Latest / The Learning Corner by Precursor / Episode #77: The Broken Seed Model, YC's Revenue Honesty Rules, China Blocks Meta's Manus Deal

Transcript

Episode #77: The Broken Seed Model, YC's Revenue Honesty Rules, China Blocks Meta's Manus Deal · The Learning Corner by Precursor

• 0:00
  When you're in a very competitive environment, standards go down.
• 0:04
  We saw that in the mortgage crisis.
• 0:06
  I think we saw that in 2021.
• 0:07
  I think we're seeing a version of that right now too, which is that I think people wanna get
• 0:11
  into hot day ideals.
• 0:12
  And if a company gives you a revenue number that looks passably accurate, if you're the
• 0:17
  person who really digs in, you might not get the deal.
• 0:19
  In the long run, that might be the optimal thing, is to actually not get the deal.
• 0:23
  In the short run, that is not what people want.
• 0:26
  It's the week of April 27.
• 0:28
  Welcome to the Learning Corner, a Precursor Ventures podcast.
• 0:33
  Welcome to the Learning Corner by Precursor.
• 0:35
  I'm Mia Farnam, and I'm joined by RGP and managing partner Charles Hudson, where we will
• 0:40
  be talking through the articles and Goodreads that caught our team's attention over the past
• 0:44
  week.
• 0:44
  Our first piece this week is a post on X by Lucas Vaz titled The Narrow Path, talking about
• 0:49
  what is actually happening to seed investing right now.
• 0:52
  The core argument is that the venture ecosystem has split into two completely different games,
• 0:56
  a top tier where the best founders raise at 40 to 100,000,000 post money from brand name funds
• 1:02
  and everything else which Lucas calls a capital black hole where money goes in but paths to
• 1:07
  liquidity are nearly nonexistent.
• 1:09
  He uses a data point to anchor the whole piece that the original angel investors in OpenAI put
• 1:14
  in roughly 10,000,000 collectively and are sitting on about 1,400,000,000 in paper value
• 1:19
  today, which sounds great until you see that it's only a a 140 x return on perhaps the most
• 1:24
  consequential company of the decade and that the traditional seed model simply cannot
• 1:28
  survive if the biggest winner of the era only returns your fund 1.4 x.
• 1:33
  His prescription is a direct challenge to the conventional wisdom that smaller fund size
• 1:37
  should drive lower entry prices and more diversification.
• 1:40
  Position concentration, he says, matters far more than ownership percentage.
• 1:44
  You have to be willing to pay up for quality, and picking has always been the job.
• 1:48
  The industry just spent a decade building models that disguised it.
• 1:51
  I read this piece and found a lot of the takeaways genuinely insightful, and I know we
• 1:56
  just talked about it together, like, five minutes ago before we started recording.
• 1:59
  But I'm not sure fully that I buy the argument that seed as a construct is the problem.
• 2:04
  There are deals happening today with structures that we've never seen before that are changing
• 2:08
  what the outcomes look like for that stage.
• 2:10
  Wanted to pause for your thoughts on this one, Charles.
• 2:13
  Yeah.
• 2:14
  I think Luca's article definitely made the rounds at a bunch of GP group chats that I'm
• 2:18
  in.
• 2:18
  Whenever you touch on, like, what works in seed, you're going to get a lot of attention
• 2:22
  and reaction from folks.
• 2:23
  I read the post and a couple things.
• 2:25
  I don't think we always know what the big outlier outcomes are going to look like.
• 2:30
  I think about Groons, like, was not a company that most people had on their bingo card.
• 2:35
  And so I think seed is, for me, if you zoom out of all of this stuff, seed is supposed to be a
• 2:40
  really high multiple, high risk, high return part of the venture asset class, which to me
• 2:47
  means you should be able to get multi 100 x returns on your single company winning
• 2:53
  positions Mhmm.
• 2:54
  On first check.
• 2:54
  Now if you put more money in, of course, the it's it's likely that the aggregate return will
• 2:58
  come down.
