Latest / The Learning Corner by Precursor / Episode #49: Making Money on AI, Consumer Deserves Attention from Investors, Nvidia Invests in OpenAI

Transcript

Episode #49: Making Money on AI, Consumer Deserves Attention from Investors, Nvidia Invests in OpenAI · The Learning Corner by Precursor

• 0:00
  I think we're in the frenzy stage because I feel like in the eruption phase, there's nobody
• 0:05
  who calls it a bubble.
• 0:06
  There's no one who says things have gotten extended because you're so focused on the
• 0:11
  optimism and the possibilities.
• 0:13
  The reason I think we're in the frenzy stage is privately, I've talked to investors who are
• 0:16
  like, look, I know that this is frothy and insane, but I'm gonna ride the wave.
• 0:21
  It's the week of September 22.
• 0:23
  Welcome to the Learning Corner, a Precursor Ventures podcast.
• 0:31
  Welcome to the Learning Corner by Precursor.
• 0:33
  I'm Mia Farnam, and I'm joined by our GP and managing partner, Charles Hudson, where we will
• 0:37
  be talking through the articles and Goodreads that caught our team's attention over the past
• 0:42
  week.
• 0:43
  Our first topic is written by Jerry Newman titled AI Will Not Make You Rich.
• 0:48
  And while extremely thorough, it turned out to be one of my favorite reads in a while.
• 0:52
  He uses economist Carlota Perez's four phase tech wave framework, eruption, frenzy, synergy,
• 0:59
  and maturity to show that some tech phases favor investors far more than others.
• 1:03
  Newman draws a compelling parallel between AI and containerization, emphasizing that
• 1:08
  transformative technologies often start slow before gaining momentum just as electricity and
• 1:14
  personal computers did, taking decades to reach widespread adoption through gradual self
• 1:19
  reinforcing cycles of experimentation and practical use.
• 1:22
  When shipping containers revolutionized global trade, they eventually created a world of
• 1:26
  perfect competition where no company could sustain protected profits.
• 1:30
  The way the article describes it is AI is following a similar path despite early
• 1:34
  speculation and intense investment, widespread adoption quickly eliminated lasting competitive
• 1:40
  advantages.
• 1:40
  Instead of protected profits for innovators, we got a competitive free for all as everyone
• 1:46
  gained access to the same AI tools.
• 1:49
  The lesson from it feels clear, transformative technologies frequently end up commoditizing
• 1:54
  the very advantages they initially promise.
• 1:58
  So, Charles, there are so many routes to go here, so I wanted to pause first for your
• 2:01
  reaction.
• 2:02
  don't know.
• 2:03
  Anybody who hasn't read the book, The Box, I highly encourage it.
• 2:06
  It's a great book on just the development of the standardized shipping container and how
• 2:10
  transformative it was.
• 2:11
  When I read it, I thought a lot about general purpose technologies and, you know,
• 2:15
  technologies like GPS and electricity where the benefits are really widespread.
• 2:22
  And there are certain elements of I feel like my thinking on this has maybe evolved a little
• 2:26
  bit.
• 2:26
  Maybe a year ago, would have said AI is all general purpose technology.
• 2:29
  It's gonna be hard for anybody to capture value.
• 2:32
  Yeah.
• 2:32
  I think there's there is the distinct possibility that the large language model
• 2:36
  providers end up having some durable lasting advantage simply because of the cost complexity
• 2:44
  and sheer scale required to develop, iterate on, and maintain those models.
• 2:50
  But I continue to believe that they will capture a relatively small fraction of the
• 2:55
  total value created by AI.
• 2:57
  And that more of the value, you know, I'm thinking very seriously about making this one
• 3:02
  of the themes for our annual meeting, later this year, is just this idea that if you think
• 3:08
  about the surplus that's created by AI, everybody's focused on the surplus value that's
• 3:12
  created.
• 3:13
  No one's really focused on, at least I feel like, enough attention being paid to the split
• 3:17
  between how much of it will be captured by the people who produce the AI Yeah.
• 3:21
  Versus people who build things using it.
• 3:24
  And again, I think it's hard for 90% gross margin software people to envision a world
• 3:29
  where the customer is the principal.
• 3:31
  And this sounds like software people are robber parents, but it's hard to envision a world
• 3:35
  where margins are lower precisely because the end user is getting more bang for the buck
• 3:41
  Yeah.
