Ep7. The Great IPO Debate, Tesla Robotaxi v Uber, Tech Check

00:00:00.000 | If you're public with $100 million in revenue and a 10% growth rate,

00:00:04.960 | your valuation is not going to be all that great. But guess what? If you're private

00:00:10.400 | at $100 million in revenue with a 10% growth rate, it's not like you're better off.

00:00:14.640 | You're just fooling yourself.

00:00:18.320 | Hey man, good to see you.

00:00:32.000 | Good to see you, Brad.

00:00:34.240 | I've been well. How was the eclipse? Didn't you have an eclipse party down there?

00:00:38.080 | We did. We had a lot of people out to a location that had a total eclipse and we got super lucky

00:00:45.440 | about 30 minutes prior to the total. It had been cloudy and clouds parted, blue sky,

00:00:52.000 | bright sun on everyone, and then a miraculous event. It's kind of strange to see wrong humans

00:00:58.480 | applauding at the sky. Was it religious for you? It really was for a lot of people.

00:01:04.080 | It felt that, you know, that's a loaded term, but yes, let's use the word spiritual. It felt

00:01:11.840 | spiritual. And certainly our forefathers, you know, equated it with something religious.

00:01:20.560 | Well, I was sorry to miss it. My mother told me, she's 88, and she said she was in tears.

00:01:27.920 | And, you know, I honestly, I was skiing with my 13-year-old son, and so we missed it. But,

00:01:37.280 | you know, the rest of my family saw it and they said it was incredible.

00:01:40.560 | Well, lots of tensions in the world, Bill. You know, it was interesting.

00:01:45.360 | Oh, no.

00:01:45.840 | I tweeted last week that on the one side, you know, I really see a lot of great things

00:01:52.880 | happening in the markets. You know, M&A, I think, is going pretty bonkers right now and

00:01:58.960 | liquidity is definitely heating up. You saw the rumors about the Salesforce Informatica deal,

00:02:04.000 | the Google HubSpot deal. I think I know at least six other deals over a billion dollars

00:02:09.360 | that, you know, are actively being courted in the M&A pipeline. So you got that on the one hand.

00:02:17.440 | On the other hand, you know, we had inflation come in hotter. We have the 10-year kind of

00:02:24.000 | back at 4.7%, so up a lot for the year. And we have these geopolitical concerns that are not

00:02:31.120 | only tragic, you know, human events, but the backdrop is getting, I think, more challenging

00:02:37.280 | for the markets at the same time that the markets have been done pretty well for the year. So that

00:02:42.640 | always gets me concerned. You know, I was with a legend investor over the weekend, and I said,

00:02:49.440 | "What was your net exposure at the beginning of '23?" And he said, "80%." I said, "What about

00:02:54.960 | the beginning of '24?" He said, "40%." I said, "What about now?" He said, "Zero." Right? And

00:03:00.080 | so whenever you have that sort of, I think, reaction, it always, you know, you have to slow

00:03:07.120 | down and think about it. So I definitely think there's a lot of increasing volatility. We're

00:03:11.760 | heading into this election here in November that most people aren't even yet thinking about, but

00:03:16.320 | we know that's going to lead to a lot more volatility. So I'm feeling, you know, looking

00:03:21.360 | at our own portfolios, I'm feeling that tension on the one hand really excited, on the other hand,

00:03:26.240 | increasingly nervous. Well, as you know, I've often sworn off the notion of macro analysis,

00:03:34.240 | primarily because I think the best investors of all time have sworn it off. And through immense

00:03:41.280 | pressure from you, I've started paying attention to these things, which I don't like because I

00:03:49.280 | think it's quite clear that right now risk seems to be on the rise, like just the term risk, you

00:03:56.000 | know, across a bunch of different vectors, which is unfortunate. It'd be nice to see it start

00:04:01.600 | moving in the other way, but with the election coming and these different conflicts around the

00:04:06.480 | world, it's hard to have confidence that something like that could happen. Yeah, no doubt about it.

00:04:12.000 | Well, speaking of the markets, maybe we just jump right into our first topic here, which is,

00:04:17.040 | you know, there's been a lot of debate over the course of the past few weeks on the IPO markets,

00:04:22.560 | the size you need to go public, why we have so few public companies in the U.S. So maybe just

00:04:29.120 | kick off with looking at a little FRED data to normalize where we are, right? The number of

00:04:34.080 | public companies has gone in the U.S. has gone from 6,500 or so down to about 4,000 over the

00:04:40.560 | last 20 years. This is, you know, a cause and concern for many because on the one hand, we have

00:04:46.960 | more innovation, we have more startups, and so you would think you would have more public companies,

00:04:52.160 | not less. And if you look around the world, that's in fact what you see. And what you saw for a long

00:04:58.160 | time in the history of public markets in the U.S. up until this recent period. But this is a big

00:05:03.680 | number. It's like over 40%. Yes. Like significant. Over probably the most prolific period of

00:05:11.920 | innovation in the history of the United States. So now, of course, Jamie Dimon in his annual letter

00:05:18.160 | just last week said, yeah, well, that is true. And perhaps concerning the number of private equity

00:05:25.600 | backed companies had skyrocketed from 1,900 to over 11,000. So, you know, I started, you and I

00:05:34.080 | started kicking around like what are some of the, you know, first, you know, are these, is this bad?

00:05:39.120 | What are some of the causes of this? And of course, there was a tweet last week about Philippe

00:05:44.400 | Lafont, who I think was speaking at the information event in New York, where Philippe said, listen,

00:05:50.160 | the markets have structurally changed. It used to be you could go public with 50 or 100 million

00:05:55.760 | dollars in revenue. And he said, today you have to have a billion dollars in revenue in order to go

00:06:01.360 | public. Gokul tweeted something similar at first, looking at some Meritech data. He said, it looks

00:06:09.920 | like unless you have 700 million in revenue, then you'll probably underperform in the public markets.

00:06:16.960 | So here's his tweet there. And then Jam and Ball on our team, you know, responded to that and showed

00:06:23.520 | some data that we'll show here that says, no, in fact, if you look at the, you know, the last 10

00:06:30.160 | years, the average software IPO has been about 185 in median revenue over the last 12 months.

00:06:41.600 | And it has over 50 percent growth. And so there's this real debate, right? Like how big do you have

00:06:48.960 | to be to go public? What is that profile look like? So maybe you could just weigh in a little

00:06:54.480 | bit with your thoughts on, you know, that debate online between Philippe and Gokul and Jam and about

00:07:00.880 | kind of what's the revenue profile that one needs in order to get out the door today?

00:07:06.160 | I'll make some very quick comments about what I kind of fundamentally believe in my heart of

00:07:12.240 | hearts. But then I think we do need to address some of the realities that are out there. So

00:07:17.760 | personally, you know, I think I was on the board of OpenTable when we went public and we did on a

00:07:24.400 | 10 million dollar quarter. So that's a 40 million run rate, far shy of where we are. And if I bring

00:07:31.040 | that up, some will say, well, that was 20 years ago. And it was. I have always believed and I'm

00:07:37.520 | not the only one there. There's there's an assortment of other people in our industry

00:07:42.240 | that being public is great for companies. It raises the bar in terms of their performance.

00:07:47.920 | It causes them to to to, I think, achieve more than they would otherwise. So I think it's very

00:07:54.800 | positive for our ecosystem. I also think it's positive for the U.S. financial markets, for

00:08:01.840 | companies to feel comfortable coming public sooner. You will hear the SEC and others worried

00:08:08.960 | that mom and pop investors don't have exposure to these names or don't have exposure to these

00:08:14.480 | trends or to these companies. And the more that that, you know, they don't go public into their

00:08:21.200 | billion dollars, a lot of those gains will be gone. But the people will miss out on this. So

00:08:27.680 | I don't think it's healthy. Now that we started with a data point that the number of public

00:08:33.520 | companies has shrunk. And so, like, there is something going on and we could talk about,

00:08:40.400 | is there something that could happen that would change that or something that that has caused

00:08:44.640 | this to happen? It's interesting to me because I think there is in the boardroom in Silicon

00:08:50.560 | Valley today, among founders, there is a lot of question like, what do I, what's the profile of

00:08:57.120 | the company need to be in order to get public? And, you know, Gokul came back and responded

00:09:03.040 | to Jammin in a way that I thought was interesting. He said, well, OK, you don't have to be at 700

00:09:07.200 | million in revenue in order to get public, but you have to get to a billion within five years

00:09:12.800 | of being public. You know, I called the heads of all the major banks who run the capital market

00:09:18.480 | groups because I wanted to get their opinion. Like, you know, they're the ones who actually

00:09:22.720 | advise. So what do you all think it needs to look like? And the heads, one of the head of capital

00:09:28.560 | market said to me, and I'll just read it here. He said, I feel the bar for size is somewhere between

