How do you quietly raise $9 billion in a world obsessed with hype? In today’s episode, we break down the rise of Lightspeed Venture Partners: the ultra-successful, strangely under-the-radar mega-fund shaping the next decade of AI, enterprise, and consumer tech. Lightspeed has posted huge exits this year while sidestepping the froth of the AI bubble… so what are they doing differently?
But first, we dig into the new rift between Amazon and OpenAI, and how shifting alliances in Big Tech are reshaping the AI economy. Why is Amazon repositioning now? What does it signal for OpenAI, Anthropic, and the broader AI stack? And who actually benefits when tech giants redraw the map?
This episode goes deep on: • What makes a mega fund — and why so few succeed • How Lightspeed raised $9B without becoming a public personality cult • Amazon’s evolving AI strategy and why it matters for everyone • Whether the “AI bubble” is real — and who’s insulated if it pops • Who really controls the future of the AI economy
If you want to understand power in Silicon Valley right now, this is the episode.
00:00:00
What does it take to be a mega fund in the world of venture
00:00:03
capital? Today, we're talking about a
00:00:05
mega fund. You may never have heard of
00:00:07
Lightspeed Venture Partners, who have done a terrific job of
00:00:10
flying under the radar despite raising $9 billion alongside
00:00:15
some strong exits this year. In this episode, we're going to
00:00:18
dive into what makes a mega fund using the example of Lightspeed.
00:00:22
We'll look at their business model, the startups they're
00:00:24
invested in, the IPOs that they're leading, and how they're
00:00:27
seeing significant growth while seemingly spitting in the face
00:00:31
of the AI bubble. But first, we're going to
00:00:33
discuss the new developments in the relationship between Amazon
00:00:37
and open AI and the shifting allegiances in AI.
00:00:40
This is the newcomer podcast. There, there, there was a time
00:00:52
where you one of the most exciting things that could have
00:00:55
happened but didn't happen in Silicon Valley would be that
00:00:58
Amazon would buy Anthropic. Amazon, you know, a big cloud
00:01:03
service provider invested in Anthropic doesn't seem to have
00:01:07
figured out foundation models. But we've really put a nail in
00:01:12
the coffin, I think this week of that idea.
00:01:16
Amazon just turned around and struck a huge deal with open AI.
00:01:22
Tom, you cover these cloud companies, the foundation models
00:01:27
and the deals. They're striking, Yeah.
00:01:29
What do you what do you make of Amazon's decision to strike an
00:01:33
enormous deal with Open AI? Yeah.
00:01:36
Well, we should have some caution here that it's just sort
00:01:38
of in discussions right now. So I don't know if it's.
00:01:40
In talks, it hasn't closed yet. Yeah.
00:01:43
But I mean, the origins of this basically was a couple of months
00:01:47
ago, Amazon and Open AI announced that they had signed a
00:01:52
multi billion dollar cloud computing deal with AWS where it
00:01:55
was going to be hosting some of Open AI models on AWS.
00:01:58
And it was sort of the first big non Microsoft deal that Open AI
00:02:02
struck once it got out of it it's exclusivity agreement.
00:02:06
And, you know, in case you were worried after Amazon and Open AI
00:02:10
struck a deal that it wasn't circular enough, fear not, fear
00:02:14
not. And, you know, people always
00:02:16
wonder with Open AI, how are you going to pay for it?
00:02:19
The answer is more money from the cloud computing companies
00:02:21
that they work with. So it looks like Open AI is in
00:02:25
talks to get around $10 billion, maybe more from AWS.
00:02:29
And, you know, just to, like, take some pride in the way this
00:02:33
is all progressed. We did a story, I think it was a
00:02:35
newsletter item a couple of weeks ago, where I talked about
00:02:39
how Amazon had really started to shift away from its anthropic
00:02:43
relationship, that for a time anthropic models were the Amazon
00:02:47
bet, and that was the way it was discussed internally.
00:02:49
I was told Jeff Bezos was going around talking about anthropic
00:02:52
as the Amazon bet, and it's clear that they have shifted
00:02:56
allegiances a little bit behind the scenes.
00:02:58
When I wrote that, AWS was very upset with me and they said that
00:03:03
to say that Anthropic is no longer, you know, our big bet
00:03:06
and one we're committed to was absolutely wrong.
00:03:08
And they pointed to all these press releases that Anthropic
00:03:10
would make claiming that, you know, you know, Anthropic is
00:03:13
still our primary provider. Sorry, AWS is still our primary
00:03:16
provider and we use their chips. Yeah, that's clearly changed.
00:03:20
That's that's no longer the case.
00:03:21
There's been a. Shift there.
00:03:22
Yeah. And, and I think as you wrote at
00:03:25
the time, that newsletter item, like I think there's been
00:03:27
shifting allegiances across the tech world where once Microsoft
00:03:31
was the exclusive provider for open AI and then they've
00:03:34
invested or planned to invest billions of dollars into
00:03:37
Anthropic. Amazon obviously made their big
00:03:39
play with Anthropic. Now they seem a lot more
00:03:41
attached to open AI. Google is kind of relatively
00:03:45
neutral, but more focused on their own models.
00:03:47
And I think it just goes to show that there is just such an
00:03:51
insane need for compute right now and an overall belief in AI
00:03:55
that there's no reason to pick a horse.
00:03:57
You really kind of just need to bet on the industry at large and
00:04:01
just take as much money and give as much money to keep it going.
00:04:04
Amazon's also trying to get Open AI to use its chips here, right?
00:04:07
Yeah, that's part of the deal. Amazon is building chips that
00:04:11
they want open AI to be using to diversify off of NVIDIA GPUs.
00:04:16
Also in Anthropic deals of AI world too, they announced a
00:04:21
couple months ago a deal with Google to use Google's chips.
00:04:24
So Google is has a partnership with Anthropic as well in this
00:04:28
even beyond, you know, focusing mostly on and their internal
00:04:31
products they do have some ties to.
00:04:33
Anthropic, maybe Anthropic somewhat jilted Amazon first by
00:04:37
doing stuff. With Google, a Google deal.
00:04:39
So then we have to circle back around here, yeah.
00:04:42
Yeah. Well, but also keep in mind that
00:04:44
Google had already invested money in Anthropic before that.