• 3:41
  More software.
• 3:42
  The power of, know, if you just think about software in terms of how much power is there
• 3:45
  per dollar, software power per dollar will go up dramatically Yeah.
• 3:50
  Which means you won't be able to charge as much.
• 3:52
  And I think it was a really good piece.
• 3:54
  Also just, you know, these things go in waves and cycles.
• 3:57
  Yep.
• 3:57
  So I don't know.
• 3:58
  I thought it was a really good piece.
• 3:59
  And it was something that four or five different people sent to me and said, you
• 4:02
  should cover this.
• 4:03
  I mean, to that same point, I wanna go back to Ashley Perez's graph where it outlines the
• 4:07
  stages of tech waves.
• 4:09
  So it's eruption, frenzy, then you have synergy, and maturity.
• 4:14
  And so I certainly feel like we're at the frenzy stage as it pertains to AI, but I know
• 4:17
  there are investors who actually believe they're at the eruption stage still.
• 4:21
  We haven't even started scratching the surface of of nearing the end of the right time to come
• 4:26
  in, which is why they're dumping so much capital there.
• 4:28
  So from your perspective, are you kind of thinking that that door has closed?
• 4:32
  I think we're in the frenzy stage because I feel like in the eruption phase, there's nobody
• 4:38
  who calls it a bubble.
• 4:40
  Yeah.
• 4:41
  There's no one who says things have gotten extended because you're so focused on the
• 4:46
  optimism and the possibilities.
• 4:48
  The reason I think we're in the frenzy stage is privately I've talked to investors who are
• 4:52
  like, look, I know that this is frothy and insane, but I'm gonna ride the wave.
• 4:57
  I'm gonna play the game on the field.
• 4:58
  What whatever analogy whatever sports or nautical analogy you wanna use.
• 5:02
  That's what they're gonna do.
• 5:04
  And if that's your plan, all that matters is that you can get in and out before the bubble
• 5:08
  burst.
• 5:09
  Yeah.
• 5:09
  And or you can get in and out while there's still money to be made during the frenzy.
• 5:14
  And I think there are certainly venture firms whose either public or private restated
• 5:19
  strategy is to do just that.
• 5:21
  Well, our next topic is from Cameron Peak over at Restive Ventures titled consumer businesses
• 5:26
  are stronger than ever.
• 5:28
  Where are the investors?
• 5:30
  The article walks through the phenomenon of a massive disconnect in consumer tech So while
• 5:35
  Chime's successful IPO signals a comeback, VC funding tells a really contradictory story.
• 5:41
  Consumer shares of venture dollars crashed from 15% in 2020 to just 5% in near in early twenty
• 5:48
  twenty five, making recent wins like Polymarket and Bill just rare exception.
• 5:53
  But public markets are thriving.
• 5:54
  Consumer tech companies crushed the Nasdaq 100 up 63% year over year, driven by strong
• 6:00
  earnings from Uber, Robinhood, and Netflix.
• 6:03
  So despite recession fears, consumer startups are actually outgrowing b to b peers,
• 6:08
  especially in fintech, where it feels like uncertainty drives innovation.
• 6:12
  Peak identifies AI powered personalization as the next catalyst.
• 6:17
  Companies using AI for hyper customized services are defining consumer tech's next
• 6:22
  trillion dollar wave with OpenAI's ChatGPT evolution proving this opportunity is really
• 6:27
  just the beginning.
• 6:28
  I was reflecting on this, and and being a generalist fund has allowed us to continue
• 6:32
  being bullish sectors that have fallen out of favor for other funds.
• 6:36
  From your perspective, do you think AI is going to give that lift on consumer to really bring
• 6:40
  it back into good greases with the majority of investors?
• 6:44
  I've read so many articles about how AI is gonna be this catalyst in the next wave, and I
• 6:50
  see a just a gigantic disconnect between the number of articles written and the number of
• 6:55
  checks written.
• 6:55
  Yeah.
• 6:56
  And I've had this theory for a while that part of what's happening in consumer is I don't have
• 7:02
  that many GP level friends of mine who are focused on consumer outside of some very
• 7:08
  specific funds or people who've been doing this for a really long time.
• 7:12
  And I've gone to a few events that are more focused on associates and principals, and there
• 7:16
  are tons of people there who are interested in consumer.