00:09:35.280 | 200 million and 300 million of revenues, premium growth rates to peers. So the median growth rate

00:09:42.080 | has been around 50 percent and attractive unit economics. I think you need to net out at greater

00:09:48.880 | than two billion. So two to two and a half billion in terms of market cap on average that generates

00:09:54.320 | an IPO of at least 200 million to 250 million dollars, which has enough float to make it

00:10:00.480 | reasonable position for investors. He said to me, the IPO market is slow in volume terms for lack of

00:10:06.880 | supply, not lack of demand. And I've been making this case that the IPO window is wide open. It's

00:10:14.160 | just a matter of price. Certainly Altimeter and Cotu and others would like to buy companies at

00:10:19.360 | this size that we think can compound at 50 percent or higher for the next five years. Right. Those

00:10:26.640 | are terrific companies to back so long as the point of entry is at a price that reflects what

00:10:31.600 | current market multiples is. So, you know, he's kind of taking the other side of the argument

00:10:36.400 | saying this is not a demand problem. This isn't that you can't get it into the market. It's a

00:10:41.040 | supply problem that's for some reason companies just don't want to come into the market at that

00:10:46.640 | size. That becomes a circular argument, right? Because if people start saying that, you know,

00:10:53.120 | externally, oh, you've got to be at one hundred or you've got to be at two hundred, you've got to be

00:10:56.800 | at seven hundred. Then people hear that and then they think they can't go. And so, you know, but

00:11:03.040 | let's talk about some things that have led to this. You know, on one hand, there's immense

00:11:08.400 | capital availability that I had suspected might go away with the market reset, but it clearly didn't

00:11:15.440 | or with interest rates going up. And so companies that don't want to be public don't have to be

00:11:21.840 | because they can get access to private capital. And at least today, that private capital will

00:11:28.560 | let them do secondaries, which solves one of the problems that that historically has brought them

00:11:35.120 | to the public market. Second, you know, I think structurally those people you call have built

00:11:42.640 | their Wall Street businesses to cater to these larger companies. And so they prefer to work on

00:11:49.600 | a bigger IPO. Way back when there was a group of bankers known as the Four Horsemen that took a

00:11:55.120 | ton of companies in Silicon Valley public. And I found some data which we can put up, but like

00:12:03.280 | the vast majority of IPOs are being underwritten by like four or five firms. And one thing that I

00:12:10.000 | think would be helpful is if some of the other firms were kind of more dedicated to a smaller IPO

00:12:16.480 | where everyone knew they were the go to for that. But that doesn't I don't think anyone's

00:12:22.640 | filling that role today. And so those are two big things. And then the third one that could

00:12:27.360 | contribute to it is just regulation. And I throw the cost of being public in with regulation. So

00:12:34.320 | that could include litigation costs, you know, the insurance you have to buy for your board,

00:12:41.120 | all those things like is the cost of being public. It's probably two to five million a year

00:12:46.480 | just to be public. So let's break those down a little bit because I think, you know, it's a

00:12:52.000 | combination of factors, right? So let's just normalize around one. So if you pull up this chart

00:12:59.920 | on the Instacart valuation, right, over the course of the last several years. So Instacart was one of

00:13:06.320 | these companies that got, you know, really high valuations during the pull forward that we saw

00:13:12.880 | in the Zerp COVID period. You know, it hit a valuation of almost $40 billion.

00:13:19.440 | They went on to raise subsequent rounds of private financing at a lot lower.

00:13:24.240 | They then went public at just around, I think, $6 billion. So here's the IPO performance chart.

00:13:33.360 | And here's the valuation trajectory of that business over the course of the last several

00:13:38.960 | years. So I think one of the obstacles, right? So it's performed incredibly well.

00:13:43.680 | Off of the bottom, it's performed really well since the IPO. But it took a board that had

00:13:50.160 | the courage to say, you know, we're going to ignore what the prior valuations were.

00:13:55.360 | And we're going to focus on getting the company public. Now, in the case of Instacart, I think

00:14:00.320 | you had, you know, the branding benefits of when they went public. I think they, you know, had

00:14:05.520 | access to, you know, to the cash that they needed to invest and grow in the business.

00:14:11.120 | But, you know, this seems to undermine this argument that, you know, that companies can't

00:14:16.400 | do well if they're on the smaller side, you know, post going public. So this is an

00:14:22.560 | early indication. But I think that there are a lot of companies in the venture capital pipeline

00:14:27.760 | bill that look like this, that had these really high marks in '20 and '21. And they need boards

00:14:35.280 | and founders and CEOs who have the courage to enter the public market and just accept the new

00:14:41.040 | set of marks. And I think this is one of the big psychic or behavioral hurdles to these companies

00:14:46.960 | getting out and getting into the public markets. Yeah. Yeah. And look, I mean, there are smaller

00:14:52.960 | companies, you know, like even in the SPAC space, look at him and hers, or look at SoFi,

00:14:59.440 | Antony Noto. Like they've actually done well over the past six months, like numbers up and to the

00:15:06.400 | right, stock up and to the right, and much smaller than Instacart. So it is doable. Yeah, I think

00:15:14.880 | you're absolutely right that the structural blocker that comes from having a previous private round

00:15:22.160 | is always a problem in these situations. And there are ways around that that can be negotiated

00:15:27.840 | around. People just need to kind of bite the bullet and take care of that. There's another

00:15:32.880 | thing that happens that relates to that, which is there's this presumption that, oh, well, you're

00:15:39.360 | not, you know, you're not in a good place to be public. Or the other thing I hear is, oh, imagine

00:15:45.840 | if you get out and you're too small in your public, how horrible that is. And to me, that's just this

00:15:53.600 | indication that there's this ostrich mentality. And what I mean by ostrich mentality is someone

00:15:58.800 | willing to stick their hand, head in the dirt, and not see anything and therefore feel better

00:16:04.080 | about themselves. So where I'm going with that is, if you're public, with 100 million in revenue and

00:16:11.920 | a 10% growth rate, your valuation is not going to be all that great. But guess what? If you're

00:16:17.680 | private at 100 million in revenue with a 10% growth rate, it's not like you're better off.

00:16:22.800 | You're just fooling yourself, right? In fact, because of a lack of liquidity premium,

00:16:30.720 | you're probably worth less than that public company. And structurally, your cap chart's

00:16:36.720 | more rigid. And so you have less flexibility. You can't do any acquisition. I mean, there's

00:16:41.840 | all kinds of reasons why that's a worse place to be. But people have this belief that if I

00:16:48.720 | can see that price and it says $2.33, oh my God, I'm in this horrible place. People have recovered

00:16:57.040 | from that. So I think we've talked about this before. Here's a chart that we put together

00:17:05.280 | that looks at the valuations, the multiples that companies were coming out in different cohorts

00:17:14.560 | by year. So you can see that the multiples of the companies that came out, not surprisingly,

00:17:21.120 | in '20 and '21 were really high multiples. And then you can see that the companies that have

00:17:26.320 | come out in more recent cohorts are at these much lower multiples. And so I really do think

00:17:32.320 | it comes down like this is one of the big hurdles that we have. And we saw this after 2006, 2007.

00:17:40.320 | We had companies funded at really high multiples. I'm thinking even about Zillow or Kayak that were

00:17:47.920 | funded then. And they waited a much longer period of time coming out of 2008, 2009 in order to go

00:17:54.640 | public because it took them a while to grow above those multiples. And I think that in that case,

00:18:00.880 | there were actually ratchets that were in those prior preferred rounds that prevented those

00:18:06.880 | companies from coming public. That's not the case in most of these deals that were done in '20 and

00:18:11.600 | '21. And so it seems to me when you look at this chart, the multiples that persist in the public

00:18:19.360 | market today, they kind of are the multiples. And so a lot of these companies will be forced

00:18:24.320 | into the-- they're either going to have to take a down round in the private markets to have access

00:18:28.800 | to cash, or they're going to have to access it by way of the public market. It gets back to your

00:18:34.640 | point. There's no hiding from whatever fair value is for these businesses. And now we're talking

00:18:40.560 | about businesses here that are doing a couple hundred million in revenue that probably still

00:18:45.840 | haven't hit profitability, where growth maybe has started to slow below that 50% annual growth rate.

00:18:53.680 | And I think we have over 1,000 of these companies that are marked over a billion dollars

00:18:59.200 | that still have to get work through the system. Now, I'm seeing some more M&A that's happening,

00:19:03.920 | so larger strategics coming in and buying these companies. I think that will be the answer for

00:19:08.080 | some. But to me, we're starting to see the IPO pipeline fill with these companies that I think

00:19:15.520 | will be on the smaller side, that will have growth rates that are still 30%, 40%, 50%

00:19:20.400 | superior to public market growth rates. But the valuations will be below, in many cases,

00:19:28.240 | their last big round of private financing. And by the way, I hope you're right. I mean,

00:19:33.680 | I hope we see that. We do have a situation where at least the big guys have been pushed away from

00:19:41.200 | M&A by regulatory, which is another reason why you need to be more serious about the public option,

00:19:48.000 | because there are fewer, relative to history, I guess, fewer options on the M&A side.