00:04:48
They're they're they're a significant investor.
00:04:49
And I actually, my understanding is that most Anthropic training
00:04:54
and inferencing, at least training I know is done on TP us
00:04:58
on Google's chips. So there was already like a deep
00:05:01
relationship there. I mean, the fact with Amazon is
00:05:03
that they're kind of like the rich kid that has to buy their
00:05:06
friends. No one likes their chips.
00:05:09
They just aren't very good. I'm sorry.
00:05:10
I'll, I'll, I'll, I'll say this strongly on the podcast.
00:05:13
I'll have anyone come in here and disagree with me, But every
00:05:15
conversation that I've had with people in the space says that
00:05:18
like the Tranium and Inferencia chips that that Amazon makes are
00:05:21
like a huge tier below certainly Nvidia's chips, but also
00:05:26
Google's Tpus. And so they got to, they got to
00:05:28
get people to use them somehow. And $10 billion is not a
00:05:31
terrible way to do it. And if there's a core financial
00:05:34
engineering risk with AI, it's that I invest in you and then
00:05:42
you use that investment to do me a partial favor by pretending
00:05:46
like there's enthusiasm for my product, even though there isn't
00:05:50
really enthusiasm, right? So in the case of chips I Amazon
00:05:55
invest in, you open AI and then you turn around and use my
00:06:00
chips, show a little enthusiasm, and then maybe I get a market
00:06:04
premium on my stock. My stock goes up.
00:06:07
That's great for everybody, you know, everybody involved and,
00:06:11
and who cares, right? And similarly this, you know,
00:06:15
this is the cynical case. The cloud companies benefit from
00:06:18
open AI, anthropic, etcetera, spending on these enormous
00:06:22
training runs, right? So they want to keep this
00:06:25
foundation model game going and so keep the money going there
00:06:28
also so it can turn around and spend, spend money on on us.
00:06:33
So there's definitely, you know, this cynical argument that
00:06:37
foundation models like Open AI and Anthropic are so valuable in
00:06:41
part because the people investing them want them to be,
00:06:44
you know, because they're, they're big, they're big
00:06:46
customers and they're validating a lot of the products they're
00:06:50
creating. And This is why to me, this is
00:06:52
the bull case for Google, because Google's on both sides
00:06:55
of it. If you think foundation models
00:06:57
are going to be revolutionary and not only revolutionary but
00:07:00
capture a lot of the economic value of what happens in AI,
00:07:03
Google has Gemini. If you don't and you think all
00:07:06
these models are sort of ludicrously spending on training
00:07:10
runs that are just like lining the pockets of cloud providers,
00:07:14
well, Google has one of those too.
00:07:15
So I I think, yeah, good to be in both camps.
00:07:20
Yeah, they have the chips, they have the models, they have the
00:07:23
customers and they can force these AI models down the throats
00:07:26
of these customers by putting it in all the different, you know,
00:07:29
crevices of their products. Sorry as a gross image but I
00:07:34
think like you know, this all leads to what I thought was a
00:07:37
fucking hilarious video that Greg Brockman the Co founder of
00:07:42
Open AI just made where he, I don't even know what it was for.
00:07:46
But it's basically this video where he talks about the
00:07:49
importance of compute and is like as time goes on we just
00:07:52
need more and more compute. The funnier part of this video
00:07:55
to me was him saying at first we tried other ways to build our
00:07:58
models. Open AI did not set out with a
00:08:02
thesis. The compute was the path to
00:08:04
progress. It's that we tried everything
00:08:06
else and the thing that worked was compute.
00:08:09
We really wanted to find another way to grow our company without
00:08:12
just demanding multi $100 billion of compute deals, but we
00:08:15
can't. This is what it is.
00:08:17
So please, it's like a, it's like APSA.
00:08:19
It was like for only a dollar a day he was getting.
00:08:21
Quirky with it, you know, he was doing like a oh, how do we do it
00:08:26
kind of tone. I don't know.
00:08:27
Yeah. It was for only for only a
00:08:29
dollar a day. You too can keep Open AI afloat
00:08:32
and give us the compute that we need.
00:08:33
Please reach deep if you have GPUs, if you have GB 2 hundreds,
00:08:37
if you have anything that you or your friends might have, please
00:08:39
donate them to Open AI because that's what we need to keep
00:08:42
going. I don't know what the fuck this
00:08:44
video was. I don't know who it was for.
00:08:46
I honestly, if you start to look back at this era from 10-15
00:08:50
years from now, if it does happen to go sideways or doesn't
00:08:52
turn out the way people expected, this video is going to
00:08:54
be one of them. To me, that's just like, what
00:08:56
were we thinking? How do we not understand that
00:08:59
there was like something slightly concerning at the core
00:09:01
of the way these companies run? To be clear, my point of view
00:09:04
remains that we are imbuing everything with the capacity to
00:09:08
be intelligent and there's going to be a huge market for that.
00:09:11
And as they get more and more intelligent, they're going to
00:09:14
get even more valuable. But what that what that means
00:09:17
for the near term value of open AI, I've no idea, you know, glad
00:09:22
that's not my job to figure out. But I do think these things all
00:09:27
come back to the actual your belief on the value that AI is
00:09:31
creating or not, right. If it's creating a lot of value,
00:09:34
then makes sense to build chips and spend money on chips.
00:09:38
It makes sense to to these huge training runs.
00:09:41
But yeah, and I'll be fair to Brockman in his argument was
00:09:46
that like, our products are so successful, they're in such
00:09:49
demand that we need compute to be able to provide power to it,
00:09:53
that, you know, we can build more of them because people want
00:09:55
them so much. It's not that speculative.
00:09:57
It's like we need this now. It's for our products to work.
00:10:00
My only counter to that would be like what products outside of
00:10:03
Chachi PT, you know, like you? Know you do these training runs
00:10:09
and then everybody catches up with you tomorrow.
00:10:13
Right. Well, I think that's also
00:10:14
that's, that's the that's the business question that is
00:10:18
completely separate from the value that AI, this technology
00:10:21
creates and is transformational technology.
00:10:23
I think we're all in agreement here that it is very useful and
00:10:26
is very impressive and there is a ton of utility for it.
