Latest / The Learning Corner by Precursor / Episode #50: What Does ARR Mean?, High Agency in Silicon Valley, Debt Fueling The Next AI Wave

Transcript

Episode #50: What Does ARR Mean?, High Agency in Silicon Valley, Debt Fueling The Next AI Wave · The Learning Corner by Precursor

• 0:00
  When I talk to a of people about this nine nine six, it's not because they're trying to keep up
• 0:03
  with customer demand.
• 0:04
  It's not like, oh my god.
• 0:05
  We have to be in here on Saturday because we have so many support tickets to deal with.
• 0:08
  It's, I think, a fear that everybody else is doing it, and am I a loser if I'm not working
• 0:13
  on Saturday?
• 0:13
  It's the week of September 29.
• 0:16
  Welcome to the Learning Corner, a Precursor Ventures podcast.
• 0:24
  Welcome to the Learning Corner by Precursor.
• 0:26
  I'm Mia Farnham, and I'm joined by RGP and managing partner, Charles Hudson, where we will
• 0:31
  be talking through the articles in Goodreads that caught our team's attention over the past
• 0:35
  week.
• 0:36
  Our first topic is a piece from Fortune titled Founders Are Using Creative Accounting to Boost
• 0:42
  Lofty ARR, the hottest startup metric in Silicon Valley.
• 0:46
  Author Ali Garpingold dives into the historically deeply understood concept of ARR
• 0:52
  and what has become murkier about the metric in the recent years.
• 0:55
  In the early days of SaaS, ARR was a clear cut indicator.
• 0:59
  It was a measure of stable reoccurring revenue readily accepted by VCs as a proxy for the
• 1:05
  startup's long term viability.
• 1:07
  Standardized contracts and annual pricing made the metric hard to manipulate.
• 1:12
  Most deals reflected real, lasting Damn.
• 1:15
  Customer commitments.
• 1:17
  But as the AI gold rush has accelerated, the industry's trust in ARR has eroded.
• 1:24
  Founders face immense pressure to show revenue early, leading to creative interpretations,
• 1:31
  pilots, short term trials, unactivated contracts, and so called booked ARR Yeah.
• 1:37
  Are counted as if they were guaranteed long term income even when customers can really just
• 1:43
  walk away at any time.
• 1:45
  So the metric is now less about actual reoccurring revenue and more about Vibe
• 1:50
  revenue, as quoted in the article, or experimental revenue, as we've talked about
• 1:55
  here on the podcast, really lacking the rile reliability it had in SAS's prime.
• 2:00
  So there is an argument in the piece that bubbling valuations are tied to the growing
• 2:06
  murkiness of ARR.
• 2:07
  Essentially, the higher a company's ARR, the bigger its valuation can be.
• 2:11
  Nothing really revolutionary.
• 2:12
  It's kind of like how it's always worked.
• 2:14
  But really based on investor hopes for future liquidity, which back up them wanting to give
• 2:18
  out those higher valuations.
• 2:20
  My question is, even if the current AI boom hadn't lifted all in venture recently Yeah.
• 2:26
  Do you think ARR would still have become as murky as it is today?
• 2:31
  No.
• 2:32
  Yeah.
• 2:33
  I don't think so.
• 2:34
  I think in SaaS land, we'd sort of decided this is contracted revenue, this is activated
• 2:39
  revenue.
• 2:40
  It felt like, you know, from predictable revenue way back in the beginning of SaaS.
• 2:44
  Today, we had pretty good conventions about what does revenue mean?
• 2:47
  What does bookings mean?
• 2:49
  What's recurring?
• 2:50
  What's nonrecurring?
• 2:50
  What's transactional?
• 2:51
  I feel like we'd sorted all of this out.
• 2:54
  And this is one of those cases where I am both sympathetic and terrified for the plight that
• 2:59
  founders face.
• 3:00
  And for me, the first time I really got a hint that something was different was 11x.
• 3:05
  Yeah.
• 3:05
  When the article came out, 11x was like, oh, we have all these annual contracts.
• 3:08
  And I was like, wait a minute.
• 3:09
  I thought we had settled this.
• 3:11
  That revenue isn't revenue when the customer has an easy opt out.
• 3:15
  If you're telling me that you signed a year long contract, but the customer has a ninety
• 3:19
  day opt out, I'm like, well, that ninety days is kind of a pilot sport where if it's I
• 3:24
  wouldn't book that as revenue until the customer has left the period where they can opt
• 3:28
  out of the deal.