00:19:53.680 | And this gets back to the point I was thinking about. There's no hiding place in being private.

00:20:02.160 | And I've often said, the minute you took stock options and started handing them to your employees,

00:20:09.360 | you're in the game, you're on the field. And it's a tough game. And the number of people

00:20:14.480 | that make it to the next tier always drops exponentially. So we often talk in Silicon

00:20:21.680 | Valley about the Googles and the Facebooks of the world, but most people don't make it there.

00:20:27.440 | And so it's a hard, steep climb. And you need to have realism about what's available to you.

00:20:35.200 | One thing quickly that I wanted to highlight, Gokul referred to this Meritech report. I don't

00:20:41.120 | know if that's public. You were able to get a copy that I was able to read. You mentioned

00:20:46.000 | unit economics. I think being honest with yourself about your own unit economics is super important.

00:20:52.960 | And at least for SaaS companies, they had a ton of great data in that report that could help one

00:20:59.520 | figure out where you are and what your real multiple is very, very likely to be. I don't

00:21:04.960 | think it has to be a guessing game. What often happens in Silicon Valley is everyone

00:21:10.400 | looks at the top one or two data points and tries to equate themselves with that.

00:21:16.080 | If you look deep in this research report, you shouldn't be doing that. But that does happen.

00:21:22.320 | And I think the number one thing, if you're a smaller company coming public, is it's about

00:21:26.160 | growth. Your unit economics have to be working. So your net dollar retention, if you're a software

00:21:32.080 | company, needs to be in a range that shows that your customers love the product, that they're

00:21:37.200 | sticky to the product, et cetera. But these companies that have waited a little bit too

00:21:43.200 | long, Bill, now their growth rate is 20%. They're marginally profitable. They have 200 million in

00:21:50.720 | revenue. Unfortunately, that is not a profile that can come public. So this question, this debate,

00:21:56.960 | like what's the magical number? Philippe, is it a billion or is it 700 million? The answer is,

00:22:02.080 | for those of us who are underwriting these businesses, we're saying, is this a fantastic

00:22:06.720 | business with great unit economics that can grow at above market rates for a long period of time

00:22:11.680 | and expand margins? If the answer to that is yes, then the public markets are wide open. But guess

00:22:16.640 | what? The private markets are wide open for those companies as well. And that brings me to this

00:22:21.280 | point. I think the biggest thing that has changed over the last 20 years, back to if we think about

00:22:29.040 | the companies that went public in those early vintages, the Microsofts, the Apples, the Googles,

00:22:36.720 | the Salesforce, et cetera, what was one commonality, Bill? They needed access to capital

00:22:44.240 | to grow, right? And if you look at the private market alternatives for them for capital,

00:22:50.720 | we didn't have sovereign wealth funds that were writing multi-billion dollar checks at that point

00:22:54.880 | in time. We didn't have deep pools of liquidity and growth equity and late stage venture that

00:23:00.240 | were writing multi-billion dollar checks at that time. And I think that is a huge difference today.

00:23:05.120 | So I talked to the CEO of a very, very large company that many are speculating as to when

00:23:11.200 | they will go public. And they said, we just don't need to go public. Marks are not a consideration

00:23:17.520 | at all. Employees get liquidity. We have access to capital. There's just no need. And so then

00:23:24.800 | I read the Jamie Dimon letter. And I'm going to read a part of his letter because he talks about

00:23:31.680 | us going from 1900 companies to 11,000 companies. And he talks about the liquidity in the private

00:23:39.600 | markets, how deep it is for these companies to be able to grow. And he said there are good reasons

00:23:44.880 | for private markets and some good outcomes result from them. For example, companies can stay private

00:23:50.160 | longer if they wish and raise more and different types of capital without going to the public

00:23:54.560 | markets. However, taking a wider view, I fear we may be driving companies from the public market.

00:24:00.960 | The reasons are complex and may include factors such as intensified reporting requirements,

00:24:06.480 | including investors growing needs for environmental, social, and governance information,

00:24:11.680 | higher litigation expenses, costly regulations, cookie cutter board governance,

00:24:17.120 | shareholder activism, less compensation flexibility, less capital flexibility,

00:24:22.880 | heightened public scrutiny, and relentless pressure on quarterly earnings. So going back

00:24:28.960 | to this company, right, if you're one of these companies, you're cash flow positive like this

00:24:34.000 | company is, you're growing fast, you have access to all the capital you need. You know, think of,

00:24:39.920 | you know, the SpaceXs, the ByteDance, the Databricks, the Stripes of the world,

00:24:44.800 | all the companies people are waiting to see when they're going to come public,

00:24:48.800 | they all fit that profile. They don't need the public markets to provide the liquidity and given

00:24:55.840 | all of the headwinds that Jamie Dimon talks about, all the reasons not to come public, you know,

00:25:01.760 | it may very well be that the very best companies choose to stay private, right, which we may even

00:25:09.520 | see some adverse selection into the public markets. And it's only the companies that can't

00:25:14.560 | rave private financing that try to get into the public markets. What do you think about

00:25:18.720 | both this idea that we have deeper and more liquid capital markets, private capital markets,

00:25:24.400 | and we've created all of these government burdens, litigation burdens for the public company?

00:25:30.240 | Yeah. Obviously, my first reaction is if he's right, if those are all the reasons,

00:25:36.320 | the vast majority of them relate to this kind of cost of being public. And if you separate his list

00:25:43.200 | between what might have changed recently from the past, to say what might have caused this,

00:25:49.760 | these changes recently, you know, it's mostly the litigation, the ESG stuff, the compensation

00:25:58.480 | flexibility, it's the activist stuff. And, you know, you and I already had a long discussion

00:26:05.200 | about SBC, but I find that it would take, if the SEC were generally worried about why more

00:26:13.680 | companies aren't going public, I think, and I don't even know if they have the wherewithal to

00:26:17.920 | do this, because some of them may live at the judge level, but it would take a reevaluation of

00:26:23.120 | why the cost of being public or the regulatory burden of being public has grown. And is there

00:26:29.040 | anything we can do to take it in the other direction? I'm particularly worried about

00:26:36.880 | things like derivative lawsuits, where you have a very ambulance chaser mindset, going back to the

00:26:43.280 | Elon compensation thing, someone with nine shares, who made money can bring that kind of legal case

00:26:51.200 | against the company where the stock went up. That's really bad. Like, that's not what I would

00:26:57.120 | call a efficient capital market. I would call that an inefficient, broken capital market. And so I do

00:27:05.520 | hope we can look at some of those things. The second thing I would say is, if you're right,

00:27:09.680 | and if this becomes some kind of permanent state, I think it's going to be really tough on limited

00:27:15.840 | partners, because you're going to have a bigger and bigger part of your assets in things that

00:27:25.760 | where we really don't know what the price is. And if you have large positions, despite the fact that

00:27:32.880 | there may be traits for employees, you probably don't have a trade at the size that would allow

00:27:38.480 | you to get. Let's hit on that. That's a super important point, Bill, because we've been talking

00:27:44.240 | about why do we have fewer public companies than we had 20 years ago? We've been talking about this

00:27:48.240 | for a long time. And people have thought, well, maybe it's just part of the market cycle.

00:27:52.960 | I'm evolving my own thinking to say, this is permanent and this is structural. Because the

00:28:00.320 | government regulatory burdens are at least semi-permanent. And the markets have responded

00:28:06.640 | by the private capital markets stepping in and providing the liquidity where companies can get

00:28:13.360 | the cash they need, can get the secondaries for employees they need, without having to go public,

00:28:17.520 | at least the best companies. So if you think about that, you made a good point. So if you're

00:28:24.480 | an employee of SpaceX, or you're an employee of Stripe, or you're an employee of ByteDance or

00:28:30.080 | these other companies, you probably have quarterly or biannual liquidity events.

00:28:35.920 | They raise a secondary tender, and those employees can sell up to a certain amount within those

00:28:41.360 | tenders, right? But in those same companies, you have massive LP gains that are locked up,

00:28:49.040 | right? And it's very difficult. You can't sell easily a billion or a $2 billion position in

00:28:56.240 | ByteDance or sell a multibillion dollar position. So all of the endowments, foundations, pension

00:29:03.040 | funds, they're in a different position than they've been in the past. They perhaps have--

00:29:07.840 | even if you did, it would be a trade by appointment, pink sheet kind of,

00:29:11.760 | likely with a banker grabbing 5% or 10% along the way. It wouldn't be an efficient transfer.