00:10:29
But where the business Moat lands and how these deals as
00:10:33
they are structured financially work out, that remains a little
00:10:37
bit still murky. And and as you'll see in the
00:10:40
segment coming up, that's why they need all the money.
00:10:42
That's what the money is for, is to make the compute pay for the
00:10:46
compute that open AI so desperately needs, and so
00:10:49
everything has to be mega sized from that.
00:10:52
Up next, we talked about Lightspeeds $9 billion
00:10:55
fundraising as a venture capital firm, so we'll dig into that.
00:10:58
Stay tuned. Hey folks, I'm Omar, product and
00:11:01
design lead at Google DeepMind. Have you ever wanted to build an
00:11:04
app for yourself, your friends, or finally launched that side
00:11:07
project you've been dreaming about?
00:11:09
Now you can bring any idea to life, no coding background
00:11:11
required, with Gemini 3 and Google AI Studio.
00:11:15
It's called Vibe Coding and we're making it dead simple.
00:11:18
Just describe your app and Gemini will wire up the right
00:11:20
models for you so you can focus on your creative vision.
00:11:23
Head to AI dot studio slash build to create your first app.
00:11:28
We're about to talk about the biggest, most important venture
00:11:32
capital firm that maybe you've never heard of.
00:11:34
Depends how much of an insider you are.
00:11:37
Lightspeed Venture Partners just raised $9 billion, which even in
00:11:44
the world of venture, where we get numb to big pools of money,
00:11:49
this is a venture firm with a lot of money to invest in
00:11:52
startups. Here at Newcomer, we have
00:11:55
published some of Lightspeed old performance numbers, so we have
00:12:00
a good sense of what companies they're investing in.
00:12:03
That gives the limited partners that back venture capital firms
00:12:06
the confidence to re up with Lightspeed.
00:12:09
And Madeline has been doing some reporting on Lightspeed to
00:12:13
understand this latest fundraising can help breakdown
00:12:17
all the numbers. And Tom will weigh in with his
00:12:20
spicy quips as. Tom can weigh in with his
00:12:23
hottest takes. I can I can really dig in on
00:12:26
Lightspeed circa 2016. Yeah, exactly.
00:12:29
You cared a lot. I mean, so if you if you are not
00:12:32
in the venture business, you might have heard of Lightspeed
00:12:34
because they were a big investor in Snapchat.
00:12:37
Jeremy Liu, who we all know invested in Snapchat then I
00:12:42
think as Tom helped brake had a falling out with lightspeed
00:12:46
right? Did you break that story?
00:12:47
I. Actually didn't break that
00:12:49
story. It was it was our departed
00:12:51
former Co host on Dead Cat, Katie Benner, which is a real
00:12:54
blast. Really.
00:12:55
Oh my God, I forgot. Yeah, it was a very different
00:12:57
era where the New York Times would have written a story about
00:13:00
like a fallout between a founder and the whole.
00:13:04
I mean, we can we should get to your story, Madeline.
00:13:05
But if you really want to a trip down memory lane, the whole
00:13:08
history between Lightspeed and Snapchat is a very interesting
00:13:10
one. Not even just the falling out,
00:13:12
but the, the way that they made that investment.
00:13:13
So we can we can talk about that then, but it's like a very.
00:13:16
Different circle back to that. Yeah, it's a very different
00:13:18
light speed was and is mostly a firm that invests in
00:13:24
enterprises, businesses selling to other businesses.
00:13:28
There are exceptions. They're a big investor in a
00:13:31
firm. Max Levchin's payment advance
00:13:35
company, but a lot of the big investments are firms like
00:13:39
Grafana and Navon and fair, you know, it's sort of are, are the
00:13:44
backbone of or not consumer facing companies.
00:13:48
Right. Well, Lightspeed, I think you
00:13:50
know, wants to be known for their enterprise deals more
00:13:54
their their ties to Snapchat, cover the press.
00:13:57
And so people think of them as a consumer first fund and they're
00:13:59
quick to tell you no, we do enterprise deals.
00:14:02
And to their credit, they had significant IPOs in the last two
00:14:07
years. They had, they were big backers
00:14:09
of Rubric, Netscope and Navon, which all went public in this
00:14:13
time of IPO drought recently. So they've had big exits in the
00:14:17
last couple years, which is impressive.
00:14:19
You know, if you say they're the firm that flies under the radar,
00:14:22
they're returning money back to their LP's, which is great when
00:14:24
you're fundraising. Madeline, you want to break down
00:14:26
this $9 billion fundraise and give us a sense of where all the
00:14:30
money's gonna go? So this $9 billion fundraise is
00:14:33
broken up across six different vehicles.
00:14:37
It's all kind of multiple funds under this one umbrella to hit
00:14:40
the $9 billion total. But the two, the early stage
00:14:43
fund is kind of split into Lightspeed 15-A and 15-B.
00:14:48
And the way you can think about that is 15 A.
00:14:50
What marketing geniuses? They've got a number and they've
00:14:53
got a naming scheme like Sony headphones here.
00:14:56
Yeah. Well, 15-A is 980 million and
00:15:00
like the title says, it is for seed up to Series A, maybe a
00:15:04
smaller B deal. Lightspeed B is for is $1.2
00:15:08
billion and that is going to go to things that are more entering
00:15:12
product market. 5th growth Series B onward side.
00:15:15
However, it's not the growth fund.
00:15:18
It's just like in the age of AI, these series BS and Series A's
00:15:22
and billion dollar seed rounds, they wanted a lot of capital to
00:15:26
be able to properly allocate in these rounds and double down on
00:15:29
concentrated bets. And when that moves earlier and
00:15:31
earlier with these AI rounds that everyone wants in, they
00:15:35
wanted capital for those two, wanted capital back those.
00:15:38
So that's how those are structured, the select those.
00:15:42
Are two earlier stages. So those are the 2 early funds.
00:15:45
There's also Lightspeed Select four and Lightspeed Opportunity
00:15:49
Fund three select four, A $1.8 billion fund that is for
00:15:54
doubling down on bets that they want to get a higher ownership
00:15:58
share in that are more growth stage.
00:16:02
So think of it as a kind of medium growth fund.
