00:00:02
After saying he was out, now Bill Maris
00:00:04
is returning to the investing world. The
00:00:06
founding CEO of Google Ventures has
00:00:08
raised $150 million for his new fund
00:00:11
called Section 32.
00:00:15
>> With a smaller fund, I have the
00:00:17
advantage to be very selective in the
00:00:20
companies that I invest in, the people
00:00:21
that I hire. We're going to invest for a
00:00:23
financial return. Any other metric is
00:00:25
impossible to measure and therefore
00:00:27
won't succeed. Think of the change that
00:00:29
has happened just in the last hundred
00:00:31
years and what's about to happen in the
00:00:32
next hundred years with the advent of
00:00:34
AI. The world is going to change by
00:00:35
orders of magnitude.
00:00:37
>> Thank you very much for that uh warm
00:00:39
welcome. I am Bill Maris. I'm the
00:00:41
founder of section 32. Prior to that, I
00:00:44
was the founder and CEO of Google
00:00:46
Ventures. I was also Google's vice
00:00:49
president of special projects where I
00:00:51
incubated Whimo and Google X Calico uh
00:00:56
uh and many other uh projects as well.
00:00:59
Uh and before that I founded a web
00:01:01
hosting and data center company uh which
00:01:03
we're going to talk a little bit about.
00:01:05
Um and uh today I think I'm going to
00:01:08
talk to you about a few of the lessons
00:01:10
I've learned on these interesting
00:01:12
experiences I've had uh in life. So,
00:01:14
we'll start. We're going to have four
00:01:15
lessons I'm going to talk about. And
00:01:17
we're going to go back to 1997 to start
00:01:20
when I was a uh fresh college graduate.
00:01:23
I had a degree in neuroscience. Uh and I
00:01:26
found myself on Wall Street. Uh somehow
00:01:28
managed to land a job there, but I was
00:01:31
miserable having to wear a suit uh and
00:01:34
trudge to work in the heat. But one good
00:01:37
thing came of that which was I looked in
00:01:40
the closet of the office one day and I
00:01:41
saw a server and I asked, "Well, what is
00:01:44
this thing beneath our jackets?" And
00:01:47
they said, "Well, that's where our email
00:01:49
and websites uh live." And and as can
00:01:53
happen to many of us, I I had a moment
00:01:55
where I felt like I was bathed in the
00:01:57
light of inspiration. And and I thought
00:02:01
I thought I think I've glimpsed the
00:02:03
future. Uh, I I I think I can maybe make
00:02:06
a business out of this because if you
00:02:09
can have our website and email in your
00:02:12
closet, how many websites and emails
00:02:13
could I put in my closet? So, I
00:02:16
immediately quit my job. Uh, because I I
00:02:18
had I had kind of glimpsed through a
00:02:21
keyhole and through that keyhole I
00:02:24
thought I saw the internet and I saw a
00:02:27
data center and it looked something like
00:02:29
this. Or maybe when I say data center,
00:02:32
you think of a something like this or
00:02:34
something like this. But in 1997, a
00:02:37
state-of-the-art data center uh looked
00:02:40
almost exactly like this. Uh we had
00:02:43
three servers. Uh a small, medium, and
00:02:46
large. Uh business grew. We eventually
00:02:49
had five servers. Uh and this isn't a
00:02:51
data center at all. This was my
00:02:53
apartment where I founded the company uh
00:02:55
with credit cards. Uh, and the servers
00:02:58
lived in one room. Uh, the work, uh,
00:03:01
happened in the other room. Uh, and it
00:03:03
would get very hot in that room. Uh, and
00:03:06
this was in Vermont. Uh, so I open the
00:03:08
windows and then it would get very cold.
00:03:10
So cold, in fact, that by noon, if you
00:03:13
had a glass of water in your desk, it
00:03:15
would ice over. Uh, you may think,
00:03:18
though, this isn't so bad. But, but
00:03:20
actually, this was also my apartment as
00:03:22
well. This was the bed. Uh, and you may
00:03:25
look at that and think, well, you've got
00:03:27
a mattress and a nice pillow and then
00:03:29
look at that nice blanket, but this is a
00:03:30
rug I got from Home Depot to keep myself
00:03:33
warm on those nights. And one day there
00:03:36
was a thunderstorm. Uh, the roof started
00:03:38
to leak. Uh, and I knew I needed to do
00:03:42
something because water and computers
00:03:44
and servers don't mix well. So, so I
00:03:46
called the landlord and said, "The
00:03:48
roof's leaking." The landlord said,
00:03:50
"Well, that happens sometimes." Uh, but
00:03:52
I knew that I needed to do something.