00:29:20.080 | Right. So that to me is if you said, what's the ultimate cause of these companies to go public?

00:29:28.720 | The ultimate cause would seem to me that at some point, the LPs, the people who put up the first

00:29:34.640 | money into the business need to get liquidity, right? But two structural changes that we've seen

00:29:40.880 | there, right? One is it just seems the best companies will stay private longer, right? And

00:29:46.720 | we've seen this trend over the course of the last 15 years. Although I will remind you that Zuckerberg

00:29:54.080 | said he should have gone public two years sooner than he did. Right. And think about how early he

00:29:59.600 | went public relative to the late stage private companies that exist today. So I would say I do

00:30:07.600 | see that these companies will eventually come public because they need to get liquidity back

00:30:15.120 | to those LPs who provided that first capital, unless the private market again responds and

00:30:20.800 | provides some vehicle for these companies to come public. It's interesting. By the way,

00:30:28.400 | another negative impact of this. I mean, I think if you're right, I think it's structurally

00:30:35.200 | disadvantageous to the venture capital world at large or as an asset class in an industry,

00:30:42.480 | because one of the things that happens if the best ones aren't doing it is there's no pressure

00:30:48.880 | on the next level to do it. And it kind of trickles on down. Or just the common belief,

00:30:57.360 | as you've said, comes to be that, oh, well, you're not eligible for that. Right. And so

00:31:02.720 | then you're going to have even more companies that kind of rot in place. I hate to use such

00:31:10.560 | a dramatic term, but they're just going to sit there and exist and probably dilute

00:31:19.520 | 5%, 7%, 8% a year on new equity and never get closer to the finish.

00:31:25.840 | Well, remember, the companies that can stay private, Bill, are only the very best companies.

00:31:31.600 | The vast majority of these companies don't get to cash flow break even, or they just barely

00:31:36.640 | get there. They need access to the cash in order to keep funding the business.

00:31:42.800 | They can get recapped. They can get rolled up in PE. There's a lot of things that can happen

00:31:48.480 | that don't necessarily equate to gains for VCs and LPs.

00:31:54.480 | Well, that's true. And I do think, like I said, we're going to see a lot-- I think this was the

00:31:58.640 | point of Gokul's tweet, actually. He's saying, hey, listen, a lot of these software companies

00:32:03.360 | no longer have the profile that allows them to ever get public. And so they're going to have

00:32:08.080 | to sell to private equity or do something else. They need to get real about their situation.

00:32:14.640 | And we are unquestionably in a place where, because of overfunding over the past four

00:32:20.240 | or five years-- I remember the one slide where there were more private unicorns than public

00:32:25.040 | unicorns-- there's a lot to clean up. And so unquestionably, that's true.

00:32:30.320 | The last thing I would mention on this topic is-- I mentioned it quickly earlier-- but there are a

00:32:36.000 | number of people-- and you'll hear people talk about-- that are worried that mom and pop or the

00:32:42.640 | average investor doesn't have exposure to these names. And if you're right-- and this is more

00:32:48.560 | permanent-- rather than back up and fix the regulatory costs that put us in this place,

00:32:55.840 | they'll try and fix it from the other side. They'll try and create ETFs for privates. Or

00:33:02.080 | they'll try and mandate that firms like yourself take on individual investors. And I think that's

00:33:10.240 | a complete rat's nest. We don't have to go into why. But it's not fixing the real problem, which

00:33:17.600 | is having a functional capital market that's open and inviting to more companies.

00:33:22.800 | It's an interesting debate and one that is also evolving. So this is the question of,

00:33:28.960 | does the little guy get access to the best companies in technology, for example?

00:33:36.640 | And if those companies are in the public markets, then clearly all retail investors can buy those

00:33:42.400 | companies. But it's very difficult for a retail investor to get access, perhaps, to a SpaceX or

00:33:48.560 | a Databricks, a Stripe, et cetera, the best of the private companies.

00:33:54.320 | But when you look at the evolution, Bill, of who is providing the ultimate capital,

00:34:02.560 | increasingly pension funds, which represent firefighters and teachers and police officers

00:34:09.920 | in the city of New York or the city of San Francisco or the state of Texas,

00:34:13.680 | increasingly they're LPs in these growth equity funds, the later stage funds like ours.

00:34:20.720 | So they are gaining access, the retail investor, by virtue of being a participant in that 401(k)

00:34:29.840 | or that pension. So I do think that there's probably more retail exposure to these late

00:34:36.960 | stage big companies than appears at first blush. But I will also agree with you, it's much less

00:34:43.200 | egalitarian than being able to open a Robinhood account and have access to those companies.

00:34:49.920 | Yes. Yeah. And my history of reading and watching the regulators, that argument won't be good

00:34:56.480 | enough for them. And they'll try and create what is, in essence, a new public market. Because when

00:35:02.720 | you want to have a market where everyone can trade and it's all fair, you need the rules that exist

00:35:07.920 | in the public market. So trying to have your cake and eat it too doesn't really work.

00:35:12.480 | Well, a couple other points that I would make. One is, as we saw in 2021, there were a lot of

00:35:18.880 | companies that were just walking under the assumption that private market capital would

00:35:23.200 | always be available to them. And so you don't go public because you say, oh, Masa will always be

00:35:30.480 | there. SoftBank will always be there to provide my next round of financing. And I think what

00:35:34.640 | people are starting to see as they burn through the capital that they raised during this period

00:35:40.000 | is that private markets are less liquid. Granted, the liquidity has gotten way deeper

00:35:46.000 | for the best companies. But for the average company in the private markets, that is not the

00:35:52.160 | same story. The negative reflexivity, the drawdown, the tightening up of those private markets can

00:35:58.240 | have draconian consequences on these businesses. So again, I think that there is, for the vast

00:36:05.680 | majority of companies, I still, I would conclude by saying I'm in kind of the jam and go-go camp

00:36:13.600 | that if you're a software company, you have $250 million in revenue, you're growing at 50%,

00:36:18.560 | you're approaching break-even, you think you're going to compound at 30%, 40%, 50% for the next

00:36:23.520 | five years, expand those margins. The public markets are a perfectly good place to innovate,

00:36:28.720 | to grow, to build your brand, to gain credibility. I do think we're going to see some acceleration.

00:36:35.280 | All those heads of capital markets that I talked to, they all told me their pipelines are filling.

00:36:39.920 | I know this because my team is spending more time on roadshows.

00:36:45.360 | I think there's also some off-the-beaten-path companies with pretty good numbers that are

00:36:51.120 | going to come as well. Ironically, not in Silicon Valley, because I think these means and rules and

00:36:58.720 | this kind of negativity about how big you have to be, I think those things echo much louder

00:37:04.560 | in Silicon Valley than they do externally. One thing that we have done, and I've encouraged

00:37:10.640 | LPs like endowments to think about, is we used to have these two buckets, venture and public,

00:37:18.800 | when I talk to technology investors. Then they would have everything in their venture bucket,

00:37:26.240 | all the way from a series A, all the way through Stripe or ByteDance, which obviously makes no

00:37:33.920 | sense. Those companies are over $50 billion in enterprise value, all the way up. What we talk

00:37:40.080 | about internally is there's a venture market. Think of that as a market with less than $100

00:37:45.200 | million in revenues, less than $1 billion in EV. There's a lot of mortality risk associated with

00:37:52.080 | these businesses. There's a lot of volatility associated with them. Then for companies that

00:37:58.160 | have a couple hundred million dollars in revenue and are well over that billion dollars in EV,

00:38:03.520 | we call those quasi-public. There's more liquidity associated with those businesses.

00:38:08.720 | There's less mortality associated with those businesses. These are oftentimes companies

00:38:12.960 | that could be public and are choosing to stay private. It has a whole different set of investors

00:38:19.840 | that invest in those companies, family offices, sovereign wealth funds, foundations and endowments,

00:38:26.000 | obviously big growth equity funds, private equity funds, and venture capital funds that

00:38:30.800 | are multistage. Then, of course, you have the public markets. They also have different

00:38:36.400 | characteristics in that the VC bucket, obviously, the VC chooses you. The quasi-public bucket,

00:38:44.720 | that's not everybody can participate, but there are 10 to 50 firms probably that show up around

00:38:50.640 | those party rounds and have an opportunity to participate. Then, of course, everybody can

00:38:56.400 | participate in the public market. I think that at different points of the market cycle, we'll see

00:39:02.720 | different appetites for going public, but these underlying dynamics, I do think the structural

00:39:08.880 | changes are here to stay and that there will be a category of excellent companies that can choose

00:39:14.320 | to stay private a lot longer. I don't necessarily think that's a bad thing. I would push back on

00:39:20.160 | that a little bit. I think history will show and unfortunately, we may all be in our graves by the

00:39:29.120 | time, the window you need to evaluate this, but that staying private forever has consequences.