• 7:20
  And so my working theory is that the people at a lot of venture firms who control the money
• 7:24
  and the investment agenda are not as interested in consumers as they once were, either because
• 7:29
  it's been a long time since they had a hit Yeah.
• 7:31
  Or they don't feel the same cultural connection to the site guys.
• 7:34
  For whatever reason, they're not as interested.
• 7:37
  And so what the firm invests in reflects what those people are invested in, and right now
• 7:41
  that's AI infrastructure.
• 7:43
  Whereas maybe some of their younger teammates either haven't been scarred by consumer
• 7:47
  investing or actually just closer to today's consumer Yeah.
• 7:51
  And can see trends and see things that others might miss.
• 7:56
  So I I was surprised that it was even 5%.
• 7:58
  If you'd asked me what it was, would have said it probably 1%.
• 8:01
  I feel like zero.
• 8:02
  I guess.
• 8:02
  Mhmm.
• 8:03
  And I look in our portfolio, I look at some of our consumer both consumer software and CPG
• 8:08
  companies.
• 8:09
  They're some of the best performing companies in our portfolio, and I I think the irony is
• 8:13
  for many of those companies, access to capital is not assumed.
• 8:17
  So you have to build a business that, if need be, can be profitable and still grow.
• 8:23
  And I think it enforces a unique discipline on those companies to both build real businesses,
• 8:28
  but not real businesses that are small, real businesses that end up being interesting.
• 8:33
  And so I'm not surprised if those companies outperform.
• 8:36
  They have it tougher.
• 8:37
  Yeah.
• 8:37
  I mean, that's the internal dialogue I have with myself.
• 8:40
  A lot about consumer.
• 8:42
  I find, to your exact point, that they have to hustle the hardest to get the smallest amount
• 8:46
  of money, and then have to constantly be thinking about other streams of capital that
• 8:52
  they can pull to just make everything come together.
• 8:55
  And even when they've hit all of the metrics and the traction that investors told them they
• 9:00
  had to hit in order to unlock capital, they oftentimes just, like, won't get it and will
• 9:04
  never get that interest.
• 9:05
  So if I'm a consumer founder that's so defeating to just know that like, has it just
• 9:11
  completely fallen out of favor no matter how I dress it up?
• 9:14
  And if that's the case, should an investor just be open?
• 9:17
  I'm doing absolutely no consumer and here's the reason why and it has more to do with scar
• 9:21
  tissue than anything else.
• 9:22
  I've had some people telling me like, hey, I'm just not doing consumer anymore.
• 9:27
  Don't send me any.
• 9:27
  We've also had other firms that have, you know, had big transitions in their consumer investing
• 9:33
  team and it's unclear who they're still focused on it.
• 9:36
  I will say this, we're probably one or two more big consumer, you know, there's some big news
• 9:41
  from Aura today like Yeah.
• 9:42
  We're probably only a few consumer wins away from people.
• 9:47
  Maybe now that you you mentioned the way you said it made me think of something.
• 9:50
  I think it's very possible that these consumer wins could start to stack up right as AI
• 9:55
  investing fatigue sets in.
• 9:57
  And it could set the stage for people to once again say, maybe we should go back and look at
• 10:02
  consumer anymore.
• 10:03
  Ninj.
• 10:04
  Maybe there are, like, interesting valuable companies here now that we've had our fill of
• 10:08
  AI.
• 10:08
  Yep.
• 10:09
  I'll be curious to see, but I find it very hard to fundraise for our consumer companies.
• 10:14
  But the ones that really do well seem to be able to attract capital.
• 10:17
  Yeah.
• 10:18
  Cameron talked a lot about consumer tech companies getting
• 10:21
  a
• 10:21
  lot of attention, specifically on the fintech side.
• 10:23
  During times of uncertainty, think a looming recession or economic peaks and valleys.
• 10:29
  Do you think that's something like, what she mentioned, AI driven personalization is a
• 10:33
  specific factor that could bring everyday consumers a sense of peace, or are there other
• 10:38
  things that we're missing that consumer startups could capture with the support of AI
• 10:41
  that could be really valuable and catapult them back into the favor of other investors?
• 10:45
  I mean, I think a lot of my friends who used to do consumer who aren't doing it anymore, their
• 10:50
  argument to me is most consumer waves are built out of distribution innovation.