• 2:59
  I think seed is very challenged for dedicated seed managers.
• 3:02
  But I think you have to go back to this other point about bifurcation, which is there are
• 3:06
  some people for whom seed is not their core business.
• 3:09
  And making money on seed is a necessary but not sufficient thing.
• 3:13
  Or it's maybe sufficient but not necessary, actually.
• 3:15
  If being a player in seed gives them visibility into the market, allows them to see what people
• 3:19
  are working on, informs some of the checks they the big checks they write at series b or c, and
• 3:25
  the price to acquire that information is a relatively de minimis part of your AUM, that's
• 3:30
  a very different strategy.
• 3:31
  And I think, really, you have two different games you can play in seed right now.
• 3:36
  One game is invest in the hottest, buzziest consensus stuff, and hope that despite the high
• 3:42
  prices, that on exit, you're still happy with the multiple.
• 3:44
  We simply don't know if that strategy works at scale.
• 3:48
  It works if you're a whiz.
• 3:49
  We know that.
• 3:49
  Like, if you do a 100 and get it at 23,000,000,000, that works.
• 3:52
  I think there's no question that that works.
• 3:54
  The question with those is, like, how representative are the outliers of the average?
• 3:58
  And I think by very nature of calling them outliers, I think kind of answers the question
• 4:02
  to some extent.
• 4:03
  The question is, is seed a good business to be in as an investor if on average you're paying
• 4:08
  high prices, but on average you're not getting outlier exits?
• 4:11
  Well, it's not a bad business, but you're also probably not a good picker.
• 4:14
  And so for me, the question is, like, if you're gonna do the high consensus deals that are
• 4:19
  expensive, they have to be extraordinary outliers, and you have to find a way to get
• 4:23
  enough of those companies that matters.
• 4:24
  But the truth is, you know, 11 people can't own 10% of a company.
• 4:27
  The math simply doesn't support it.
• 4:29
  And so I think that's the challenge with that strategy.
• 4:32
  The other strategy, which I think is closer to what we do, is if you pick things that are
• 4:35
  maybe less sought after initially, you have follow on fundraising risk, or you have the
• 4:40
  cocktail party risk where you're in a bunch of companies that you tell people that you're an
• 4:43
  investor, and they look at you blankly and say, I've never heard of those companies, and I have
• 4:46
  no idea why you'd wanna invest in them.
• 4:48
  So to the extent that people thought seed was easy, I hope that this post disabuses them of
• 4:52
  that notion.
• 4:53
  And I think seed is hard.
• 4:54
  You have to pick really well, and you have to own enough of the companies that you choose.
• 4:59
  And part of this is seed funds have gotten bigger, which means owning enough, in quotes,
• 5:03
  has become even more important because the scale of the fund and the amount that a winner
• 5:08
  needs to return has gone up.
• 5:09
  I think we talked about this before with Micah's post.
• 5:12
  It's a really bad idea to apply, like, seed math to a series a fund.
• 5:16
  They're they're designed to do different things.
• 5:18
  And if seed becomes buying big chunks of these companies at high prices and you lose the
• 5:23
  ability to get the really high multiples, I'm not sure it's the same business.
• 5:26
  And if you're more of, like, a tourist in seed and you make most of your money in alpha in
• 5:32
  other stages, then I imagine that maybe the point of participating seed also maybe is not
• 5:37
  for all of the returns.
• 5:38
  It's all having a pick for what's coming up and then being able to push in more capital in
• 5:43
  larger quantities in later rounds.
• 5:45
  I think the biggest bifurcation I see in the market right now is what chunk of venture
• 5:51
  returns are going to come from consensus, deals that are born consensus hot.
• 5:55
  And I think all of these structural incentives as a fund manager are to pile into consensus
• 6:01
  hot deals.
• 6:02
  Because if you do 10 deals with the big multi stage funds, when you go out to go raise your
• 6:06
  next fund, you will probably have markups, and you will probably be able to say, hey, I'm in
• 6:10
  10 deals with these great firms that you know.
• 6:11
  That has nothing to do with company quality.