00:03:54
So, when you don't know what to do, you
00:03:56
go to Home Depot. I got a bucket of tar
00:03:58
and a mop and I went up on the roof and
00:04:02
there was lightning and there was rain.
00:04:04
And I went up there and I I tarred the
00:04:06
roof. And I did not glimpse the future
00:04:09
in that case because I didn't know when
00:04:12
you're taring the roof that you should
00:04:13
start at the far corner and work towards
00:04:15
the door rather than the reverse. And I
00:04:18
tarred myself into a corner. But the
00:04:20
choice that I faced was either the
00:04:22
servers get electrocuted or perhaps I
00:04:25
get electrocuted. But as an
00:04:26
entrepreneur, I was willing to take that
00:04:28
risk, which you know, news flash, I
00:04:30
survived. Uh my shoes though are still
00:04:33
stuck on that roof uh in Vermont, which
00:04:35
takes me uh to uh lesson two, which is
00:04:40
to see the future, sometimes you need to
00:04:43
be a little bit insane. Uh, it may
00:04:46
appear to those around you that you are
00:04:49
taring the roof in a thunderstorm. And
00:04:51
to that point, I'm going to share a few
00:04:53
slides here that a friend named Stuart
00:04:55
Butterfield was kind enough to share
00:04:57
with me. And here's the inauguration,
00:05:00
1989.
00:05:01
And there's someone taking a picture.
00:05:04
That makes sense. It's probably a film
00:05:06
camera. And 2005 is not very different.
00:05:09
There's still someone back there taking
00:05:10
a picture. And then let's go just four
00:05:13
years later to another inauguration.
00:05:16
And if we look closely, it's quite a bit
00:05:18
different because now everybody's got a
00:05:21
camera. Everybody's got a camera. And
00:05:23
this was kind of before cameras were
00:05:25
mushed into cell phones. It was kind of
00:05:27
around that time it was starting to
00:05:29
happen. But but that's not the most
00:05:30
interesting thing about this photo
00:05:32
because in this crowd is someone who to
00:05:34
his friends, I'm sure seemed insane, who
00:05:37
also did glimpse the future. If we look
00:05:39
closely, this gentleman has decided to,
00:05:42
I don't know, live stream or record the
00:05:44
inauguration on his laptop. Uh, he knew
00:05:47
something that those around him didn't
00:05:50
know, which is one of the things that
00:05:51
I've always looked for in entrepreneurs
00:05:53
is they know a secret about the future
00:05:56
that most of us don't believe.
00:05:58
Let's fast forward to 2007. I find
00:06:01
myself somehow at Google. Uh, and a
00:06:04
challenge was given to me. Uh the
00:06:06
challenge was Google needs a venture
00:06:08
fund.
00:06:10
Uh we were starting to make some
00:06:11
investments. Uh we didn't have a
00:06:13
coherent strategy. There were no
00:06:14
budgets. I had to figure out what to do.
00:06:17
Uh so I first found a friend Rich Miner
00:06:20
who's the co-founder of Android. Uh and
00:06:22
he became my partner in crime as we
00:06:25
conceptualized what what could Google
00:06:27
Ventures be. Uh we went up and down Sand
00:06:29
Hill Road and we we talked to everyone.
00:06:34
anyone that was willing to talk to us
00:06:35
and have a conversation, we were willing
00:06:37
to talk to to see what we what we could
00:06:39
learn. Uh we came up with a plan. Our
00:06:43
plan was to obtain all the data of
00:06:46
venture that we could find. And being
00:06:49
Google, you can imagine it was a lot of
00:06:50
data, historical data, you name it. Uh
00:06:55
then we decided we would as step two use
00:06:57
AI. But at that time, Google would not
00:07:00
let us use the term AI. And this
00:07:02
persisted for many years. Bill, AI is
00:07:06
science fiction. It is it's a hundred
00:07:08
years away if it's ever going to happen.
00:07:10
Uh let's stick to machine learning. By
00:07:12
the way, when you say AI, it freaks
00:07:15
people out. So stop freaking people out.
00:07:16
So we had to call it machine learning.