00:39:37.680 | Unfortunately, the way it'll show up is some people will have kept some private marks in their

00:39:46.720 | very large positions and very large portfolios for a very long time and then they'll be corrected all

00:39:53.040 | at once. The learning window on that is so long that there's not a self-correcting mechanism,

00:40:04.480 | not in the near term for sure. Well, I would say the single greatest

00:40:07.760 | input to valuation and technology is growth. Number two is margin. The danger in technology,

00:40:16.800 | which is a highly disruptive industry, is that you stay private, your LPs and your investors

00:40:23.280 | don't get liquid. During that period of time where you have ultimate confidence,

00:40:27.920 | a structural change occurs in the market dynamics such that your growth rate goes down. Maybe you

00:40:33.840 | have to spend a lot more money to defend your market position. The multiple for the business

00:40:38.560 | goes down a lot. You and I remember this. There were companies in 2010, this is really pre-mobile

00:40:45.120 | taking off in the search industry that had vertical search engines that had really high multiples

00:40:51.280 | that were demolished by the transition to mobile. When that happens, those things represent

00:40:58.240 | permanent capital loss to the investors. I do think that getting people liquid,

00:41:03.120 | performing and competing in the public markets, it's certainly a terrific choice.

00:41:09.200 | I think Jamie Dimon's right. We've got way more liquid private markets and the government

00:41:14.960 | has really mucked up through all of these regulations, allowing this excess litigation.

00:41:21.760 | It has mucked up and made it less attractive to be a public company today, but let's move on.

00:41:26.160 | Let's move on. The next thing we're going to talk about, I want to kick it off. We're going to talk

00:41:33.600 | about autonomous vehicles and whether or not and how they would compete with ride-sharing services.

00:41:43.840 | The thing that I think is bringing this topic to the forefront, obviously you and I have talked

00:41:49.200 | about this for I don't know how long now, many, many, many, many, many years as both being earlier

00:41:56.240 | investors in Uber. I think the thing that's bringing it to the forefront are twofold.

00:42:02.240 | One, Tesla has had some breakthroughs with FSD-12 and they've already announced they're

00:42:09.600 | going to start talking about their RoboTaxi initiative in a more public way. Then Waymo

00:42:15.440 | is open and available in more cities and people are riding in them and so they're having experiences.

00:42:20.880 | I'll just stop there as a kickoff and let you go next.

00:42:27.040 | Great. I was up in the city last night. I was shocked by the number of Waymos I saw.

00:42:34.720 | I saw them everywhere. It reminded me, none of them had drivers. None of them had anybody in

00:42:42.480 | the front seat. They had riders in the car. I think in the city of San Francisco and these

00:42:47.120 | other cities, that was just mind-boggling when you saw it the first time. There was at one point,

00:42:53.120 | I had three cars around me and they were all Waymos.

00:42:56.720 | People have eclipse-like experiences when they get into these things.

00:43:02.960 | Listen, I think part of the setup here is it's been a tough week for Tesla. They're

00:43:09.120 | laying off 14,000 people or about 10% of their workforce. Car demand is clearly weak. That's

00:43:15.200 | on the one hand. On the other hand, there's some transformative things going on that you and I've

00:43:18.960 | talked about. These imitation models are clearly better than expected at scale. They're having a

00:43:25.520 | larger rollout of FSD. Now, I think they pushed FSD out to everybody who has the technical

00:43:31.600 | capability on their Tesla in order to receive it. They've talked about dramatically lowering the

00:43:36.240 | cost of FSD to 99 bucks a month. I think they've hit over a billion miles driven now with FSD.

00:43:43.440 | By our estimates, they're adding over a billion miles a month. I think that their confidence that

00:43:49.200 | their data advantage is increasing dramatically is going up. There's some speculation out there,

00:43:55.920 | "Oh, Elon's throwing the long ball on RoboTaxi and all this." I don't believe any of that.

00:44:01.200 | I think that the conditions that get the company re-excited about RoboTaxi is that FSD is going way,

00:44:10.480 | way better than they expected it would 18 months ago in terms of the technology, Bill, in terms

00:44:15.600 | of the technology that they have available. They've reprioritized RoboTaxi. They said,

00:44:21.360 | "We're going to move it to the top of our priority list." I think Elon announced that they're going

00:44:27.760 | to have an announcement in August. Now, of course, immediately after he does that, analysts dig in,

00:44:33.920 | and they called the regulators in a couple of different states, and the regulators said,

00:44:38.480 | "We haven't heard from Tesla." Of course, there were tweets that went out and said that a couple

00:44:42.880 | of states' regulators had not heard from this. I don't think that really means anything because,

00:44:47.920 | of course, they could launch this in any city on the planet, right? From Abu Dhabi to South

00:44:55.520 | America. Certainly, there's a city on the planet that wants to be first to have Tesla RoboTaxis.

00:45:01.040 | Go ahead. I thought, because I know you and I've talked about this, there are a number of topics

00:45:10.000 | that I think are important to consider and discuss that live beyond the technical feasibility. I know

00:45:16.480 | there are people that would say, "Okay, even if they're at four nines, they're not enough. It's

00:45:20.880 | got to be six nines," or, "Will it work in snow?" and all that kind of thing. I would say for this

00:45:27.120 | discussion, let's put that aside. Let's assume Waymo and Tesla have both achieved the quality it

00:45:34.960 | takes for the vehicle to move around by itself. What are the other things that these companies

00:45:40.800 | need to think about in order to have a successful robo fleet, if you will? Well, I'm going to touch

00:45:46.960 | on the technical thing real briefly, and then go to the other elements required to have a successful

00:45:54.640 | robo fleet. If you think about the way Waymo, you and I did this breakdown on FSD-12. We said one of

00:46:04.080 | the downsides of Waymo and Cruze is they have these deterministic models, right? That they have

00:46:09.440 | hundreds of thousands of lines of code, and they all are a rule about how they expect the car to

00:46:14.720 | behave. Well, it turns out in ride-sharing, cities, airports impose a lot of specific rules on the

00:46:24.640 | ride-sharing companies. They tell them where you can drop off, where you can pick up, how those

00:46:29.920 | cars have to behave in certain circumstances. That would seem to lend itself to a deterministic model

00:46:36.240 | where you could just write a line of code that says, "Here's what you do at an airport in San

00:46:40.960 | Francisco." One of the things I had a question on is just, can an imitation model that's imitating

00:46:48.160 | five-star drivers, can it easily have these deterministic components to it? I talked to some

00:46:56.480 | friends who are working on this, and they said, "Think of it like once the user inputs, 'I'm going

00:47:05.440 | from San Francisco to the San Francisco airport,' that they would drop that information into the

00:47:11.760 | prompt, right? That would tune the general model so that they actually have been thinking about

00:47:16.960 | how to solve that technical problem," because it certainly was one of the things on my mind.

00:47:23.200 | But okay, let's assume that they both have this cracked and that this becomes standard and

00:47:29.360 | ubiquitous that these cars can drive around. Well, one of the reasons you invested in Uber is

00:47:35.200 | that as Uber grows, the power of the network effects gets even bigger, right? And so they

00:47:42.480 | have more riders, which leads to more drivers, which reduces wait times, which creates better

00:47:49.840 | experience, which leads to more riders. And you have this virtuous cycle that's created by that.

00:47:55.760 | Obviously, that's made it very difficult on Uber's competitors, and they're achieving the natural

00:48:01.440 | market leadership, economics, and share position that you would expect vis-a-vis Lyft,

00:48:06.560 | particularly now that access to capital to Lyft got a lot harder. So we can start by just asking

00:48:16.400 | the question, "How do we think Tesla will go to market with its RoboTaxi?" In the case of Waymo,

00:48:22.720 | they've chosen to go to market, both directly, you can book it on a Waymo app, or in Phoenix,

00:48:30.320 | for example, you can book it on the Uber app. So you can book your Waymo on the Uber app.

00:48:34.800 | And so in that case, Waymo recognizes it's a very small fleet, right? In that case,

00:48:41.120 | Waymo owns the fleet, so huge cost of ownership there for Waymo. They want to drive utility of

00:48:47.520 | that fleet, so they're using Uber as third-party demand generation into that, and they have an

00:48:53.760 | economic revenue-sharing relationship between those two companies. Now, of course, that's not

00:49:00.000 | the way I would think that Elon might approach it, right? He has the potential to be much bigger.