• 6:13
  But in the short term, if you're trying to fundraise or you wanna be cool, like, that's a
• 6:17
  very good strategy to pursue.
• 6:19
  But in the end, it doesn't mean that you're gonna actually make any money on those
• 6:23
  companies.
• 6:23
  But I think there's a lot of people who said, oh, well, this has worked in spurts before, so
• 6:28
  maybe this is the new way to play CWIC, but maybe not.
• 6:30
  He kind of makes this claim that the era of easy, low priced, diversified venture investing
• 6:37
  is over.
• 6:38
  I think that only makes sense if you believe that just because you're coming in low price,
• 6:43
  it means that you have no downside to everything, and you kind of just pick a bunch
• 6:47
  of things in a range of price.
• 6:48
  You still have to pick Yeah.
• 6:49
  Even if things are at five post.
• 6:51
  That's just part of the game.
• 6:52
  There's a subargument in here that I read, and I'm like, maybe I'm the only person who thinks
• 6:56
  this.
• 6:57
  If the belief is that the blue chip assets are obvious and knowable seed.
• 7:01
  You don't actually need seed funds.
• 7:03
  You don't need them.
• 7:03
  I think if you're like, yes, if we know at seed all of the companies that will matter, then the
• 7:08
  only strategy that's worth pursuing is to get into those companies and buy as much as you
• 7:12
  possibly can.
• 7:13
  I just reject the premise.
• 7:14
  We've just also seen it work in other ways too.
• 7:17
  Yeah.
• 7:17
  If we didn't see it working, then I think you would have fewer data points to make that
• 7:21
  argument, but but we do.
• 7:22
  Our next good read is a post on X from Gary Tan, CEO of Y Combinator, sharing YC's official
• 7:28
  guidance on being truthful and precise about revenue metrics.
• 7:31
  The crux of his message that nonbinding LOIs, GMV, bookings, and contracted ARR,
• 7:37
  transactional revenue, and true monthly or annual reoccurring revenue all completely
• 7:41
  different things and should never be used interchangeably.
• 7:44
  Gary is direct that this is not just a communication preference.
• 7:47
  It is a matter of trust with investors and founders who blur these lines even accidentally
• 7:52
  creating real problems for themselves down the road.
• 7:55
  The post reads almost like a field guide.
• 7:57
  It's ranked from least to most impressive with the explicit call out that annual revenue run
• 8:01
  rate should never be abbreviated as ARR since ARR means something specific and reoccurring.
• 8:07
  There was a lot of discourse in the comments if you looked in x that the fact these were
• 8:11
  written out was offensive to founders, and there was a lot of finger pointing to investors
• 8:16
  and media players for creating an ecosystem where founders feel like they have to become
• 8:20
  creative with their metrics in order to stand out.
• 8:23
  Wanted to hear your thoughts on this all.
• 8:26
  First of all, I'm glad they wrote this because YC saying it is authoritative and people will
• 8:31
  pay attention.
• 8:32
  I'm embarrassed for our industry that this needs to be written.
• 8:35
  These are not judgment calls.
• 8:37
  Yeah.
• 8:37
  These are terms that have specific meaning and have had specific meaning for a long time, and
• 8:42
  we shouldn't even have to have this conversation.
• 8:45
  Yet, here we are.
• 8:46
  Mhmm.
• 8:46
  A lot of the comments are like, well, whose fault is it?
• 8:48
  I'm like, well, it's everyone's fault.
• 8:49
  I think there are some founders who are not as well steeped in math and in the meaning of
• 8:56
  these terms, in some cases, who might have genuinely gotten them wrong.
• 8:59
  I think the AI hype cycle has made really vertical revenue the sort of lingua franca if
• 9:05
  you wanna get media attention or investor attention.
• 9:07
  And we talked about this maybe a year ago.
• 9:09
  The predictable outcome is that people were gonna try to find a way to fudge, shape, or
• 9:15
  present their numbers in such a way that they can fit under whatever the high growth scenario
• 9:20
  umbrella is at the time, just so that they can get considered.