00:07:18
Uh and we used machine learning to do
00:07:21
two things. design the ideal portfolio
00:07:24
construction by running millions and
00:07:26
millions of simulations uh and back
00:07:28
testing and all of the things you can
00:07:30
imagine that data scientists uh would do
00:07:33
uh and and to determine what the ideal
00:07:35
fund size uh would be and people were
00:07:38
excited. Here's a headline from
00:07:41
TechCrunch at the time uh uh and and
00:07:44
people inside of Google were also pretty
00:07:46
excited. This is one of the senior execs
00:07:47
I later learned uh had this to say. Um,
00:07:51
and you know, I I have to admit it
00:07:54
seemed crazy. The plan seemed uh crazy
00:07:56
at the time, but let's look at how it
00:07:58
turned out. So, over this time period,
00:08:01
2009 to 2018, top cortile VC returns
00:08:04
looked like this and top desile looked
00:08:06
like this. Uh, using publicly available
00:08:09
information, I'm not sharing any uh
00:08:12
non-public proprietary Google
00:08:13
information. We would estimate Google
00:08:15
Ventures returns at about 4.1x.
00:08:18
and I
00:08:20
adhered more closely to the strategy and
00:08:23
the investments that I led and the
00:08:24
investments that I led turned out like
00:08:27
this which takes me to lesson three
00:08:29
which is don't bet against computer
00:08:32
science. I've seen it happen many many
00:08:35
times in many many fields. If you apply
00:08:37
the right kind of computer science at
00:08:39
the right time to the right problem uh
00:08:41
you will get to the right answers. I
00:08:43
would not bet against it, even if it
00:08:46
looks like you're taring the roof in a
00:08:47
thunderstorm. So, let's fast forward to
00:08:49
2017. Uh, I decided to start my own
00:08:53
fund. And again, those around me said,
00:08:55
"You're insane. Why would you do that?
00:08:57
You're in the warm womb of Google. Lunch
00:08:59
is free and massages are a plenty." And
00:09:02
so forth. Uh, but after the idea, you
00:09:04
know, sunk in, I the advice turned into
00:09:06
raise as much money as possible. Uh, you
00:09:09
know, that's the right way to run a
00:09:10
fund. you'll get a big management fee,
00:09:13
you'll be happy, things are going to
00:09:14
work out really well for you. Uh, and I
00:09:16
thought about that relative to
00:09:18
everything I had done up to that point.
00:09:19
And I decided to to not uh take that
00:09:22
advice. Uh, and over the course of my
00:09:26
time at section 32, we've had uh six
00:09:28
funds. We've invested in companies like
00:09:31
Crowdstrike and Coher and Coinbase. uh
00:09:34
and all six of those funds have average
00:09:38
uh about 400 million in size and all are
00:09:40
performing in their top decile and to
00:09:43
the extent there is DPI to measure uh
00:09:45
that's the only measure as far as I'm
00:09:46
concerned in venture that counts as DPI
00:09:50
uh which takes me to lesson four that uh
00:09:52
this is this will be heresy to some but
00:09:54
small funds outperform large funds uh
00:09:57
this is simply the math this is not an
00:10:00
opinion I'm trying to convince you of uh
00:10:02
but there are many reasons for this
00:10:04
smaller funds you can have more focus
00:10:06
you have I mean I've I've already
00:10:08
managed a multi-billion dollar fund with
00:10:10
hundreds of employees it's distracting
00:10:12
you cannot uh give the attention to
00:10:15
founders that I would like to give uh it
00:10:17
it uh there there are many reasons for
00:10:19
this uh and if we look at top desile
00:10:22
performance of DPI um funds smaller than
00:10:26
750 million average return of 4.76x and
00:10:30
funds larger than a billion 2.42x 42x
00:10:34
funds below 750 million across that time
00:10:37
period represented 95% of top decile
00:10:40
performers with discontinuous return
00:10:42
compression above 750 million. Why is
00:10:47
this? There's a lot of reasons for this.