00:49:07.360 | Tesla is a much better-known brand. There are millions of people with the app already on their

00:49:11.680 | phone. So I think most of the people I talk to, Bill, suggest in the first instance that he'll

00:49:18.320 | build an owned and operated fleet, right, where he may have outside financiers or rental car

00:49:23.920 | companies that take ownership of the car, but insofar as booking that fleet, you would go

00:49:29.520 | through the Tesla app. So, you know, we could talk about it. I throw it back to you. Let's just say

00:49:36.480 | that that's the approach that they take. What are Tesla's chances of breaking down the network

00:49:45.760 | effects that Uber currently has on rideshare demand? Well, I think you've hit on a key point,

00:49:52.720 | which is I think you have to separate Tesla from everybody else. Waymo, and I don't know if Cruz

00:49:59.600 | and Aurora still aspire to compete with Waymo or not, but they all have this very high-tech

00:50:05.360 | infrastructure with LiDAR that we've talked about, and they're all here today working on a model

00:50:11.520 | where they own and operate all the cars, which we should get into in a little bit because

00:50:16.720 | financially I think that's a complete disaster. Tesla has the benefit that they can utilize,

00:50:27.840 | theoretically, the cars that are already owned by their customers and not have to front the capital

00:50:35.200 | for each and every vehicle. If you look at the data we already have, Cruz was losing, what,

00:50:43.760 | $3 billion? Aurora's public lost $2 billion two years ago, a billion last year, and virtually

00:50:50.880 | no revenue. I have zero reason to believe, zero reason to believe, and if anyone wants to share

00:50:57.520 | data with us to correct this, that Waymo's financials don't look exactly like both of

00:51:02.560 | the ones. What do you think Waymo's worth today? They did a, we can look up, I mean,

00:51:09.440 | it's a private company. No, I'm asking you. All we can look up is what it traded for last time.

00:51:15.040 | Well, even Aurora's trading at, what, $4 billion here with no revenue and a billion dollars a year

00:51:20.640 | of losses, so there's Wall Street speculating that the IP has value. So keep going on Waymo.

00:51:28.320 | Yeah, but here's the thing. I think it really comes down to supply and demand. There is a

00:51:35.600 | beautiful emergent quality of the Uber model where supply and demand are matched and it's enhanced

00:51:46.480 | with price through surcharges, but there are plenty of people who want to make incremental

00:51:54.800 | money that know when to get on the road. I have a chart which we can put up which shows at one

00:52:00.800 | point in time the weekly pattern of ride sharing, and it's remarkably non-linear, right? It peaks

00:52:10.720 | in the morning, it peaks in the evening, it really peaks on Friday and Saturday, and Waymo or anyone

00:52:17.920 | like them with an owned fleet has, here's a huge conundrum. Are you going to build the fleet to

00:52:24.320 | average or peak? And by the way, you lose whatever your answer is. If you build it to average,

00:52:30.880 | you're not going to be able to serve your customers during the peak at all. They're only going to be

00:52:35.440 | disappointed. And if you build it to peak, you're going to have a bunch of very, very, very expensive

00:52:41.440 | CapEx sitting around doing nothing most of the time. Okay, so I think this is the most critical

00:52:46.160 | point, right? Because this is the same challenge that Amazon faced, which led to the creation of

00:52:52.800 | AWS. This is the same challenge that most businesses face, which is they have very spiky

00:53:01.360 | demand periods. And if you want to be an incredible customer experience, you have to have something

00:53:06.720 | that can regulate to those higher periods of demand. Like you said with Uber, they have a

00:53:12.560 | natural thing. They can just charge a little bit of peak pricing. They got more people who get out

00:53:17.680 | of their houses, drive their cars, pick you up, wait times don't go up that much. And they benefit

00:53:22.640 | from that. And as the network expands, it gets even better. Well, and you remember, it goes back

00:53:28.160 | to what was that word when Airbnb first came out? It was asset sharing. It was a different word that

00:53:33.360 | was used. But there's a huge amount of cars, especially in the US, they're sitting mostly

00:53:40.880 | on it. And so Uber gets to take advantage of that and utilize that CapEx. And Tesla may have a

00:53:49.280 | similar advantage just because they have enough customers. Someone trying to build a completely

00:53:54.720 | separate fleet, I just don't think it works. And I would encourage anyone that believes in it or

00:54:00.400 | anyone that's a, let's call them a Waymo advocate, let's build a public Google Sheet of the 20 year

00:54:09.120 | financial statements for this thing. And let's put it out there in the public. You're going to need

00:54:13.280 | hundreds and hundreds of billions of dollars. And I don't know how you answer this peak versus

00:54:19.200 | average question. I don't know how you do it. So let's talk about some of the ways that might

00:54:24.000 | happen. As a Tesla owner, if I had the opportunity to put my Tesla in a pool and allow it to be

00:54:32.240 | pulled out during peak periods of time for ride share, I'd have a problem with that because I

00:54:37.600 | need my car during those periods of time. Well, it's very likely, here's another way of saying

00:54:44.480 | what you're saying, which is it's very likely that the average Tesla owner's usage map is

00:54:51.040 | similar to the Uber map that we just put on the screen. Of course, this is my point. So here would

00:54:57.600 | be a straw man. Tesla launches with an owned and operated fleet to prove the model, to start

00:55:04.720 | building the customer affinity with the product, the customer trust. Which would unquestionably

00:55:11.520 | work because there's so many Tesla fans that would want to do it just to be supportive of the team.

00:55:19.840 | And by the way, it's a great product. It feels great. You probably already have a Tesla,

00:55:24.560 | so you feel good about FSD. So my hunch is they launch it on... You might even launch it at first

00:55:31.040 | with only allowing Tesla owners or something like that. Oh, interesting. It's almost like a benefit

00:55:37.600 | of ownership. So then I think what they do is, my hunch is that they will open it up and partner

00:55:46.080 | with Uber. I think Uber would like that. I think Tesla would like it. And here's my argument for

00:55:52.960 | that. The only way you solve the peak demand problem, Bill, is that you have to basically

00:56:02.160 | have high utilization at lower periods of demand. You have to reduce the delta between your trough

00:56:09.280 | and your peak. And the way you do that is you pull in the Uber demand when your demand is lower.

00:56:16.160 | And so an example of that would be, if you think about in the city of New York,

00:56:21.760 | McDonald's has an app. You can go on the McDonald's app and you can order a burger,

00:56:26.320 | right? Or you can go on the Uber Eats app or the DoorDash app and order a burger. And so

00:56:34.480 | McDonald's recognizes that they want to try to flatten that demand out over the course of the day.

00:56:40.720 | And so they can yield manage that by having both a 3P and a 1P strategy. So it seems to me that the

00:56:47.600 | only way you really crack this is have a 3P and 1P strategy. And the upside to that, in addition,

00:56:54.800 | is that now you have a flywheel of Uber drivers potentially building some mini fleets of their

00:57:03.760 | own. You could imagine somebody who says, hey, this is pretty accretive. We saw this with black

00:57:08.880 | cars, I know, in the early days, where you had one operator that may have two or three or four

00:57:14.160 | black cars. You could imagine somebody who says, listen, I can make money by arbing this system

00:57:19.440 | and by having multiple Teslas in that fleet. So I think they probably launch it on a standalone

00:57:30.160 | basis. They get a lot of excitement and a lot of customer love for it. But my sense is, by the time

00:57:35.120 | you scale this, it probably does have to be an open system. We'll see. Yeah. And I mean, that's

00:57:40.640 | clearly up to the powers that be at Tesla to make that decision. The two companies have partnered

00:57:46.880 | to date already. They have quite a number of initiatives, including like free charging for

00:57:53.760 | people that buy them and put them on. So there's a there's a close partnership already. But,

00:57:58.720 | you know, as as this plays out, we may see, you know, reasons why they may not want to.

00:58:04.720 | Can we talk a little bit just about the unit economics, you know, of these different businesses?

00:58:10.640 | Yeah, there's one thing I want to bring up specifically that that'll be interesting to

00:58:16.560 | see how it plays out. And that's insurance. And so last time I had specific knowledge,

00:58:23.680 | and it's been four or five years, but I doubt it changed that much. The cost per mile of insurance

00:58:30.000 | for Uber in the U.S. is dramatically higher than the rest of the world. And if you look at just

00:58:37.440 | the cost of consumer car insurance, the same thing's true, way higher in the U.S., which gets

00:58:43.200 | to another problem with regulatory costs that our government should try and fix. But anyway,

00:58:52.400 | it strikes me and I would be worried about this as an investor. It strikes me that the

00:58:59.680 | that a particularly aggressive litigator lawyer would chomp at the bit to get in front of a jury

00:59:09.280 | on a robot killing a human and maybe even create more litigation costs and a higher cost of

00:59:17.680 | insurance for a robo-taxi than a driver. Now, I know how people, our listeners, are going to react

00:59:24.400 | to that. They are convinced in their heart of hearts and in their brain that these are safer

00:59:30.160 | than human drivers. And I am saying this even with that potentially being true. We may have

00:59:39.760 | just such a broken litigation situation in the U.S. that the actual cost of underwriting the

00:59:48.320 | insurance for a robo-taxi may be higher than a real car. And time will tell. And as we know,