00:16:05
And then opportunity funds have been actually the past
00:16:09
opportunity funds was where they first backed Anthropic because
00:16:11
they did not get into Anthropic until the Series D, but then
00:16:16
they ended up Co leading and then leading later Anthropic
00:16:20
rounds as they kept raising. So they did get a larger
00:16:23
ownership share later on, but that's what the opportunity fund
00:16:26
would be for to double. Right now seem like great
00:16:28
investments. I know they've been marked up a
00:16:30
lot. There's been, you know, this
00:16:31
chatter in the zeitgeist right now about like, if you're not in
00:16:35
these key AI deals, you're failing behind, but they're
00:16:39
mostly mentioning, you know, Cursor Mercor, Open AI,
00:16:43
Lightspeed is not necessarily in those companies, but they are a
00:16:49
pretty significant shareholder in Anthropic, which for as you
00:16:51
can tell by now on this podcast, we're kind of fans of anthropics
00:16:54
style at this point, you could say in how they they do things.
00:16:58
And it's a it's a strong growing business with top models.
00:17:01
So the two new funds they raised with this fundraise are the
00:17:04
Lightspeed Co investment fund, which is about $600 million and
00:17:08
a single investor vehicle that was $1.25 billion.
00:17:12
Those exist for existing LP's to tack on more money to double
00:17:18
down on specific bets so in and those basically exist because of
00:17:23
how big AI rounds are getting. There are certain LP's who want
00:17:26
to have even more exposure to top AI deals and those funds
00:17:31
exist so they can opt in to give Lightspeed even more money on
00:17:35
ones that they think Lightspeed has an advantage of getting into
00:17:37
further rounds on and getting additional ownership.
00:17:40
So those are what those two additional structures are, which
00:17:42
total it up to $9 billion, which is pretty hefty fundraising.
00:17:46
What was one of them? So sorry, One of them is a
00:17:48
sidecar. There are two kind of sidecars.
00:17:52
Like are these real funds? These are real funds, but they
00:17:56
basically exist because certain LP's wanted to double down more
00:18:02
on later stage deals and it's a way of balancing the pro rata
00:18:06
rights in those deals to where the LP's wanted to participate
00:18:09
more. So all the.
00:18:10
Different terms across some of the, you know, some of the more
00:18:13
favorable. To but you have to be an LP in
00:18:16
the early and growth funds if you want to get into these
00:18:19
special two additional funds. So it's the same LP base.
00:18:22
What's interesting to me with this and like I don't have the
00:18:24
perspective, but like do other funds portion out their funds or
00:18:30
do other firms portion out their funds in such a regimented
00:18:32
segmented way? Or is this like basically a new
00:18:35
marketing scheme for LP's where you're just like you can have
00:18:38
access to the select fund, but if you want further exposure to
00:18:42
later stage rounds, you can be select plus, you know, the
00:18:45
sidecar, whatever they call it. Like, is this all just kind of a
00:18:47
different way of divvying things up so you can essentially get
00:18:50
more from an LP? Well, there's a lot going on
00:18:52
that we don't know, but one reason you'd want separate funds
00:18:55
is venture. How are venture capitalists
00:18:57
compensated? I think that's that's worth a
00:19:01
discussion here, right. So venture capitalists are
00:19:04
compensated based on management fees and carry management fees.
00:19:09
That's the two of the famous expression 2 and 20 means that
00:19:13
they get 2% a year of the assets that they've raised.
00:19:19
So you have bigger and bigger funds, you get larger and larger
00:19:22
management fees. And so part of what's appealing
00:19:25
about mega funds is that you have so much money that the
00:19:28
management fees alone become very valuable.
00:19:31
And who cares how well your investments do because you can
00:19:35
make, you know, if you do the, we should do the math off 9
00:19:38
billion. I mean, it depend not all these
00:19:40
funds are the same fee structure and we don't they don't like
00:19:43
announce the fee structure, but they're making hundreds of
00:19:47
millions certainly in management fees alone.
00:19:50
Now you have to run a bigger firm.
00:19:52
You pay employees right The. Overhead goes up among you know,
00:19:56
the number of partners that sort of thing.
00:19:57
So it does get diluted out, but it's it's a pretty good market
00:20:01
if you can make sums like this off your management fees.
00:20:03
So if you're thinking about management fees, you want to
00:20:05
have a big, you just want to raise a bunch of money, right,
00:20:08
That's great for management fees on the carry side, you like
00:20:12
having different funds, right. So if my venture fund does
00:20:18
really well and I only get paid as a venture capitalist, if I,
00:20:23
you know, double the value of my holdings or if there's some
00:20:27
hurdle that I have to reach, I don't want that to be dragged
00:20:30
down by the fact that then we double down later at bad
00:20:33
valuations. And so we're not going to make,
00:20:36
you know, we, we have a bigger and bigger pool of money we have
00:20:39
to pay out until we reach our hurdle.
00:20:42
So there to hit carry, it can be advantageous to split out funds
00:20:50
so that. Meaning like having early,
00:20:52
having late. Having right early could be in
00:20:54
the money, but opportunity might not, you know.
00:20:57
And so then it's like, oh, we make money off.
00:20:59
Whereas if early and opportunity were the same fund, it's
00:21:02
possible we don't make out and LP's know this.
00:21:08
And this also could be in Lightspeed's case, they have,
00:21:10
you know, 15-A and 15-B, which even splits the early.
00:21:14
Level. Farther so there's like seed A
00:21:16
and BC. So before you even get into
00:21:20
growth, they're tearing it out that.
00:21:21
Way it's possible they have some overall promise to LP's.
00:21:25
I don't know like how much these things are operating stand alone
00:21:31
versus they're saying overall we will make the money back.
00:21:35
But but yeah, I don't, I don't think it's mostly about
00:21:38
marketing and they're probably let you know some LP's
00:21:42
concentrate in areas they like better and you know there
00:21:45
there's, there's a lot of negotiating.
00:21:48
But you don't think there's a way that you can, you know, get
00:21:51
L like more money from an LP by offering them exposure to
00:21:54
multiple funds within the firm? Like it just feels like it's
00:21:58
like one. Yeah.