• 10:55
  Social media, mobile phones, there's the app store, something happens that creates a new
• 11:02
  channel to distribute stuff that people didn't understand before.
• 11:05
  And there are companies that rush in and take advantage of that void to build and they're
• 11:09
  like, we haven't had one of those.
• 11:11
  I think there's some hope perhaps that LLM based search will create something akin to SEO
• 11:17
  as an industry that will allow people to get discovered and there'll be a way to, you know,
• 11:21
  be seen by the LLM search engines in the same way that people figured out how to work through
• 11:26
  and work with SEO.
• 11:28
  But I think barring a new platform or some big new chance to get distribution, I think a lot
• 11:35
  of people are gonna say that that's a necessary condition for me to get excited about consumer.
• 11:39
  Yep.
• 11:39
  Well, last topic is a New York Times article highlighting major tech news.
• 11:44
  NVIDIA will be investing a $100,000,000,000 into OpenAI.
• 11:48
  This deal allows OpenAI to use NVIDIA's AI semiconductors and its data centers instead of
• 11:54
  spending tens of billions to build most of their own facilities.
• 11:57
  This follows OpenAI's pattern of partnering with companies whose services they can
• 12:00
  leverage, Microsoft being another investor that provides computing power with Azure's data
• 12:05
  centers, Oracle for additional data center infrastructure, and SoftBank to help keep The
• 12:10
  US ahead in the AI arms race.
• 12:12
  This investment is completely separate from OpenAI's recent $40,000,000,000 funding round
• 12:18
  that valued the company at around $300,000,000,000 at that point.
• 12:22
  I know it's shifted and changed a little bit today.
• 12:25
  The interesting takeaway is that NVIDIA has made several investments into companies that
• 12:29
  rely on their products and chips, like CoreWeave and Lambda.
• 12:33
  We actually talked about CoreWeave three or four months ago in that connection and calling
• 12:37
  into question that relationship.
• 12:39
  This raises just questions about whether NVIDIA is artificially inflating its financial health
• 12:44
  by investing in companies that then buy its products.
• 12:46
  And again, we've discussed that all through and through, but we're back in this interesting web
• 12:52
  between data centers, chip makers, and foundational model companies.
• 12:56
  Why, from your perspective, is it important to understand how interconnected all of tech's
• 13:03
  largest players are with one another?
• 13:05
  When I first read about this deal, I thought about and maybe this was like a knee jerk
• 13:10
  thing.
• 13:11
  Back when I first got into Internet investing when AOL and the big search engines would do
• 13:16
  these these ad exchange deals where no money would change hands, but both companies could
• 13:21
  claim offsetting value in these round trip deals.
• 13:24
  I'll give you a $100 to spend a $100 on my platform.
• 13:28
  I get the logic of it.
• 13:29
  If you're NVIDIA, you're like, look.
• 13:31
  I wanna make sure that OpenAI doesn't feel like they have to go somewhere else to get chips.
• 13:37
  Like, they need they need they're gonna need silicon to do their compute.
• 13:42
  We would like to keep them as a customer.
• 13:44
  We'd like them to not have to go out.
• 13:45
  And if you're Sam Altman, I'm sure you're like, I'd like to not be in the data center building
• 13:50
  business and have to do all of this other infrastructure.
• 13:53
  So why not?
• 13:54
  And in most businesses, it generally pays for a big customer and a big supplier to have a good
• 13:59
  healthy relationship.
• 14:01
  I think an investment that also allows your customer to purchase almost an equivalent
• 14:07
  amount of services from you in the future, it's kind of an interesting structure.
• 14:11
  But I think right now so much stuff is blocked by Nvidia.
• 14:16
  I don't mean in a bad way.
• 14:17
  They're just at the center of so much with on both the software and the hardware side of the
• 14:24
  current state of AI.
• 14:25
  It reminds me a lot of again, I keep using this, like, late nineties analogy.
• 14:29
  In that Internet era, it felt like Cisco was at the center of everything because we were
• 14:33
  building these huge networks and everybody needed routers and routing.
• 14:37
  And Cisco was this, like, really important router, no pun intended, of value through the
• 14:43
  whole ecosystem.
• 14:44
  And it feels like so much of the value right now is tied up in NVIDIA.
• 14:49
  And so if you're Sam Altman, you're like, should have a close relationship with
• 14:53
  Microsoft.