• 9:23
  Some founders I've talked to, well, investors are smart enough to know, to see through this
• 9:27
  once they get in there.
• 9:28
  But I if I don't have high growth, I don't get the meeting.
• 9:30
  And I think some investors are like, yes.
• 9:32
  I am smart enough to see through this.
• 9:34
  But this is a high growth company.
• 9:35
  If I push hard and due diligence and I don't get in and it continues to go up, I was like
• 9:40
  morally or intellectually right.
• 9:41
  I didn't get into a deal that matters.
• 9:43
  And plus every other company out there is maybe engaging in similar shenanigans.
• 9:47
  And you end up with a system where, like, nobody wanted to pull the compound cord and
• 9:51
  say, hey, let's just stop this.
• 9:52
  I don't know that this will change much, but at least it creates the ability to have a
• 9:57
  conversation about, like, what do these numbers actually mean?
• 10:00
  And I think a lot of investors were aware that founders were calling ARR was maybe contracted
• 10:05
  ARR, but not deployed ARR, or maybe it was services revenue from Ford to put.
• 10:10
  I don't think investors were, like, being duped or naive.
• 10:12
  I think founders might have been in some cases.
• 10:14
  I think it's bad for communication industry when revenue and ARR don't mean the same thing
• 10:19
  to everybody.
• 10:20
  I've had tons of conversations with founders where it's certainly not malicious, but I mean,
• 10:24
  it's now become and it should have always been part of practice to ask additional questions
• 10:29
  when you mean contracted ARR.
• 10:30
  What's the length of the contract?
• 10:32
  Is it a month long contract, or is it a multiyear contract?
• 10:35
  Contracts right now in today's AI age have completely changed in language.
• 10:39
  Most of them are like actually pilots, but you call them contracts so that
• 10:42
  you can put it
• 10:43
  into contracted ARR language.
• 10:44
  And they have kill clauses and no re ups involved, so it does muddy the water.
• 10:49
  But as investors, you should always be double clicking.
• 10:51
  But so early on, you would have thought it's not as murky as before.
• 10:55
  I think when you're in a very competitive environment, standards go down.
• 11:00
  I think we saw that in the mortgage crisis.
• 11:03
  Standards went down.
• 11:04
  I think we saw that in 2021.
• 11:05
  I think we're seeing a version of that right now too, which is that I think people wanna get
• 11:10
  into hot day ideals.
• 11:11
  And if a company gives you a revenue number that looks passably accurate, if you're the
• 11:16
  person who really digs in, you might not get the deal.
• 11:18
  And in the long run, that might be the optimal thing, is to actually not get the deal.
• 11:22
  In the short run, that is not what people want.
• 11:25
  One thing I was thinking about the other day is we mentioned in this piece, investors are
• 11:29
  implicated here, founders are implicated here.
• 11:31
  You know, how much do you think imprecise revenue language actually ends up hurting
• 11:36
  founders and our teams internally?
• 11:37
  Like, how they track their own progress and goals for their internal team?
• 11:41
  If you're, like, employee number three, four, five Yeah.
• 11:43
  Like, you're the first sales hire, like, what happens there?
• 11:45
  Or do you think oftentimes you're too far removed that those things don't actually
• 11:48
  trickle down in front of you?
• 11:49
  I think what ends up happening I've seen this happen.
• 11:51
  This is one this is one archetype.
• 11:53
  The company hears about somebody else's revenue number and says, how on earth are they doing
• 11:57
  that?
• 11:57
  We're so behind.
• 11:59
  And then you have a board meeting and someone goes, company x is doing a 100,000,000 in ARR,
• 12:03
  and we're at 10.
• 12:04
  We're so behind.
• 12:05
  How is this possible?
• 12:06
  And there's all these, like, convincing and, like, strategy, meetings, and how are we gonna
• 12:10
  grow faster?
• 12:11
  And it turns out that company is only doing 15,000,000 in a in true ARR.
• 12:14
  And we're like, okay.