00:21:59
It's just like 1 chunk of money going to 1 fun versus like, you
00:22:02
know, it's just like a psychological thing of little
00:22:05
bits will add up to a greater number, but it doesn't feel as
00:22:08
psychologically, you know, like a huge chunk that's coming out
00:22:11
of your fund like you're, you're I don't know you're.
00:22:14
Doing different strategies, yeah.
00:22:15
Yeah, yeah. It just feels like there's an
00:22:17
appeal there. It also if you wanted to get a
00:22:19
larger ownership share of one specific winning bet, which
00:22:22
seems to be what everyone wants to do in the.
00:22:24
AI. Race right, like everyone wants
00:22:26
to be an open AI, everyone wants to be an anthropic with these
00:22:29
kind of winner companies, you could have exposure as an LP
00:22:32
from the growth fund, but you could also be in the Co
00:22:35
investment fund and get more that way.
00:22:37
So you could say, you know, they could sell it like, hey, if you
00:22:40
want to get even more in Anthropic, you could opt into
00:22:43
this additional fund. And it's doing so well.
00:22:46
Why wouldn't you do that? You know, I think this new fund
00:22:50
was broken in the New York Times and the way that the Times frame
00:22:55
the story was like AI investment is going crazy right now.
00:22:59
You need to be able to raise huge amounts of money in order
00:23:01
to participate in these mega round seeds or you know,
00:23:04
hundreds of millions of dollars and billion dollar valuations
00:23:06
right off the bat. And this is just a response to
00:23:10
the size of rounds that AI companies are are demanding.
00:23:13
Is that basically what's going on here?
00:23:16
When you see Lightspeed doing this, you know, astronomically
00:23:19
large round comparatively to the other things just like this is
00:23:22
AI, this is the bubble. This is just like the table
00:23:25
stakes for participating. I I think, you know, there are
00:23:27
two things going on in terms of why Lightspeed is able to raise
00:23:31
this amount of money. 1 Limited partners, the the people and
00:23:35
firms and institutions that invest in venture capital firms
00:23:38
or sort of fleeing to the the best of the best related to the
00:23:43
second phenomenon, which is companies are staying private
00:23:46
forever and so there are these clear winners or apparent
00:23:50
winners companies like open AI, Anthropic cursor clean data
00:23:57
bricks didn't. Data bricks just raise like a
00:24:00
series L. Recently, yeah.
00:24:01
It's like a. 160. Billion or something?
00:24:04
I'd never heard of a series L before that.
00:24:06
Right. So they're staying private,
00:24:08
which means that only sort of connected investors can really
00:24:12
get access. And so firms like Lightspeed are
00:24:15
well positioned to do those deals.
00:24:17
And so they need way more than traditional venture capital
00:24:21
money to invest. But what about AI specifically?
00:24:25
Like is it just, I mean, is it, are we seeing rounds that are
00:24:31
fewer and larger, like they're not really passing the hat
00:24:33
around that widely? And so you just are entirely
00:24:36
reliant on like you're saying these big name firms to give you
00:24:40
allocation. You just don't have that much
00:24:41
auction as an LP. Yeah, that's basically part of
00:24:45
this too. There.
00:24:48
He's hardly ever been, at least in my time reporting VC, just a
00:24:52
clear kind of consensus view of what to invest in from all these
00:24:55
firms, right? Like people have collectively
00:24:58
decided that there are the winners in AI not for
00:25:01
applications, the jury's still out there, but on the foundation
00:25:04
model level, a lot of people think there's still a ton of
00:25:06
value in these companies. There's not that many of them.
00:25:09
They're very expensive operating companies, so the money they
00:25:13
need to raise and keep growing is more.
00:25:16
And so then the big players need more money to fund them because
00:25:19
there's a sort of consensus view that these are the companies
00:25:22
that we need to be backing right now.
00:25:23
So it's kind of reinforcing. Yeah, I mean, you know, there's
00:25:27
been some, it's almost like a meme at this point, but I've
00:25:30
just seen it enough on on X to decide.
00:25:32
This is like, I think people are discussing that.
00:25:34
They're like only so many firms or or sorry, there are only so
00:25:38
many start-ups that really matter right now in AI.
00:25:41
And the ones that typically get names are like Anthropic cursor,
00:25:45
open AI, Mercor for whatever reason gets gets Yeah, people.
00:25:49
Love Mercor? Well, OK, I get it.
00:25:52
I mean, you know, there's value in sort of, we basically do like
00:25:55
reinforcement learning in human, like kind of they're basically
00:25:58
like replaced scale as like kind of scale.
00:26:02
Yeah, that's like the go to service.
00:26:04
So sure, whatever. I don't really see those as tech
00:26:06
companies, but they certainly have venture backing.
00:26:08
But you basically if you're an LP, you want exposure to those
00:26:11
companies and you're going to get pissed off if you don't get
00:26:14
it. And so like the value that all
00:26:16
of these firms basically have is say like, oh, we can get you in
00:26:19
it, don't worry, don't worry, we'll get you in there and we'll
00:26:21
get you in there in a big way. And it is incredibly
00:26:23
concentrated. I mean, I don't know, Eric, like
00:26:25
you were, you know, covering late stage venture in like the
00:26:28
Uber era when there really weren't, you know, it was Uber
00:26:32
and Lyft. I mean, was it the similar sort
00:26:35
of competitiveness where there was just like a small number of
00:26:38
firms that had access to these crazy Uber rounds?
00:26:41
You know, in in the hedge fund world, people, you know, think
00:26:45
it's sometimes crazy that it's like, you can raise this money
00:26:48
with great fees and then you turn around and you invest in
00:26:52
Meta, in Alphabet, you know, it's like anybody could do that.
00:26:56
It's like what you're a genius. You like pick which one of the
00:26:59
seven most important companies and go bigot it.
00:27:01
And then you collect fees off of that.
00:27:03
And I think there's a degree to which a only marginally more
00:27:07
sophisticated version of that is now happening in venture
00:27:10
capital, which is, oh, you're going to invest in open AI.
00:27:14
Now VCs have the advantage of saying, well, you schlub can't
00:27:18
get in open AI only I can. I, I don't totally, even though,
00:27:24
you know, I follow this stuff very closely.