00:10:49
Um uh you can use your own numbers but
00:10:51
I'll just do a little thought
00:10:52
experiment. If you have a $500 million
00:10:54
fund and let's say on average these days
00:10:57
you can own 10% of a company uh you need
00:11:00
$5 billion of exits uh to get your money
00:11:03
back. Let's just remind ourselves that
00:11:05
the 75th percentile of venture loses
00:11:07
money and there is persistence of
00:11:10
performance of the top quartile. So, so
00:11:13
if you need 5 billion to get your money
00:11:14
back and and if you want to be in this
00:11:17
business for the long term, let's say
00:11:18
you set your your goal at 3x, you you
00:11:21
need to return $15 billion of exit value
00:11:23
in your companies. Now, if you have a $7
00:11:25
billion fund, and we do the same math,
00:11:28
you you know, you've got to return 210
00:11:31
billion. 7 billion to to 70 * 3x is 210
00:11:35
billion which uh exceeds the total
00:11:39
venturebacked M&A and IPO exit value in
00:11:41
most years. Uh this year may be an
00:11:44
exception but I that is something I'm
00:11:46
looking forward to talking about when we
00:11:47
sit down. For those of you we've
00:11:49
crunched the numbers, we've done all the
00:11:50
math. Those are Bill's four lessons for
00:11:52
today. I hope that they're somewhat
00:11:54
useful. There's a lot of stories behind
00:11:55
all this and I'm looking forward to
00:11:57
talking about them for a few minutes
00:11:58
with the guys. Thanks so much.
00:12:02
You guys are old friends. Yes, we are.
00:12:04
We go way back. Well, Bill's um when he
00:12:06
started Google Ventures,
00:12:08
>> I was the first ex Google company you
00:12:11
invested in.
00:12:12
>> That's correct. And
00:12:12
>> how did it go?
00:12:13
>> Climate Corp. a billion dollar exit to
00:12:15
Monsanto.
00:12:16
>> What was your multiple? What was the
00:12:18
return?
00:12:18
>> O, I don't know.
00:12:19
>> It was actually good for you guys.
00:12:20
>> It was It was quite good. Yes.
00:12:21
>> Yeah. You guys were in the B and the C.
00:12:23
>> Yeah, it was one billion dollars was a
00:12:25
lot of money then.
00:12:26
>> That back then that was a good deal.
00:12:28
Now, that would have been the C round.
00:12:29
Now it's like an A round.
00:12:30
>> Yeah, that would have been your A round.
00:12:32
>> Now we're going to do it again with Oh.
00:12:33
So,
00:12:34
>> now we're going to do it again with Oh.
00:12:35
>> Um, so, you know, I just want to
00:12:38
juxtapose what you said with what Thomas
00:12:41
shared. They've got a very large kind of
00:12:44
capital base that they invest and
00:12:46
they're investing significantly in these
00:12:48
later stage rounds of these wellproven
00:12:50
companies where it's, you know, the data
00:12:51
he shared is that you can get
00:12:53
significant multiples to get to that
00:12:55
next phase. you know, you're more likely
00:12:57
to go from a billion to 10 billion and
00:12:59
then you're more likely to go from 10
00:13:00
billion to 100 and 100 to a trillion,
00:13:02
trillion to whatever. Um, you know,
00:13:04
doesn't that justify an alternative
00:13:07
strategy to what you're saying of having
00:13:08
smaller funds focused on venture that
00:13:11
you can maybe barbell it, have smaller
00:13:13
vehicles focused on venture and then
00:13:15
very large vehicles that bet on the sure
00:13:18
things that have that durability and
00:13:20
that compounding advantage and you can
00:13:21
kind of have the two together both be 3x
00:13:24
return. So my observation on that would
00:13:26
be one I haven't seen the data science
00:13:28
to support that second conclusion of the
00:13:30
latest stage companies that that can be
00:13:32
a an ongoing trend other than this one
00:13:35
moment this weird moment in time with
00:13:37
these multi- kind of trillion dollar
00:13:39
exits that are coming that that would be
00:13:41
kind of observation one two um would be
00:13:44
at a certain point and this is not a
00:13:47
negative it's just an observation if
00:13:49
you're an RAIA and you're you know
00:13:51
collecting assets that is not venture
00:13:54
you know venture as I practice it at
00:13:56
least is a different craft where you are
00:13:59
making concentrated bets of your time
00:14:02
and capital on entrepreneurs and helping
00:14:04
them build a business and there's
00:14:07
nothing wrong with latestage investing.
00:14:09
However, I also have a an observation
00:14:12
that a a uh a bit of an objection to uh
00:14:17
companies that wrap themselves up in
00:14:20
public benefit language and then uh keep
00:14:25
the value creation uh to themselves and
00:14:28
an elite group of investors through a
00:14:32
big part of the curve and then say,
00:14:34
"Well, we're here to benefit humanity."
00:14:35
Well, what humanity needs is money. So,
00:14:37
it would have might be better to go
00:14:39
public sooner because we'll see how
00:14:42
these multi-t trillion dollar IPOs go.