• 12:15
  So we weren't even behind.
• 12:16
  Freaked ourselves out, tried all these crazy things to grow faster because we thought we
• 12:19
  were behind, but we weren't.
• 12:20
  But part of the problem is, you know, if you are the perceived winner and capital flows to
• 12:25
  you based on, we'll say, misunderstood metrics, there's no take backs.
• 12:29
  There are some.
• 12:30
  There are there are relatively few take backs.
• 12:31
  And then the question is, like, what do you do if your competitor used misunderstood metrics
• 12:36
  to raise a bunch of money and put distance between you and them?
• 12:39
  I think that's hard.
• 12:39
  So I think almost everybody has an incentive to keep going.
• 12:42
  I had a conversation with one of our teams a couple weeks ago.
• 12:45
  We, like, sat for an hour to look across their financials, and one thing that I saw that came
• 12:48
  from an investor update is there was so much volatility month over month in how they were
• 12:52
  reporting their revenue.
• 12:53
  And I was like, oh, this is a problem.
• 12:54
  We walk into the meeting, and I bring it up.
• 12:56
  There was a steep decline here from, like, month over month.
• 12:58
  They're like, oh, we bill quarterly, so it doesn't come out even.
• 13:01
  So, like, this 40 k drop from February to March actually wasn't somebody churning.
• 13:06
  It's just the way you build.
• 13:07
  They're like, yeah.
• 13:07
  I'm like, oh, these are things that we need to just tweak in language to make sure everyone
• 13:11
  understands this on the same page.
• 13:12
  Because I think one thing that founders don't realize is if a net new investor saw that, you
• 13:16
  pass immediately.
• 13:17
  Or you'd be like, there's something weird going on here.
• 13:19
  I'm not interested.
• 13:19
  Even though it's not the fault of the founder.
• 13:21
  And if the founder massaged the numbers to meet them halfway, I think an investor would also be
• 13:26
  well, you're kind of tweaking the numbers here.
• 13:28
  That's actually not how you built.
• 13:30
  We had an hour long conversation, basically, of when we come up for the next raise, how do we
• 13:34
  wanna talk about this?
• 13:35
  Because you guys have amazing traction, and I don't want it to get lost in somebody thinking
• 13:39
  that you're portraying it where it's not actually as great when, like, it truly is.
• 13:43
  And I think that's actually the the conversation that most founders live in from
• 13:47
  time to time.
• 13:47
  Well, I think the other thing, most founders like, I wanna get a meeting so I can get into
• 13:51
  the room and tell somebody.
• 13:52
  And I have a VZs are just like, oh, the growth in that blurb is too slow.
• 13:55
  I don't wanna meet the
• 13:56
  Right.
• 13:56
  Of course.
• 13:57
  Yeah.
• 13:57
  Company's gonna be like, well, if revenue growth or revenue run rate is what gets me in
• 14:02
  the door, I want my company to be presented in the best light possible.
• 14:06
  So people will take the meeting, and then hopefully I can dazzle them with my, like,
• 14:09
  winning personality and, like, amazing story.
• 14:11
  But if I can't get in the door because I'm growing at 75% year over year Yeah.
• 14:17
  And everyone else is advertising 500% growth or a thousand percent growth, people aren't gonna
• 14:21
  take the meeting.
• 14:22
  But wouldn't you argue that if you get into a tough situation coming into the meeting, and
• 14:26
  then the founder's like, oh, I found something weird.
• 14:28
  You're in the hole already.
• 14:29
  There's no back and forth of like, oh, let's hop on the phone and walk through like how we
• 14:32
  have a discrepancy of a million dollars in error hard difference here.
• 14:35
  Like, you're in the hole already, which is a bummer.
• 14:36
  I think most people are like, just get me in the room.
• 14:39
  I'll do whatever I have to do to get in the room.
• 14:41
  Which is also things that investors say about trying to get into hot deals, I will say also.
• 14:45
  Our last piece this week is a TechCrunch report from the past weekend.