• 3:29
  Yeah.
• 3:29
  And I think that worked for a while.
• 3:32
  And what I've seen is people getting progressively more aggressive.
• 3:35
  Everything from, well, we gave a bunch of our customers credits for future usage, and we're
• 3:40
  gonna count that as revenue.
• 3:41
  I'm like, well, that's not revenue.
• 3:42
  That's, like, credits you gave them for future usage.
• 3:45
  If anything, that's more like a liability than revenue.
• 3:48
  And I understand why this is happening.
• 3:50
  It's because you read these blog posts about companies that have gone from zero to whatever
• 3:56
  in these impossibly short periods of time.
• 3:58
  And I think founders say, well, that's the standard that I need to show in order to get
• 4:03
  attention from investors.
• 4:05
  And even there was a, like, a sub conversation that, you know, the old triple triple double
• 4:10
  double, which used to be really good performance.
• 4:12
  I was like, that doesn't matter anymore.
• 4:14
  That's not good enough.
• 4:15
  I think all of the messages from VCs are like, if you don't look like Lovable, if you don't
• 4:19
  look like Cursor or Replit, one of these companies that goes from nothing to something
• 4:23
  in a very compressed period of time, you don't matter.
• 4:27
  Yeah.
• 4:27
  And so I think the rational, albeit bad, response from most founders is why need to make
• 4:32
  my business look like that.
• 4:34
  And I've met people who've told me, like, all kinds of different things.
• 4:38
  So I never thought I'd to spend this much time yet.
• 4:40
  And talking to people about what does that revenue number you've told me, What does it
• 4:44
  actually mean?
• 4:45
  First of all, is it even revenue?
• 4:47
  Is it money in from a customer or is it something else?
• 4:50
  And then what are the terms?
• 4:52
  And I hope that this fever breaks soon because I think it doesn't help anybody to spend that
• 4:57
  much of a first introductory pitch meeting trying to sort through where is this company
• 5:02
  actually.
• 5:03
  The flip side is the companies that can make those claims and sell them raise really
• 5:07
  quickly.
• 5:08
  So there isn't always a lot of time as an investor to diligence.
• 5:10
  And if you're like, hey.
• 5:11
  I wanna see all the contracts, and I wanna, like, understand, somebody else might just say,
• 5:15
  oh, I'll take your word for it.
• 5:16
  So in some ways, you might be back to some of the bad habits of 2021.
• 5:19
  It doesn't help that investors aren't spending time digging into it.
• 5:23
  And I don't again, I don't blame investors to it all part of, like, the larger dynamic of
• 5:28
  everyone wanting to seek liquidity and typically ARR being a way to do risk if this
• 5:33
  company is going to provide liquidity in a shorter period of time.
• 5:36
  So if you told me you went from 0 to 7,000,000 ARR in a week, I try to look through and find
• 5:41
  what that is.
• 5:42
  But if you're an investor and you're like, this is a time of competitive deals.
• 5:45
  If I spend too much time digging into this and this actually could be the next thing, why
• 5:49
  don't I just write it up at some crazy valuation to show support and then move on and
• 5:55
  hope that it all works out?
• 5:56
  We were just talking about this with one of our interns, Missy, going into kind of the loop of
• 6:01
  if you're a user of these different AI tools, it is fascinating.
• 6:06
  I've never been so disloyal as a subscriber than now.
• 6:11
  So I'm not surprised that these new tools have such an uptick in revenue month over month.
• 6:18
  But I've churned an outrageous amount of times
• 6:21
  Of course.
• 6:21
  For so many of them.
• 6:22
  So from an investor standpoint, do you feel like they're asking about churn and retention
• 6:27
  as much?
• 6:27
  Or do you think it's just no?
• 6:28
  I think what you said is right, which is the competitive deal dynamics make asking a lot of
• 6:33
  probing questions difficult because it is a little bit of game theory.
• 6:37
  If you're like, well, if all of the investors ask the same probing questions, we'd all be on
• 6:41
  equal footing.
• 6:42
  But there might be someone else who's like, I really wanna win this deal.
• 6:45
  Yeah.
• 6:45
  If the path to winning this deal is asking fewer questions, I'm gonna ask fewer questions.