• 14:53
  We should have a close relationship with Mazda.
• 14:55
  You should have a close relationship with the US government.
• 14:57
  You should have a close relationship with NVIDIA.
• 14:59
  These are all people who could have an outsized impact on your business, and it's probably
• 15:04
  better to keep them close.
• 15:05
  Yeah.
• 15:05
  I mean, we are VCs in this space, and all of our investment end up being intertwined.
• 15:10
  Whether you like it or not, it feels like down the road.
• 15:13
  What are things that we have to kind of consider day to day, especially if we're
• 15:16
  deploying a lot of money in this market and being very bullish on outsized returns in these
• 15:21
  pockets?
• 15:22
  Yeah.
• 15:22
  I I wonder if some of these businesses that are built on arbitrage or hard to get access to
• 15:30
  things, NVIDIA chips, like I wonder if some of these businesses go away because they were
• 15:35
  important in a moment in time.
• 15:37
  And then we solve the supply issue.
• 15:39
  Like, I think it's interesting that Andreessen has their cluster of GPUs that they operate and
• 15:45
  make available to portfolio companies.
• 15:46
  I never thought VCs would be in that business.
• 15:49
  I'm not sure that that's and I'm not sure that that's a long term business Yeah.
• 15:52
  That Andreessen or any other VC fund will be in or should be in.
• 15:56
  But there's probably a moment of time in scarcity where that is a very useful asset to
• 16:00
  have.
• 16:00
  And if you think about, you know, what's one of the arguments for having a lot of management
• 16:04
  fees, it's the ability to do creative things that will both help your portfolio and
• 16:09
  hopefully give you an edge in winning deals.
• 16:11
  So I I still think this goes back to your the first article.
• 16:15
  I still think we're in this phase where, like, we don't have enough of all the stuff we need.
• 16:19
  We don't have enough power.
• 16:20
  We don't have enough data centers.
• 16:21
  We don't have enough GPUs.
• 16:23
  We don't have enough AI researchers.
• 16:25
  So people are doing all kinds of strange things to get their hands on scarce assets.
• 16:30
  That's why Zuckerberg is paying hundreds of millions of dollars for these very scarce AI
• 16:34
  researchers.
• 16:35
  And that's why Google is doing deals with power companies around power usage for some of these
• 16:41
  data centers.
• 16:42
  I think it's partially why nuclear power or nuclear, whatever whatever your chosen
• 16:47
  pronunciation is, is is back on the agenda because people realize like we don't have
• 16:52
  enough of the stuff that we need Yeah.
• 16:54
  To power all the AI that we want.
• 16:57
  And so people are gonna get creative until those things are available in roughly
• 17:03
  appropriate supply relative to the demand.
• 17:06
  Yep.
• 17:06
  My worry is humans are very bad at forecasting demand at times of rapid change.
• 17:13
  And while we eventually needed most of that fiber we put in the ground in the late
• 17:18
  nineties, we didn't need it in the February.
• 17:20
  Yeah.
• 17:21
  And I worry that what happens if it turns out we don't need all of these data centers that
• 17:26
  were built?
• 17:26
  That's a whole that's probably another podcast about like private credit data centers and
• 17:30
  power.
• 17:30
  And the the infrastructure, real asset investments that are happening outside, that
• 17:34
  are all predicated on the ability to accurately predict future demand for AI.
• 17:39
  And I'm I'm not so sure we're we've ever been good at doing that.
• 17:42
  Yep.
• 17:42
  My one last thought is, I guess, if you're a fund manager that's early stage and you just
• 17:47
  create an entire thesis around AI, is your assumption that those things will be figured
• 17:53
  out by the time that your companies grow and grow in scale?
• 17:56
  Yeah.
• 17:57
  I think either that or you hope that your companies get big enough that they can get a
• 18:01
  deal like this Yeah.
• 18:01
  If they need it.
• 18:02
  Yeah.
• 18:02
  I think the hard thing is if you're big enough to need this deal, but not big enough to matter
• 18:06
  to the party on the other side of the table, that's gotta be awkward.
• 18:09
  Well, we are all out of time for today.
• 18:11
  Thank you, Charles, for stopping by.
• 18:12
  Thank you to our listeners for hopping in.
• 18:14
  Please like and subscribe, and we will talk with you all soon.
• 18:17
  Thank you.
• 18:17
  Bye.