• 14:49
  China has blocked Meta's proposed $2,000,000,000 acquisition of Manus AI after a
• 14:54
  month's long government review.
• 14:55
  Manus is a Chinese developed, now Singaporean based autonomous AI agent that went viral
• 15:00
  earlier this year for its ability to perform complex multistep tasks without human input at
• 15:05
  each step.
• 15:05
  And the deal would have been one of the most significant acquisitions of a Chinese AI
• 15:09
  company by a US tech giant to date.
• 15:11
  Regulators blocked it on national security and data sovereignty grounds.
• 15:15
  What makes this one worth discussing is the broader signal it sends as Achentic AI moves
• 15:20
  from interesting demo to genuinely strategic infrastructure, the geopolitical walls around
• 15:25
  it are going up pretty fast.
• 15:27
  And the window for cross border AI m and a may be closing in real time.
• 15:32
  This does feel like a preview of a much bigger dynamic.
• 15:34
  It's just being treated more as strategic national infrastructure, like I mentioned
• 15:38
  before, and not just as this standalone product that people are excited to make money off of.
• 15:42
  How do you think about what that means for the investment landscape around AI companies that
• 15:47
  have meaningful exposure to one side of this geopolitical divide.
• 15:52
  I was very surprised when this deal happened that everybody assumed it would go through.
• 15:57
  And I think, ironically, most people are like, well, The US is gonna allow this to happen.
• 16:00
  I was like, but don't the Chinese regulators out of Of course.
• 16:03
  The US will allow that.
• 16:04
  Of course.
• 16:05
  And, you know, I think when all the CFIUS and reverse CFIUS stuff started to happen, I think
• 16:08
  a lot of people looked at, you know, Singapore washing or how how can we still get access to
• 16:12
  the amazing innovation and technological development that's happening in China without
• 16:17
  having to deal with the regulatory environment.
• 16:19
  I'm like, guys, those regulatory environments exist because of great power competition
• 16:22
  between United States and China.
• 16:24
  For a lot, I feel really bad for the people who are involved in this transaction.
• 16:27
  I think they all thought it was a done deal.
• 16:29
  Including the employees.
• 16:30
  Sitting at Meta.
• 16:32
  And to me, we just live in a world where technology is part of national security, and
• 16:39
  companies view technology as part of their national interest.
• 16:43
  And I think it just increases the riskiness of any cross border deal.
• 16:48
  And not just any cross border deal.
• 16:50
  I think about Adobe FIC, like other deals where the company is a global player, combining with
• 16:55
  another meaningful or significant company is gonna invite regulatory scrutiny even in where
• 17:01
  you might say these people have no standing to even make an argument.
• 17:05
  And technology is global, and it touches some of the most important things in the world.
• 17:09
  And so I think this is the world we live in now, which is that whether it's semiconductors
• 17:13
  or AI or defense or energy tech, I think a lot of these things are gonna face multi party
• 17:19
  regulatory scrutiny when it comes to M and A or general business strategy.
• 17:23
  I was gonna ask you, did this block deal change how you think about the exit landscape for AI
• 17:29
  companies that have any international complexity on their cap table?
• 17:32
  Or I feel like you've thought this way probably consistent.
• 17:35
  I think every government thinks that they have a national interest in AI in some element of
• 17:40
  it.
• 17:40
  And the notion that companies were gonna be able to either skirt regulation by doing clever
• 17:45
  domicile stuff or that, like, other regulators weren't gonna weigh in.
• 17:49
  I thought, well, that seems a little self serving.
• 17:51
  That's not the world we're in.
• 17:52
  Maybe if we get back at some point to a place where there's, like, less great power
• 17:55
  competition, a greater sense of collaboration at the national scale, some of the stuff would
• 18:00
  maybe be less heady, but not anytime soon.
• 18:02
  We are all out of time today.
• 18:03
  Thank you, Charles, for jumping in.
• 18:05
  Thank you to our listeners for joining as well.
• 18:07
  Please like and subscribe, and we will talk with you all next week.
• 18:10
  Thank you.
• 18:11
  Bye.