• 6:49
  And while you're off in the data room doing due diligence, I'm gonna sign a term sheet.
• 6:53
  Right.
• 6:53
  We're done.
• 6:54
  We talked about this on the pod before about how gross margins are another metric that could
• 6:58
  start to drift away back in, you know, heat of the SaaS era of 75, 90% margins.
• 7:03
  That's just what was considered good.
• 7:05
  And now with AI, it's drifting sub 30%.
• 7:09
  And I think people are thinking it's it'll correct itself.
• 7:11
  At what point do you think investors will establish a clearer set of metrics to guide
• 7:15
  this?
• 7:15
  Or do you think we don't care?
• 7:17
  Not we.
• 7:17
  The royal we.
• 7:19
  I think when I ask people the the margins question, and when I ask, they laugh at me the
• 7:24
  way I just laughed out loud.
• 7:25
  Yeah.
• 7:25
  I think there's a belief that compute is going to get cheap enough that a lot of the cost in
• 7:31
  scaling these things so I would say a lot of the companies that are growing quickly don't
• 7:34
  appear to be very efficient in compute usage because you can't be when your usage is going
• 7:38
  like this.
• 7:39
  And compute is still a significant input to cost.
• 7:41
  But to me, the question is like, well, yes, in a perfect scenario, prices that you can charge
• 7:45
  stay high, compute costs come down, and that drops the margin.
• 7:49
  But I'm like, what if, because of competition, compute comes down, but the price you can
• 7:53
  charge also comes down?
• 7:54
  Yep.
• 7:54
  And then all you're doing is you're not getting any of the you're not getting much of the
• 7:57
  benefit of the drop in compute because it's being eaten away by the competitive dynamic.
• 8:01
  And there's a lot of products I see where I'm just like, oh, this is like that other thing I
• 8:04
  used, but it's cheaper.
• 8:06
  And it seems to be just as good, so why won't I churn and use the cheaper version?
• 8:10
  And I think that person's like, oh, once I get market share, I'll have some pricing power, and
• 8:15
  I will either raise prices or drop cost.
• 8:17
  But I'm like, but what if as soon as you start to get market share, somebody else comes and
• 8:21
  does the same thing to you?
• 8:22
  And if you're an investor, again, to our question about portfolio conflict, you've
• 8:26
  already dumped a lot of money into this and then you see that happen.
• 8:28
  Do you abandon that investment or just start being like, let me just slide in an investment
• 8:32
  over here because it feels like that's gonna be the next thing for the next handful of months?
• 8:36
  Well, maybe if you have a growth fund, you could say, yeah, the growth fund's gonna do
• 8:38
  this.
• 8:39
  Light a 100,000,000 on fire.
• 8:40
  Any last thoughts on this one?
• 8:42
  So I was talking to someone this morning.
• 8:46
  I think in 2021, everybody knew that they were overpaying, but they understood what they it
• 8:52
  wasn't like, oh, I don't know what this revenue means.
• 8:54
  Like, I'm gonna put a 200 x multiple on this 1,000,000 of ARR To win it.
• 8:58
  To win it.
• 8:59
  I know I'm paying an aggressive multiple, but I think I understand the underlying thing I'm
• 9:04
  basing this multiple I'm very curious to see what happens maybe two years from now.
• 9:10
  And when some journalist says, let's go back and look at all those companies, and I think
• 9:15
  there will be some actual fraud.
• 9:17
  Yep.
• 9:17
  I think there'll be some actual fraud that we find.
• 9:19
  That's what Missy also asked.
• 9:21
  She's like, when do you think the next big thing that's gonna be a documentary in six
• 9:25
  years is gonna happen?
• 9:26
  And I said, a little bit longer than you,
• 9:27
  but one
• 9:28
  to two years is jarring.
• 9:29
  Oh, yeah.
• 9:30
  Our next topic is a recent piece by Jasmine Sun talking through the new key terms flooding the
• 9:36
  Bay Area in this post recession renaissance fueled by AI's rapid rise and permeating every
• 9:42
  corner of culture.
• 9:44
  This piece spotlights Silicon Valley's obsession with agency.
• 9:47
  The ability to act independently and shape one's destiny in a world where AI can answer
• 9:53
  questions but not develop goals.
• 9:55
  Being high agency means resourcefulness and initiative, while the NPC, it's like non player
• 10:02
  character, label mocks those stuck following routines, oblivious to broader tech shifts.
