cost in the early days. And so there will be a lot of opportunity to work through those portfolios, and I think private equity will have a run there.

I also think big M and A is coming back. I do think that the incumbents responding to the cloud AI imperatives are going to need to get back in the buying game. The FTC is finally taking a more rational posture on antitrust and and blocking deals and these things. I think they’re going to let market forces operate there again. And so there will be this buying imperative that rolls through the public markets. And companies like SAP and Oracle and IBM need to buy or they’re going to get crushed.

And then I do think that the IPO markets are going to open up again, and we’re going to see a pull through there. And the big wild card, this fourth bucket is what we talked about briefly, and I thought one of the great questions you asked was, will secondary liquidity start to flow through from different, you know, capital providers, and will crossover investors be comfortable doing more secondaries and things? And I and I think, yes.

I’m not totally sold on these exchanges for private markets or these sorts of things, but I do think incrementally we’re gonna see people get more comfortable with secondary transactions, stop looking for signal risk in that, and we’re gonna take pressure out from all these sources, which will then get capital flows coming back, which ironically may then just feed the engine at the front end even more. Did Tiger do more to help

Oh, boy. The foie gras analogy comes back to me there. Incredibly mixed. I think that there was a lot of happy disruption and big thinking and creativity that that I applaud, and there was a lot of reckless overfunding and deal work and nongovernance that we’re gonna have to clean up for quite a

I mean, that’s the amazing thing. Over these cycles, an index of the venture industry may be fine. Historically, it has underperformed the S and P and probably will again in this case, but you’ll make money. That very well could be the case there. It certainly wasn’t an optimal portfolio construction, but but at the end of the day, I think you’re right that they’re gonna have some home runs in there more than make up for a lot of the the zeros. And then in the middle of the pack, there’s gonna be a lot of one to three x’s that carry the load.

Oh, boy. Everyone on in our partnership is constantly outbound. I would probably put Jeremy Levine there in terms of just his ability to see unique deals that are contrarian. And I think, you know, part of sourcing, we think of it as just, hey, outbound aggressive, creative, etcetera. But I think the insights to see what others don’t is is the essential part of that that’s often overlooked. And and I give many of my partners credit for that, but but you asked for one, so that’d probably be my answer.

David Cowen. Steadily finds great people in weird places doing bizarre things. Mean, Rocket Lab is one, know, Peter Beck sitting in New Zealand, you know, recently or Auth0 or some of these things. He just has incredibly high conviction and will pound the table at times, you know, things that may seem non obvious and just has an incredible nose, has been doing this a long time.

boy. And to be clear, you gave me no prep, so I’m gonna give this on reaction. Seed Fund. I mentioned Pete Sassini before. What they’re doing at Laud Ventures, I’ve got immense respect for, where they’re going in at inception stage. They’re literally sitting in the classrooms and labs at professors. Pete’s done this many times already with with Perplexity and Databricks and many others where I mean, he literally was there at the inception stage of these Decacorn businesses. Any one of those deals would be a great firm run or a great career for an investor, and he’s done multiple.

So from a seed stage, I think they’re going to do some really special things. Early stage, oh, God, there’s so many. I think that the first round folks, I think Oh, this is horrible. I’m forgetting uncorked new name. Remind me. Oh, no, Uncorked is a new name. That’s right. The press says, I think uncorked does awesome things and actually is very aligned in kind of investment style and approach. You mentioned, you know, Lemkin and what he’s done in very selective ways. I just I have a lot of appreciation.

I mean, it’s

Harry, you suggest that that history is a negative. Quantitatively, history is a positive. I think the data shows that you actually benefit from some pattern of success in platform. And each of those firms has to go through some cycle. And both first round and uncork are certainly going through generational transitions, but I think the next generation are incredibly promising. I think the founders would even say they’re they’re proud that they, you know, they’ll carry on and and exceed them. And so I think those firms are getting stronger, not weaker.

And and that is that takes nothing away from a number of the upstarts that you highlighted. I think they’re less proven. And so we’re also trying to work with them, but we’re figuring that out.

I mean, Maritech’s fantastic. I love their culture. I love their style. I mean, quantitatively, think that, you know, Sequoia continues to to mint platinum records. And so, you know, probably some combination of the two.

Oh, boy. My partners would probably say I should be more detail oriented. I am much more team and gut, and so I’ll confess I don’t intend to change it. But but I do miss some of those small things at times. Like, I’m not a guy that’s going deep in documents and details and things. Like, I’m not I’m not gonna be hung up on some usually deal term or financial thing or whatever. Like, I I’m all about working with great people this phase in life. It’s too short.

I love this job, but I just I wanna have fun. I wanna do big things with great people, and and that’ll get me in sometimes I’ll do crimes of of commission where I’ll just fall in love with some things and miss some details, but I don’t think I’m gonna change that.

One, I I wish you’d name him because that’s incredibly flattering. I I I appreciate that, and I love hearing that. I do think I try to live by that. I mean, I’ll confess, when I sold my business long ago, my wife and I sat back and said, okay, like, hey, it’s not life changing money, but actually, we could retire on that money. Do we wanna move somewhere, raise a family, unplug? And I couldn’t conceive of doing that. We are all so damn fortunate. Like, there’s a lot of great economics in this industry, and we get paid way more than we deserve.

It’s just so damn fun to be in the middle of what we’re doing. I can’t imagine not doing this. It’s an incredible privilege to sit down with awesome people who wanna change the world and help them do it. Like, it is it is the coolest job on the planet. And so, like, I have no intention of stopping anytime soon. And and the money is,

you know, is a nice byproduct of it, but, it’s a cool gig.

The what can be from the scale of the opportunities. Like, I I thought we understood this next phase we are going into, how big this was gonna be. And very sincerely, we’ve probably added a zero to everything. I think there’s going to be a lot of trillion dollar businesses that are created from this. And I said it was kind of embarrassing when you look at our scenario analyses and our memos that we published on our website before, because we talk about a billion dollar outcome as a big deal and the great success case in these businesses went on to become 10 or $100,000,000,000 companies.

That exact thing’s gonna happen where we talk about some of these investments and we hesitate and you know, we get really close and we invest, but we’re anxious and nervous and whatever, and at the end of the day, things just blow through everything that it can be. And we’re seeing it real. Like, when you actually see the Anthropic numbers that he’s now sharing, this is real. You know, their path to billions and billions in revenue and really compelling unit economics, we have never seen this in the history of our industry, and it’s playing out

where our mouth is. We’ve been a buyer so over. And that’s, again, like, a year ago, when we were buyers, we still wouldn’t have didn’t conceive of that. We thought that where they’re at now would be the exit, not another entry point. And that is incredibly awesome.

One of the beauties of this business is

can you know, fortune favors the bold, and there will be those outcomes ahead. And obviously, we placed our bets where we think they’ll happen, and there’ll be many more.

Great to see you again. Great to be back. I look forward to seeing you in person. But until then, thank you.