00:59:55.280 | Tesla's already brought some of their insurance operations internally. And if they're willing

01:00:01.040 | to underwrite it themselves and fight it themselves, maybe that's something they'll do

01:00:06.080 | as a part of this rolling out. I got to take a tangent here for a second and then I'll come back

01:00:11.280 | to these unit economics. You know, I was with a certain investor, a group of investors in the

01:00:16.400 | Midwest, let's just say, over the weekend that are very big investors in the insurance business. And

01:00:21.200 | I asked him the question, I said, what's up with car insurance being up 22% in the CPI on a year

01:00:27.120 | over year basis? I mean, it was crazy. Car insurance up 22%. And he said, here's how it

01:00:32.320 | works, Bill. In 20 and 21, we have very predictable and very steady number of miles driven in this

01:00:40.320 | country on a year over year basis. And the accident rate remains very steady. So the cost

01:00:44.400 | of our auto insurance has largely been pretty flat, growing with GDP. What happened in 20 and 21,

01:00:51.360 | the number of miles driven fell a lot. And so it fell so much that those companies were prone to

01:00:59.760 | over earn, right? Because there were just, there's less chance of an accident with fewer people

01:01:05.040 | driving, or at least that was the assumption. But what turned out to be true is that the miles that

01:01:09.840 | came off the road for some reason were safe miles and the accidents per mile driven actually spiked

01:01:17.040 | a lot. So much so that the insurance companies lost a lot of money during this period of time.

01:01:22.480 | So they went to the state regulators who control insurance pricing and they said, hey, we got to

01:01:27.280 | raise the price of insurance because we're losing money. Now, remember, auto insurance companies

01:01:32.960 | try to price the insurance to make about, you know, 3, 4, 5% on the insurance and the rest of

01:01:39.040 | the money they make is on the float. Okay. But in those years, they were losing 5 to 10% on the

01:01:45.520 | insurance, right? So the companies were actually losing money even with the float. How long do you

01:01:51.680 | think it takes the California state regulator to pass along an insurance increase, right? You have

01:01:56.640 | to threaten them that you're going to leave the state and all this other stuff, and then eventually

01:02:00.160 | they do it. So now they passed on this insurance increase and it just kicked in, which is what you

01:02:05.280 | saw in the CPI, right? And ironically, and not surprisingly, the patterns of driving have now

01:02:12.240 | returned to pre-COVID levels. So we went from a period where the auto insurance were not charging

01:02:17.040 | enough because they had to beg the state regulators and they were slow, and now they're charging too

01:02:21.440 | much. But set that aside for a second. Let's come back to the unit economic. And by the way, I think

01:02:26.720 | this cost, this is a lot of guesswork. I apologize. People should treat it as such. But I think in the

01:02:33.280 | Uber case in the US, about 5% of gross revenue goes to insurance. Yeah. So it's an expensive

01:02:41.840 | input. And we took a crack. Okay. So let's just say at the high level, most people think that

01:02:47.440 | taking the driver out of the Uber will lead to an explosion in profit margins because it represents

01:02:53.840 | 70% or 75% of the cost of the ride. But we actually took a crack at these unit economics. And what I

01:03:03.040 | think that fails to understand is a couple of things. Number one, if you just break this down

01:03:08.400 | per car, we estimate that there's about $140,000 of revenue per car driven on the Uber network. Now,

01:03:16.080 | of course, this is blended across tons of different markets, tons of different cars, etc.

01:03:20.480 | So $140,000 of GBV. When you do that in self-driving, we think it goes down to about $100,000. Why?

01:03:27.840 | Well, you have actually more rides in the self-driven car, right? But it actually comes

01:03:33.840 | in at a lower cost because you put so many cars on the road that the charge to the consumer gets

01:03:40.240 | shared back with the consumer, right? Margins come down as you have more competition because you have

01:03:44.880 | more cars on the road. Next up, the revenue per car. Well, of course, you take out 70% of the cost,

01:03:53.040 | the driver cost, on the driver-driven car. So at a revenue level, we think that if you were

01:04:01.360 | comparing apples for apples, in the case of the one with a driver, they make about $40,000 a car.

01:04:09.920 | Without the driver, they would make about $90,000 per car. This is just revenue.

01:04:14.560 | So about double the revenue. But the costs are way higher in the full self-driving case.

01:04:21.120 | Obviously, first up is you have to pay for the car, right? There are massive costs associated

01:04:26.400 | with that. Maintaining the car, the insurance on the car, all the things the driver currently does

01:04:32.400 | now is dropped upon-- Well, yeah, and just on that first one, because

01:04:38.320 | I can go buy a two-year-old used car for $25,000, $30,000 and get on the road,

01:04:45.920 | and then you add in the cost to the human. You need the fully autonomous car to be really cheap

01:04:55.440 | in order-- because if it-- and I don't think anyone thinks the current Waymo when you were

01:05:01.520 | surrounded by the other day is anywhere close to even $100,000. No, those cars cost probably

01:05:05.120 | $150,000. It's not viable. So that car would need to get down towards this $30,000 number for that

01:05:14.720 | disruption to matter. If it's in the $45,000, $50,000, and it's an alternative to $30,000 plus

01:05:20.720 | a driver, you have no-- But again, where Tesla has an advantage, they probably can put a car

01:05:25.920 | on the road at those lower price points. So when we shake it all out, and I'm happy to--

01:05:31.280 | But by the way, there's other things the driver does. I mean, the driver maintains the car.

01:05:35.920 | So once again, this is in the Tesla case, if it's not an owned fleet, it's a customer-owned fleet,

01:05:42.960 | they can clean it, and that'll just be a cost to putting your car out on the road.

01:05:47.280 | But in the case of a Waymo, like that car has to come back to headquarters at some point.

01:05:54.480 | Someone's got to get in it. Someone's got to clean it. There's eyes that are going to be needed.

01:06:00.080 | So there's going to be a customer service staff that's going to have a camera because you're

01:06:04.480 | going to have to settle disputes. You're going to have to deal with customers. And the human does

01:06:09.680 | that in the case. So there are going to be other costs that exist. So the point is, you have to

01:06:14.640 | take the 70% that you save on the driver, and then you have to add all these new costs that you're

01:06:20.560 | going to incur. We still think that the margin goes up from, let's call it, 7% of gross bookings

01:06:28.720 | to something like 12% or 14% of gross bookings, even taking into account all of that bill.

01:06:34.240 | But that's what we're all trying to do. We're all trying to build a forecast and understand,

01:06:39.040 | you know, assuming that this could work, assuming that regulators pass it, assuming you could build

01:06:44.320 | the $25,000 or $30,000 car, assuming that you could enter into a commercial relationship with

01:06:49.840 | Uber and yield manage the peak and, you know, demand. All of that is to say, there are a lot

01:06:55.920 | of needles that have to be threaded in order for this to work at scale. And I think if it does work

01:07:02.000 | at scale, I don't, you know, obviously, talking our book a little bit here, I don't think this

01:07:07.520 | is much of a threat to Uber. In fact, I think if it works for Tesla, I think there's an opportunity

01:07:13.360 | for both companies. And by the way, like another way I'd restate what you just said, which is,

01:07:22.080 | it's not going to be financially disruptive at the beginning or in the middle. It might

01:07:28.160 | theoretically be financially disruptive at the very end. But there's way too much capital that

01:07:34.960 | and the price of the car has to come down. There's too much that has to happen. There's no way for

01:07:40.640 | someone's not going to come out tomorrow with a robo fleet and underpriced Uber.

01:07:46.080 | Well, and let's put one final point here. Remember this deal that Tesla entered into

01:07:51.920 | with Hertz where Hertz bought a bunch of Teslas? It turned out that the resale value on those cars

01:07:56.960 | was dramatically less than they had forecast. And the CEO lost his job over this. I think

01:08:02.160 | the write down for Hertz was like a billion dollars over this mis-forecast. So if you're

01:08:08.160 | in the business of buying and maintaining that fleet on behalf of Tesla, you're doing so with

01:08:13.840 | full knowledge that all of this is hard to forecast. So you're going to bake into your

01:08:17.680 | model a huge margin of safety, right? So that you're not going to have the problem that Hertz

01:08:22.880 | just ran into. I would say based on everything I know and we're not in the boardroom with these

01:08:31.760 | guys, but like letting the customers own the Teslas that own the fleet is a better alternative.

01:08:39.360 | Much, much, much, much better. And I will say this, the fact that we're even talking about a

01:08:45.520 | robo taxi fleet shows you the incredible progress that Tesla has made. And what we talked about on

01:08:53.600 | the show a few weeks ago, the month to month improvement in the model, the data that they're

01:08:58.160 | collecting. And just the anecdote I shared from last night, I am more convinced than I ever have

01:09:04.320 | that we're going to see tons of cars on the roads with no drivers in the years ahead. You and I are

01:09:11.520 | just speculating as to how the business model might come together. But by the way, one last

01:09:16.720 | thing, I'm so sorry. I do think there's an argument that you might see this be prolific

01:09:23.440 | in a different country soon. Of course. Of course. Both China and India have much higher road deaths

01:09:29.840 | per mile. So you can see the government wanting it to happen. China being more autocratic could

01:09:37.840 | just wave their hand and change insurance rates, mandates. They could just make it a reality.