00:27:27
I it's hard to come up with a non cynical reason why these
00:27:32
companies, you know, gatekeep their investor base and
00:27:35
therefore allow their investors to basically collect fees on
00:27:40
what are, you know, you know, opening eyes desperate for any
00:27:43
money can get right. So why does it restrict the
00:27:45
investor base? I think the answer is just you
00:27:48
create FOMO, you create this like sense that I need to get in
00:27:52
by making it a little bit difficult by having prestigious
00:27:55
people invest. And the people investing are,
00:27:58
you know, bullish optimists who have an incentive to run around
00:28:02
and say this is a great company. And so instead of just doing a
00:28:05
fair market price, you sort of restrict the investor base and
00:28:09
you create FOMO. And, you know, if you wanted to
00:28:12
to to do it the normal way, you'd go public.
00:28:14
So yeah, it's been this weird situation where, you know,
00:28:18
insiders are getting a pretty good deal.
00:28:20
Now, sometimes they get it wrong, right.
00:28:22
If you dump all the money into the wrong one of these winners
00:28:25
like, sometimes it doesn't work as well.
00:28:29
There are, there are. Certainly concentration risk,
00:28:32
right? And the traditional venture
00:28:33
model was that you had a bunch of different bets that you've
00:28:37
based on your research that could grow.
00:28:39
But if a bunch of them didn't work, you had, you know,
00:28:42
diversified and likelihood was that one or two definitely would
00:28:45
go to the moon and become unicorns and exit for you and
00:28:48
return your fund and make you rich and then some.
00:28:50
And what's different now is you have really capital intensive
00:28:54
businesses that also everyone wants to get in and it's
00:28:56
mutually reinforcing to where there's this hype cycle and it's
00:29:00
really exclusive to get in these deals.
00:29:02
So you want to get on the cap table because it's going to be
00:29:03
once in a lifetime company. And also if you're true
00:29:06
believer, it's like, yeah, this is going to be maybe
00:29:08
generational company and the highest technology and my LP's
00:29:11
want to get it too. To bring it back to a earlier
00:29:14
question of Tom's and to Lightspeed, This is why you have
00:29:16
different funds. You know, you have traditional
00:29:19
venture funds with diversified strategies, with premium fees
00:29:23
that, you know, are looking for unexpected winners.
00:29:26
And then you have these opportunity funds and sidecar
00:29:30
funds that are basically like, yeah, we're gonna, you know,
00:29:33
dump a bunch of money into the names that you know, and if
00:29:36
you're a savvy LP, you already understand them pretty well.
00:29:39
And the fees are gonna be lower, but still like you're, you're
00:29:43
getting to make money by by getting your limited partners
00:29:46
into pretty well known companies.
00:29:48
It's almost like like a cable package or something, right?
00:29:50
It's like if you're gonna get cable these days, it's like,
00:29:54
yeah, it's like, I know I need ESPN.
00:29:56
So like, that's a base level thing.
00:29:57
If I'm gonna get a cable package, you gotta give me ESPN.
00:30:00
Everything else on top of that is like, do I want the TBS
00:30:02
package? Do I want like the Food Network
00:30:05
package? All that stuff is just like
00:30:06
spice on top of it that provides mild differentiation.
00:30:09
But it's like you can't sell cable right now without ESPN.
00:30:13
You can't get, you know, a VC fund without open AI.
00:30:15
Anthropic. But you you also need to put
00:30:17
your head more in that in the mind of the ultra rich, right?
00:30:21
Their problem isn't, oh man, I just want ESPN and they're
00:30:24
making me buy all this other stuff.
00:30:26
Sometimes it's the opposite. It's like, I don't want to think
00:30:29
about all this piddling little stuff.
00:30:31
Like I want to write one huge check, you know, I don't know
00:30:34
who the LP's are, but it's like, I'm Saudi Arabia.
00:30:37
I don't want to see her worrying about like, oh, your venture
00:30:39
strategy or I just want to be like, here's a bunch of money,
00:30:41
like make me some money, you know?
00:30:43
And like, part of the challenge, sincerely, of having huge pools
00:30:46
of capital, even for professional investors like
00:30:49
pension funds or sovereign wealth funds, is you, you want
00:30:52
to put all your money in one place and get a substantial,
00:30:55
reliable return. And firms like Lightspeed NEA,
00:31:00
you know, is the sort of historical example of this where
00:31:03
it's not necessarily aiming for classic venture capital
00:31:07
performance, but saying we can do the market or better.
00:31:11
And it's a huge amount of money, you know, don't worry, leave it
00:31:14
with us. We'll we'll consistently return
00:31:17
and. That's kind of what the mega
00:31:19
funds are replicating. We talked a lot on this show
00:31:22
before about the kind of maturation of venture as an
00:31:25
asset class. And as it becomes more serious,
00:31:27
these mega funds are functioning more like this mature asset
00:31:33
class that these huge LP's, like you said, want a reliable return
00:31:37
and put their money in. And traditional venture was of
00:31:40
course, you know, moon shot ideas that make you really rich
00:31:42
when they work. And when you're dealing with
00:31:45
sums of money this big with institutional LP's with billions
00:31:48
to park, they don't really want to risk losing all of it on the
00:31:52
chance of like a 20X return. You know they would like
00:31:56
reliable 3X4 X 5X which is not venture returns that
00:32:01
historically people would be like.
00:32:02
These are good but not like the crazy amazing returns, but it
00:32:06
would be a great deal for your money.
00:32:08
Well, the other evolution to that also seems to be Thrive,
00:32:11
which, you know, I don't know how much they define this era of
00:32:16
mega funds and acting as quasi private equity firms, but their
00:32:19
involvement like fun to portfolio company is such a
00:32:23
weird change from the way things were in the past.
00:32:26
I mean, this, we weren't planning on talking about this
00:32:27
in the episode, but just interesting to me that like I
00:32:30
think it was this week or last week, you saw Open AI
00:32:32
effectively taking some sort of a stake in Thrive.
00:32:36
So like, you know, where one succeeds, the other one does
00:32:38
too. The whole thing is hand in hand.
00:32:39
I've been reporting on this a bit.