• 10:09
  These memes reflect deeper worries about becoming part of the, quote, permanent
• 10:14
  underclass as AI divides economic winners from losers, pushing returns to capital, and
• 10:19
  threatening traditional middle class security.
• 10:22
  It's, like, actually a big existential word.
• 10:24
  Yes.
• 10:25
  Other cultural terms dissected include nine nine six, so grueling work culture, taste,
• 10:32
  knowing what's worst working on, and decel, doomer, labels for those really skeptical or
• 10:39
  resistant to unchecked technological acceleration.
• 10:43
  Ultimately, she highlights how despite the wealth and bravado, status anxieties run deep
• 10:49
  and everyone, founders, engineers, VCs, are looking for meaning beyond the next big AI
• 10:54
  leap.
• 10:55
  It feels to me that there's a common thread feeling right now that everything in tech feels
• 11:00
  so serious.
• 11:02
  Like, one wrong move could drastically change an entrepreneurial journey.
• 11:06
  One diligence misstep could completely upend your venture career in this AI market.
• 11:12
  What were your reactions when you read the piece?
• 11:14
  I feel like every now and then, somebody writes a piece that totally captures the moment.
• 11:20
  And I was like, oh, this is what it feels like to live in San Francisco Yes.
• 11:23
  Right now.
• 11:24
  Shout out to one of our favorites, Hunter Walk, for flagging this Like, one for there's so many
• 11:29
  things.
• 11:29
  One, I'd say I've never met more of my friends who are incredibly intelligent, talented people
• 11:35
  who are working really hard out of fear.
• 11:38
  Yeah.
• 11:38
  Out of like fear that their competitors are working hard.
• 11:42
  You're talking investor side or founder Founder side.
• 11:44
  Yeah.
• 11:45
  Who are working hard out of fear.
• 11:46
  And I think like fear and anxiety is not a good reason to work hard in startups.
• 11:51
  Personally, that's like my belief.
• 11:52
  I could go longer than that.
• 11:53
  But I think when I talk to a lot of people about this nine nine six, it's not because
• 11:56
  they're trying to keep up with customer demand.
• 11:58
  It's not like, oh my god.
• 11:59
  We have to be in here on Saturday because we have so many support tickets to deal with.
• 12:02
  We have so many it's, I think, a fear that everybody else is doing it, and am I a loser if
• 12:07
  I'm not working on Saturday?
• 12:08
  Yes.
• 12:09
  And what will happen if OpenAI drops something on Friday, and we don't have a new version of
• 12:15
  our product out by Monday?
• 12:16
  Like Sora.
• 12:17
  Like Sora.
• 12:18
  Yesterday.
• 12:18
  Yeah.
• 12:18
  Like, maybe nothing or maybe everything.
• 12:20
  Yeah.
• 12:21
  And like nobody knows.
• 12:22
  What I found is when you put high achieving people in a highly anxious moment, working
• 12:29
  harder is a natural control response that people it's like, I can control the number of
• 12:34
  hours I pay.
• 12:35
  One of my family members read the nine ninety six article in New York Times.
• 12:38
  They're like, are people in San Francisco having any fun?
• 12:40
  And I think this, like, weird mix of stoicism and work culture, it's been fascinating to
• 12:47
  watch this, like, life is also about enjoyment and fun.
• 12:51
  Yeah.
• 12:52
  Yes, you should work a lot.
• 12:53
  And yes, work is serious.
• 12:54
  I hate to say it.
• 12:55
  Some of these AI companies are deeply unserious and unimportant.
• 12:58
  Yes.
• 12:59
  And I don't mean that as an insult.
• 13:00
  Mean that as a statement of fact.
• 13:01
  And I wonder how some of these people are gonna feel two or three years from now when they look
• 13:06
  back and say, spent 40 in 2025 working on some random AI thing that I can no longer describe,
• 13:14
  and I didn't enjoy the time I spent with my coworker.
• 13:16
  Yeah.
• 13:17
  And to be clear, I've had jobs where I've worked a lot of hours, and not all those hours
• 13:21
  are productive.
• 13:22
  What I will say is many of my best startup friendships came from periods of intense work.
• 13:27
  But we were doing that because we had something that was working, and we were trying to keep up
• 13:31
  with demand.
• 13:33
  And I always worry when people are applying a lot of hours, pre product market fit, or in the
• 13:37
  absence of demand, because that's a good recipe for burnout.