01:09:44.000 | Whereas we in the U.S. with the way our litigation and regulation work are very unlikely to clear the

01:09:51.840 | decks and make way for something. You know, I also think, you know, as I mentioned earlier,

01:09:58.160 | Tesla and Uber coexist in a ton of markets around the world. And there are a lot of markets that,

01:10:04.240 | frankly, are more innovative than the United States, have less regulatory headwinds. I could

01:10:09.520 | see this happening in Abu Dhabi or Dubai. I could see this happening in Singapore. I could see this

01:10:15.280 | happening in South America. I could see this happening in Mexico. There are a ton of jurisdictions

01:10:20.880 | where I expect this, unfortunately, before I expected in the U.S., you know, at scale.

01:10:27.600 | All right. Before we wrap up, let's check in on markets since we talked last. What are

01:10:33.360 | you seeing out there, Brad? What do we need to be worried about?

01:10:36.560 | You know, if you telescope way out, we're about ready to enter earnings season, right? Next week,

01:10:42.720 | we have Tesla report. We have Facebook reports, Microsoft reports, Google reports. So, you know,

01:10:50.720 | thinking about the setup heading into earnings. If you telescope way out, the S&P and the Nasdaq

01:10:56.480 | are up about 6% year-to-date, Bill. The small caps are down about 3%. And it really comes down

01:11:05.680 | to a lot of those, the companies that are up in tech have beat their numbers, right? NVIDIA's

01:11:11.440 | beat their numbers. Microsoft's beat their numbers. Google's beat their numbers. And the companies

01:11:15.680 | that are down, Tesla and Apple have largely disappointed. You know, but against this

01:11:21.600 | backdrop of the Nasdaq being up 6%, we've had a huge move in the 10-year, right? We've talked

01:11:27.520 | ad nauseum over the course of the last couple of years about how interest rates are the economic

01:11:32.800 | law of gravity for the financial markets. And since the start of the year, the cost of that 10-year

01:11:39.120 | is up 20%. We've gone from 3.9 to now 4.7 today. So, just all else being equal, right? You would

01:11:47.200 | expect that growth company multiples would be down 10% to 15% because the 10-year is up a lot.

01:11:55.440 | So, I think the thing that's making a bunch of people nervous is usually you have this inverse

01:12:02.080 | relationship. As rates are going up, you know, that's a headwind for tech. You might expect

01:12:07.680 | that the Nasdaq would actually be down a little bit on the year in relationship to the cost

01:12:14.960 | of the 10-year going up. So, if you dig in a little bit to, you know, to these businesses and

01:12:22.080 | say, "What has to happen in order for these stocks to work next week?" It really comes down to this.

01:12:28.880 | The margin in 2023, all you had to do is be less bad, right? It was that reversion to the mean

01:12:34.640 | trade. But now, the only way your stock's going up is you got to beat numbers, right? Your numbers

01:12:40.080 | have to get revised us upward. And if you miss numbers, it'll be brutal to the downside. You

01:12:45.600 | know, the team's made a chart on, an interesting chart on cloud acceleration. So, what might be

01:12:51.040 | driving that acceleration? We've talked a lot on the show about AI, but check out this chart.

01:12:57.040 | So, you can see for hyperscalers, you can see that COVID bump. So, this is absolute

01:13:03.760 | dollars that were added to the three large clouds over the course of the past several years.

01:13:09.600 | So, you see how that really bumped up during COVID and ZURP. And then, you can see the belt

01:13:15.360 | tightening period as interest rates went up, people got a lot more nervous about the economy.

01:13:19.920 | These are CapEx numbers?

01:13:21.920 | Right. No, these are the absolute dollars of revenue that were added to those companies,

01:13:27.280 | right? So, think of this as year-over-year revenue growth. And so, we dipped. It started to go up

01:13:33.760 | over the course of the last couple quarters. And now, consensus is forecasting that it continues

01:13:39.200 | to go up, you know, above trendline. So, those companies are going to have to deliver. That's AWS,

01:13:46.720 | that's Microsoft, that's Google. And, you know, our belief is that they will. We think they will

01:13:51.840 | beat those forecasts because we do see that real acceleration coming out of AI. Now, look at this

01:13:57.760 | chart. This is that same chart, Bill, for Datadog, Mongo, Confluent, Elastic. Think of it as the data

01:14:05.360 | infrastructure companies. They got a much bigger benefit during COVID, but they also had a much

01:14:10.800 | bigger pullback. Now, look here what the consensus is forecasting. Unlike for the hyperscalers,

01:14:16.560 | the consensus is saying, "No, we don't buy that they're going to re-accelerate and get back above

01:14:21.360 | trendline." We think they still have, you know, they have some challenges ahead, and they're not

01:14:26.960 | going to get back up to those COVID levels of absolute revenue dollars that were added on a

01:14:32.160 | year-over-year basis. So, I think that's an interesting divergence, you know, that you'll

01:14:37.040 | have to keep your eye on when Snowflake and Mongo report. We actually think a bunch of those

01:14:43.200 | companies are seeing the flow through out of AI as more enterprises are moving more of their data

01:14:48.720 | into the cloud, but there's clearly tension in the market about that. You know, one other thing that

01:14:56.720 | is really interesting to me is we universally see companies holding the line on expenses,

01:15:04.880 | right? And so, think of this as, you know, we had real questions, was the age of efficiency

01:15:10.640 | going to be, you know, a moment in time, or were companies going to continue to do that?

01:15:15.520 | I was talking to the CEO of a major company the other day, and, you know, with tens of thousands

01:15:22.000 | of employees. And he said, over the course of the next three years, we'll grow our top line 50%,

01:15:27.360 | and we'll reduce our personnel costs by 10% to 20%, right? That's extraordinary. Like,

01:15:34.960 | I don't think we've had a moment like the moment we have today in terms of margin expansion,

01:15:40.960 | right, in any time frame that I can remember where it was voluntary. It's not because we're

01:15:47.920 | in a recession, it's not, but it's because companies really do see the opportunity,

01:15:53.120 | both they had some, you know, they have room to cut because they got excessive in 20 and 21,

01:15:59.680 | and because they now have these tools, co-pilots, soon to be co-workers in engineering,

01:16:04.800 | that enable them to be more productive, more efficient in engineering,

01:16:08.800 | and also allow them to be much more efficient. And that includes Slack and Zoom and things that

01:16:16.160 | allow people to have more employees outside of the Bay Area. I mean, every one of these CEOs

01:16:22.400 | I talked to are intentionally trying to get headcount out of the Bay Area.

01:16:28.640 | Which, in response to our SBC pod that we did, I had a, you know, a CEO of a major

01:16:35.120 | company call me and say, "Hey, we're rebalancing how we're thinking about SBC, you know, across

01:16:42.320 | the business." And I said, "Doesn't that make you nervous that you'll lose your best engineers,

01:16:46.960 | you know, if the total cost of SBC comes down?" And he said, "No, because we don't need as many

01:16:53.280 | engineers. The labor market for engineers is softening. It's easy for us to keep and retain

01:16:58.880 | the best talent. And so now is exactly the time to get our comp plan reset for the next five to

01:17:04.880 | seven years." So, Net, I would say that I'm really excited about both the reaccelerating

01:17:12.880 | top line for some of these businesses. Again, this is very unique to tech. You know, I see

01:17:18.400 | other pockets in the economy, like we just talked about autos. I mean, it's a tough story out there.

01:17:24.000 | Higher interest rates, less consumer demand, people have burnt through their STEMI checks,

01:17:29.520 | they have less savings. And so I think it is a tale of these two worlds. But I think one of the

01:17:35.040 | key drivers over the course of the next several years in tech is going to be these expanding

01:17:40.240 | margins caused by, you know, the tool set that they're investing in around AI.

01:17:45.600 | Well, here's the good news. You've now set the stage for earnings. So when we get back together

01:17:52.640 | in two weeks and this stuff starts coming through, we'll have a lens and a framework

01:17:57.760 | for what to talk about. And I'll get a report card.

01:18:00.320 | Okay, fair enough. Great seeing you, man. A lot of fun.

01:18:04.480 | Good to see you. Take care. Bye-bye.

01:18:05.840 | As a reminder to everybody, just our opinions, not investment advice.

Ep7. The Great IPO Debate, Tesla Robotaxi v Uber, Tech Check | BG2 with Bill Gurley & Brad Gerstner

Chapters