00:32:41
It's through Thrive Holdings, their holding company vehicle,
00:32:44
which Josh Joshua Kushner is also the CEO of, but is not the
00:32:49
main Thrive fund, although it's under the Thrive umbrella.
00:32:51
But it's like in a billion dollar holding company to double
00:32:54
down on these AI bets. That's separate from the venture
00:32:57
model. And this deal is structured to
00:33:00
where engineers from Thrive Holdings and engineers and
00:33:03
researchers from Open AI can be embedded with each other to
00:33:06
learn how best to use the models to grow their other startups,
00:33:11
like in their roll ups business for example.
00:33:13
What? I'm sorry, engineers that
00:33:14
thrive? What are their engineers that
00:33:16
thrive? They have engineers at Thrive if
00:33:18
that's what they say, and they're getting to learn from
00:33:22
the best and brightest models from up in AI.
00:33:25
And I guess the engineers of the company, yeah, I don't know.
00:33:27
And the engineers? Go in and then work on these,
00:33:29
you know, roll up companies where they're like figuring out
00:33:32
how to use AI services to roll up, you know it.
00:33:37
Makes sense for everybody involved.
00:33:39
First of all, self dealing is the name of the game these days.
00:33:43
Josh Kushner is running a great well respected firm with Thrive.
00:33:48
He stood by Sam Altman when the board tried to go after Altman.
00:33:51
He keeps betting on open AI successfully.
00:33:54
Now. What does open AI want?
00:33:56
They want fucking customers. And Thrive is like, oh, we're
00:34:00
going to invest in a roll up vehicle of companies that could
00:34:03
embrace foundation models and use them to revolutionize
00:34:07
plumbing or whatever vertical it is.
00:34:10
And so then Open AI wants those customers and wants to reward
00:34:14
Josh Kushner for being a loyal ally.
00:34:17
So they they share brands and partner.
00:34:19
Yeah, I mean, this is this is how business gets done.
00:34:22
Not even, not even necessarily cynically.
00:34:24
I mean, it's just it, it makes sense for both parties.
00:34:27
But yeah. All this sort of deliver on that
00:34:30
end is I guess TBD, but it sounds.
00:34:33
I know. Well, I'm interested to see this
00:34:34
roll up story there. I mean, there's so much energy
00:34:37
around rollups, but it feels like everybody's interested in
00:34:40
the strategy. You know, we have to see it, you
00:34:44
know, play it out in the the companies.
00:34:46
Yeah, it's too early for a lot of these companies that were
00:34:48
really formed in some cases by the venture with the venture
00:34:52
firms as basically the creators who then bring in a CEO who are
00:34:55
workshop this concept and then acquire an existing service
00:34:59
business or ones where the they meet a founder that wants to
00:35:04
expand their services and say here's some capital, why not do
00:35:07
a roll up and buy this other business.
00:35:09
So they, we will see how these do.
00:35:12
It's only been really in the last year or two that these
00:35:15
companies have started. So the jury's still out on how
00:35:20
successful this will be. Although of course, investors
00:35:23
I've talked to, granted ones who've made these kind of bets,
00:35:26
say that they're doing quite well, but of course they'd like
00:35:28
to. Say that I want to make three
00:35:29
quick points on the mega funds and Lightspeed before we wrap up
00:35:34
this this portion. One, the mega funds, certainly
00:35:41
Andreessen Horowitz and Lightspeed did really well in
00:35:44
the beginning like we published in February Lightspeed's
00:35:47
performance and you know, their first couple funds are first
00:35:52
quartile venture funds. You know, they 3X TVPI, their
00:35:58
first fund, you know they're outperforming other venture
00:36:03
capital firms. And so it is the firm's
00:36:06
obviously you know what you'd expect that did well in the
00:36:09
beginning that said, oh, give us a lot of money and we'll do it
00:36:13
at scale. And then you know, they've had
00:36:15
successful funds, but some of them are second quartile.
00:36:20
But, but what's really important here is that these venture firms
00:36:24
got really big and it takes a while to prove that the really
00:36:28
big strategy works as well as the venture strategy.
00:36:31
So I think we're still in a period we're waiting to see
00:36:35
whether they can do as well with a lot of money as they did with
00:36:38
a little bit of money. The one thing that's happened
00:36:41
since February when it was like, oh wow, they're, they're raising
00:36:44
on an old strategy, while their strategy is much bigger is they,
00:36:48
they've dumped a lot of money into anthropic.
00:36:50
Even at the time, I think it was like their second largest
00:36:52
holding or sign. They're really escalating their
00:36:55
anthropic position. And that's only gotten marked
00:36:57
up. So the AI bubble in some ways
00:37:00
has only supported he's mega fun.
00:37:03
So that's that's one. I think 2:00 we're just going to
00:37:06
be very interested in like, you know, general catalysts, like
00:37:09
will it go public? Andreessen, you know, I think
00:37:12
Ben and Ben and Mark Ben Horowitz and Mark Andreessen are
00:37:15
very savvy marketers. I think they had an LP meeting
00:37:19
where they they opened it up. We're not thinking about going
00:37:23
public, which of course, you know, raises the like, right.
00:37:26
Oh, was it you bring? Up.
00:37:28
Right. You know, it's like we're not,
00:37:29
we're not doing it, but oh, was it on people's minds, you know?
00:37:32
Questions about my Not Going Public shirt are answered by my
00:37:34
Yeah, I've answered. By not public shirt so so you
00:37:38
know, they're all professionalizing, they're
00:37:39
hiring, you know, people from, you know, Blackstone and big
00:37:43
private equity. They're getting wealth managers,
00:37:45
they're pursuing multiple strategies.
00:37:49
And so it's just going to be interesting to watch this asset
00:37:52
class grow up. I do think, you know, small
00:37:54
venture capital firms are so important for the VC ethos that
00:37:59
that the sort of power struggle between mega funds and and small
00:38:04
firms will be interesting to watch.
00:38:06
And then the final point is just if you're in Silicon Valley, you
00:38:10
got to sort of breathe a sigh of relief that firms like
00:38:13
Lightspeed have all this money because their job is to deploy
00:38:16
it. And so even if we're in a
00:38:17
bubble, if the public market bursts tomorrow, Lightspeed
00:38:22
still has all this money pretty locked up now.