• 13:40
  It also doesn't help that you have I mean, all over social media, I think I've seen it from
• 13:44
  four or five different massively successful Fortune 500 CEOs that say, like, I've never met
• 13:49
  a successful person who wasn't grinding every day in their twenties.
• 13:53
  And then pair that with an influx of young founders who don't have other Yeah.
• 13:58
  Work experiences to point to of knowing what's a good use of time.
• 14:03
  There are obviously wildly successful young founders who all they did was work and and it
• 14:06
  really did pay off.
• 14:07
  I feel like you're one of the biggest San Francisco stans.
• 14:10
  Do you feel like this is one aspect of living in SF that's actually a negative if you're a
• 14:15
  founder?
• 14:16
  I've thought about this a lot.
• 14:17
  I think it's just a fact.
• 14:19
  And I think, like, your competitors are going to do this.
• 14:21
  But if they have no good insights, those incremental marginal hours when I graduated
• 14:25
  from business school in o five, I had a handful of friends who worked at Facebook right out of
• 14:30
  undergrad.
• 14:31
  They were at Facebook all the time.
• 14:33
  Yeah.
• 14:34
  I would meet them for dinner in Palo Alto, and they would go back to Facebook.
• 14:37
  But that's because Facebook was vertical, and that's because Facebook was solving real
• 14:41
  problems.
• 14:42
  The marginal value of an hour in a company that's scaling like that is incredibly high.
• 14:46
  Mhmm.
• 14:47
  And the marginal value of getting everybody in the company to put in the marginal hour, I
• 14:51
  don't wanna get the coordination costs, incredibly high.
• 14:55
  And I understood why they were spending so much of their lives at Facebook.
• 14:59
  Uber was the same way.
• 15:01
  Yeah.
• 15:01
  Uber was scaling.
• 15:02
  My friends at Uber worked like dogs.
• 15:04
  They were always in the office.
• 15:05
  But it's because they were winning.
• 15:06
  They were scaling and growing, and there was so much productive work to do.
• 15:11
  And I don't wanna get too, like, philosophical.
• 15:14
  I feel like a lot of high achieving smart people, when they get anxious, they throw more
• 15:18
  hours at problems.
• 15:19
  Instead of stepping back and I tell our founders this all the time.
• 15:22
  If you're working seventy hours a week, it's a lot.
• 15:25
  That's the same objective.
• 15:25
  It's a lot of hours.
• 15:27
  And the problem isn't yielding, what are the odds that the seventy first hour that you work
• 15:32
  is actually the magic unlock?
• 15:33
  Usually, at that point, you're missing something.
• 15:36
  Product doesn't work.
• 15:37
  Wrong customer profile.
• 15:39
  Someone else is winning, and you haven't figured out why.
• 15:41
  It is rare that the raw application of additional hours Yeah.
• 15:45
  But what the raw application of additional hours can do with an in office culture is build
• 15:50
  really strong friendships and connection and trust, and that is super valuable.
• 15:54
  But I'd rather see people make those investments when they have something that has
• 15:58
  momentum and it's working.
• 15:59
  We had our pre seed boot camp in New York City last week, and this was actually such a thread
• 16:03
  that all of our founders talked about was it's difficult to maybe have those conversations
• 16:07
  with yourself of maybe I'm not working on the right thing.
• 16:10
  What all you see in the media is that eighty fifth hour was the unlock.
• 16:15
  Media just doesn't wanna focus on the more relatable story.
• 16:19
  That's just not interesting, and so I don't blame them.
• 16:21
  One more thing I'd My friend Lan from Basis said, had a really good LinkedIn post, which
• 16:25
  was everybody should have one or two times in your life of intense work before you have other
• 16:31
  obligations Yeah.
• 16:31
  For children, spouses.
• 16:32
  For some people, that's consulting.
• 16:34
  For some people, that's investment banking.
• 16:36
  For some people, that's working in the zero to one startup.
• 16:38
  I do think it's an important and useful if you wanna go far in white collar work, it's good to
• 16:44
  have one or two of those experiences and just know what it feels like to kinda not have a
• 16:48
  life Yeah.
• 16:49
  For a year.
• 16:50
  And then you can wrestle with, like, is this something I want more of?
• 16:52
  Does this trade off for success make sense?
• 16:55
  So
• 16:55
  It'll be interesting to see that same thought for a generation of people who all they've
• 17:01
  known is being a founder.
• 17:02
  I know.
• 17:03
  That'll be interesting.
• 17:04
  I think in a year, I'm waiting for the blog post in a year.
• 17:07
  We did nine ninety six for a year and it wasn't worth it.
• 17:10
  Will they publish that?
• 17:11
  They'll write The people who do the nine ninety six, will they know to publish Some
• 17:14
  people will write it because they'll want the clout.
• 17:16
  Last question here because I think it's a good one.
• 17:18
  Between the different terms named in the piece nine ninety six, taste, decel doomer, agency,
• 17:24
  and then permanent underclass Yeah.
• 17:26
  Which one do you feel the strongest about, either in agreement or disagreement?
• 17:30
  I think for a lot of people in San Francisco, the fear of falling back and being in the
• 17:35
  permanent underclass is the dominant motivating theme.
• 17:38
  Because the city's full of really high achieving people with amazing academic and
• 17:44
  technical pedigree who are used to winning and being on top.
• 17:47
  Yeah.
• 17:48
  And if they're not on top, they're on top adjacent.
• 17:50
  Like, they're very close.
• 17:51
  And the idea of, like, falling back into some permanent underclass kind of would invalidate
• 17:57
  very many of the things that they've spent their life pursuing, whether that's higher
• 17:59
  education, good grades, prestigious jobs.
• 18:02
  And I do think some of this is like a fear of falling behind.
• 18:05
  I feel it here.
• 18:06
  And permeate underclass for people who are in SF, does it truly mean that they're not
• 18:10
  included in like the big tech companies or that just they truly are working in in a completely
• 18:14
  different sector?
• 18:15
  I think it's irrelevant.
• 18:17
  I think it's like working on things.
• 18:18
  Not a specific job.
• 18:19
  It's not a specific job.
• 18:20
  It's just like you're just part of this vast, like I talked to someone this week who's been
• 18:25
  avenger to habits for a long time, and I asked him, like, You're very rich.
• 18:28
  Why don't you retire?
• 18:29
  He goes, I don't enjoy my job, but I I fear irrelevance more.
• 18:34
  And I was like, well, you're still a person without your job, but I know what he meant.
• 18:39
  Yeah.
• 18:39
  Like, a lot of the construct of that person's life was very work centric.
• 18:43
  We have a whole podcast about how you deprogram that in but your it's it's real.
• 18:48
  Well, final topic, it actually came up in good timing.
• 18:52
  So it's a recent Wall Street Journal article shedding light on the massive amounts of debt
• 18:56
  taken on by large companies to build AI infrastructure.
• 18:59
  We talked about this last week, and then you mentioned that maybe this should be an entirely
• 19:03
  new topic for a whole podcast.
• 19:05
  So here we are.
• 19:07
  Debt is quickly becoming the engine behind much of the AI industry's rapid growth.
• 19:11
  So while tech giants like Google and Microsoft built their early AI ambitions on profits from
• 19:16
  existing businesses, some players like Oracle and CoreWeave and their peers, are fueling
• 19:22
  expansion by taking on enormous loans
• 19:24
  Mhmm.
• 19:24
  And betting big on this future demand.
• 19:27
  So more specifically, Oracle recently signed a massive $300,000,000,000 five year contract
• 19:33
  with OpenAI, securing huge amounts of data center power.
• 19:37
  To deliver on this, Oracle has to borrow around $25,000,000,000 per year, pushing their debt to
• 19:43
  equity ratio to about 450%.
• 19:46
  It's far higher than the financial structure of companies like Alphabet and Microsoft.
• 19:50
  Similarly, Quarweave is making bold infrastructure bets by leveraging contracts and
• 19:55
  anticipated revenue.
• 19:56
  Investor enthusiasm remains high.
• 19:59
  Analysts mentioned that stock prices were gonna triple in recent months.
• 20:03
  However, there are some groups who are skeptical about whether these companies will
• 20:07
  see the expected growth.
• 20:09
  Yeah.
• 20:09
  Ultimately, the article warns that the debt fueled race to build AI infrastructure could
• 20:13
  turn into a plastic tech bubble If consumer adoption slows or contracts are postponed and
• 20:18
  canceled, if growth stalls, many of these ambitious plans and the companies behind them
• 20:23
  could be forced to reconsider their strategies in the face of mounting financial pressure.