00:38:25
You know, they can slow down for six months, but they're probably
00:38:29
going to deploy it at a fairly steady pace.
00:38:32
So the fact that all this money has been put into a venture
00:38:35
capital firm is a sign that as much as people yelled out, yell
00:38:39
about a bubble, at least on the private markets, there's plenty
00:38:42
of money to keep investing in startups and keep keep some of
00:38:46
the activity level. If you know Anthropic, Open AI
00:38:49
or NVIDIA, you know, correct, Maybe that means, you know,
00:38:53
Lightspeed and others shift more money to new companies or, you
00:38:56
know, the big companies stay private and they gobble up a lot
00:38:59
of the money and then not as much for startups.
00:39:02
But it's it's a large pool of money available to the startup
00:39:05
ecosystem, which gives some confidence that, you know,
00:39:08
Silicon Valley's not going away anytime soon.
00:39:11
At the very least, it kind of papers over this narrative that
00:39:14
we keep coming across, which is that we're actually in a down
00:39:16
cycle for startups. And you know, a lot of people
00:39:19
are are like companies are struggling to raise funds.
00:39:21
And we had Kate Clark on here a couple of months ago talking
00:39:24
about, you know, the plight of a lot of these early stage, you
00:39:28
know, a round funds, which is a real thing, but no one really
00:39:31
pays attention to it. When you can see the light
00:39:33
speeds of the world raising multi $1 funds and
00:39:36
startups reaching these half trillion dollar valuations, it's
00:39:39
like that's the recession right now.
00:39:42
It's like companies worth hundreds of billions of dollars
00:39:44
is a recession, which, you know, there's a very strong case to
00:39:46
make that it is one, but it just doesn't compute with, you know,
00:39:50
money still flooding into a small number of megaphones.
00:39:53
Well, before we wrap up on Lightspeed, Tom, what's your
00:39:56
Snapchat story? Sure.
00:39:58
So famously Snapchat their first investor was was Lightspeed and
00:40:03
the story actually goes that. It's the Barry Eggers see like
00:40:08
one of the early kind of Cove Cove cofounding partners of of
00:40:10
Lightspeed. His daughter was in high school
00:40:14
and she was using Snapchat and she would come home and be like
00:40:17
dad, dad, dad. There's this funny app that
00:40:18
everyone's using where he message and it disappears and
00:40:20
it's like he's like, isn't that for sexting?
00:40:21
She's like, dad, there's this really cool app that everyone's
00:40:24
using. And so they were begging Evan,
00:40:28
who didn't he, he didn't like the Evan Spiegel.
00:40:31
He hated VC firms. And so he never wanted to
00:40:33
provide any sort of access to, to, to a lot of funds.
00:40:37
But eventually they broke him down.
00:40:38
And Jeremy Liu, who is now kind of somewhat associated with
00:40:42
Lightspeed, but fairly not. He's he's mostly retired.
00:40:45
Yeah, he was able to convince him to invest, but Evan hated
00:40:48
the terms and he kind of felt that they had him like bent over
00:40:52
a barrel and got too much control, which, you know, that
00:40:55
whole mindset from Evan ended up being like kind of a horrible
00:40:59
aspect of Snapchat. I mean, he maintains.
00:41:01
The company public performance has been terrible, right?
00:41:04
Because. He has total control over it and
00:41:06
so basically. No one can do anything.
00:41:08
Yeah, no. Shareholders, you can't fire the
00:41:10
guy and he's running the company in the way that he's running.
00:41:12
I'm not going to make a judgment, but but anyway, it's
00:41:15
kind of like a seed of destruction anyway.
00:41:17
It's worth it was once like, you know, one of the sexiest social
00:41:20
media companies in the world. He could have sold it.
00:41:22
It's worth $13 billion. He probably could have sold it
00:41:25
for 13 billion. In like 2016 or 2017, yeah.
00:41:29
I mean, Facebook was going to buy it for like a single digit
00:41:32
billions at one point. But at the top, at, you know, at
00:41:34
the top of the market during the, the, the pandemic, I think
00:41:37
it was close to 100 billion. So it's it's a disaster.
00:41:40
But anyway, you know, Lightspeed ends up having a huge falling
00:41:44
out with Evan because he hated the terms.
00:41:46
It was kind of like a Motown artist who, you know, felt that
00:41:49
he was being taken advantage of by, you know, Berry Gordy.
00:41:53
Good reference to anyone who cares that.
00:41:55
Was a. Great reference.
00:41:56
All the best references you know are good only to you, but yeah.
00:42:00
Sure. But anyway, and so, but the
00:42:03
funniest part of this whole story to me is that as a way of
00:42:05
like saying thank you because this is 1 of Lightspeed's
00:42:07
biggest early successes. Like it kind of puts them on the
00:42:10
map and to a point that I think they're annoyed because they're
00:42:13
not really a consumer tech company is they end up.
00:42:17
So this is a high school Barry. Barry Edgar's daughter goes to
00:42:20
this high school in Silicon Valley and they happen to have
00:42:23
like some sort of an endowment or or some sort of fun
00:42:26
associated with the high school. And so they actually give the
00:42:29
endowment some equity into Snapchat from like an A stage
00:42:34
round. And when the company goes
00:42:35
public, that stake ends up being worth like 10s of millions of
00:42:40
dollars. And so there's like random high
00:42:42
school, relatively random high school at the.
00:42:46
Time Snapchat. Yeah, it was 1 on
00:42:48
Snapchat. Hopefully they sold at the IPO,
00:42:50
but. Yeah, anyway, they made so some
00:42:52
high school made a a shit ton of money off of Snapchat because of
00:42:55
Lightspeed, which is a firm that Evan Spiegel hated.
00:42:59
And it's, it was one of my favorite stories that I wrote
00:43:00
covering Snapchat back in the day.
00:43:02
So anyway, there's my Lightspeed Snapchat story.
00:43:05
Cool. Well, that's our episode.
00:43:06
Thanks so much. Thank you for tuning into this
00:43:08
week's episode of the podcast. If you're new here, please like
00:43:11
and subscribe. It really helps out the channel.
00:43:13
Listen in for new episodes every week, wherever you get your
00:43:15
podcasts.