• 20:28
  Yep.
• 20:28
  So wanted to revisit what you said last week and see where your head
• 20:31
  is I talked to a friend of mine in private credit, and he's like, well, data centers equal
• 20:36
  private credit.
• 20:36
  Like, what do mean?
• 20:37
  He's like, all of the deals that I'm doing Yes.
• 20:39
  Are data centers.
• 20:40
  I'm always hesitant to draw 99 analogies, but in '99, we put a lot of fiber in the ground
• 20:45
  with the expectation that we would need it soon.
• 20:48
  And good news is we ripped up all the streets.
• 20:50
  We put down a lot of fiber.
• 20:51
  We eventually needed it.
• 20:53
  We just didn't need it right away.
• 20:54
  And right now, AI is going like this, vertical.
• 20:56
  People are trying to forecast how much capacity are we gonna need in the future.
• 21:00
  And right now, everyone's like, well, a lot more.
• 21:03
  And the thing that always scares me about debt is it's really hard to blow up an economy with
• 21:07
  equity.
• 21:07
  You wipe it out.
• 21:08
  Equity is risk capital.
• 21:10
  Equity markets are tiny compared to debt markets, and debt expects to get repaid.
• 21:15
  And what if we're wrong about the timing of AI demand?
• 21:19
  I right now, I'd say this is a low probability event.
• 21:22
  But my learning in life is you should actually pay attention to low probability, high
• 21:27
  consequence events.
• 21:28
  And I'm like, if we are wrong on the shape of AI demand and we overbuild data center which
• 21:34
  right now most seem like that's impossible.
• 21:36
  Like, demand is is growing so quickly.
• 21:38
  How can we overbuild it?
• 21:39
  We've done it before.
• 21:40
  We already built houses in Las Vegas.
• 21:42
  We eventually needed them, but there's no other town.
• 21:44
  We had too many of those.
• 21:45
  We put too much fiber in the ground for that moment in time.
• 21:47
  We needed it eventually.
• 21:49
  I'm pretty convinced that all this data center capacity will get used eventually.
• 21:52
  But if we overbuild and these data centers aren't utilized and people can't make the
• 21:57
  returns and there's default, that to me feels like the thing that not enough people are
• 22:02
  thinking about because it seems so unlikely to happen.
• 22:04
  And I actually think it's important to break down the dynamics.
• 22:08
  Like, think people are like, oh, boohoo, insert name of Massive's credit fund goes out of
• 22:12
  business.
• 22:13
  But their LPs are pension funds.
• 22:14
  That's right.
• 22:15
  They are the retirement plans for everyday teachers.
• 22:18
  Credit is notoriously a less risky vehicle to move money if they start moving into a space
• 22:23
  that we know is still quite young.
• 22:25
  The same model that you've been reusing for real estate does not apply here.
• 22:29
  There are there are massive consequences beyond super wealthy people not becoming wealthier.
• 22:34
  I mean, you know, when you get defaults like that, lending standards go up, cost of money
• 22:38
  goes up, you can choke credit out of an economy really quickly.
• 22:41
  Yes.
• 22:41
  And look, I think part of what's happening is if you were doing data center deals five years
• 22:45
  ago, you are, like,
• 22:46
  stunned.
• 22:47
  Right.
• 22:47
  Those data centers are built to realize right.
• 22:49
  Yeah.
• 22:50
  Everyone's like, well, it was great five years ago.
• 22:51
  What if we 10 x the capital?
• 22:53
  What could go wrong?
• 22:54
  I'm like, than everything, nothing.
• 22:55
  This is the thing I'm paying attention to because this is the one thing I'm like, not
• 22:59
  enough people are talking about this in the circles that I'm in.
• 23:02
  And the consequences of getting this wrong could be, like, catastrophic.
• 23:05
  Yeah.
• 23:05
  And there's a disconnect between the people who run these large credit vehicles and people who
• 23:09
  sit at the earliest stages of venture Okay.
• 23:11
  Knowing what's to come and what things to hedge.
• 23:14
  Well, are all out of time today.
• 23:15
  Thank you, Charles, for filming in person Always in San Better.
• 23:19
  Thank you to our listeners for shopping in.
• 23:21
  Please like and subscribe, and we will talk with you all next week.
• 23:23
  Thank you.
• 23:24
  Bye.