00:14:45
However, if I'm Google and I don't speak
00:14:47
for Google and I decide to arbitrarily
00:14:49
cut the cost of, you know, tokens to
00:14:53
80%. I'm going to cut them, what happens
00:14:55
to the business models of OpenAI and
00:14:57
Anthropic at that point?
00:14:58
>> What happens? Tell us actually what
00:15:00
Yeah. What does happen? Well, if you're
00:15:01
a company and you can go to Google and
00:15:03
Gemini and you can pay 80% less for that
00:15:08
basically identical product, why
00:15:11
wouldn't you do that? And then the
00:15:15
compression and the pressure on those
00:15:17
other businesses goes super critical.
00:15:19
>> What are the chances that you don't
00:15:20
think that that fall that might happen?
00:15:22
>> If I were Google, that's what I'd do.
00:15:24
Walk us through the scenario where
00:15:25
Google decides with their war chest,
00:15:28
with their money printing machine. You
00:15:30
know what, their margin is my
00:15:33
opportunity. I'm going to give tokens
00:15:36
out 20 cents on the dollar. Every time
00:15:38
they lower their price, I lower our
00:15:39
price. What happens on the playing
00:15:43
field? Walk us through that.
00:15:45
>> Would that not be the rational thing for
00:15:47
>> It's clear they're going to do it. It
00:15:49
may not be a margin though to the they
00:15:51
may be burning investor cash sort of
00:15:53
like an Uber type model. Grab market
00:15:55
share growth
00:15:56
>> capital as a weapon tokens as a weapon.
00:15:58
>> Token as a weapon grab market share grab
00:16:00
an install base on consumer and
00:16:01
enterprise but fundamentally at some
00:16:03
point you got to have cash generation.
00:16:06
>> So that's 100% possible.
00:16:08
>> It's 100%. Look, I'll just, it's been
00:16:10
said before, a trillion for spend
00:16:12
commitments on $60 billion of revenue,
00:16:15
and now you're going to go to the public
00:16:16
and hope that retail is going to pick
00:16:19
that up.
00:16:20
>> Yeah. Tell us about companies staying
00:16:23
private longer and how unfair that is to
00:16:27
the bottom half of society who don't get
00:16:29
to participate.
00:16:30
>> Speak for those 99% who are mostly not
00:16:33
us, right? So, so your 40, you know,
00:16:36
those 401ks, those retirement plans to
00:16:39
get into those companies now, which are
00:16:42
getting bizarre exceptions to S&P 500
00:16:45
rules that all of the rules are being
00:16:47
broken. Uh, the passive funds, the ETFs
00:16:51
are going to have to pick them up. And
00:16:53
where do you think we are on that curve
00:16:55
of value creation? Could they go 3x from
00:16:57
here? Sure. But they
00:16:59
>> so the
00:17:00
just to say it as plainly as possible,
00:17:03
we're going to force overpriced products
00:17:07
on the 401k holders of America who
00:17:10
didn't get to participate early. This is
00:17:11
your position that this is profoundly
00:17:13
and creates more wealth creation for the
00:17:16
people who don't need it. And it makes
00:17:17
the people's retire accounts the bag
00:17:19
holders.
00:17:19
>> There there's a lot of risk in that. And
00:17:21
my my my objection is don't say you're
00:17:25
doing this for the benefit of humanity
00:17:27
and do the other thing.
00:17:28
>> Make the public's retirement accounts
00:17:30
the backholders
00:17:32
>> or just say this is how we're running
00:17:34
our business and this isn't for the
00:17:35
benefit of humanity.
00:17:37
>> Bill, do you think that um what happens
00:17:39
to venture? I asked Thomas this
00:17:41
question. When these dollars get
00:17:43
distributed, there's going to be a
00:17:45
handful of funds that have ginormous
00:17:48
returns. I mean just unbelievably
00:17:50
excessive. Founders 2, you know, is
00:17:53
going to print a hundred billion dollar
00:17:55
return on $200 million of invested
00:17:57
capital. But that's one fund in
00:17:59
isolation. Right.
00:18:00
>> Right. And there'll be a few your funds
00:18:02
when you were at GV are going to print
00:18:04
an enormous upside.
00:18:08
>> And so if you don't look closely though
00:18:10
at beyond the averages, venture is going
00:18:12
to look incredible. If you look past the
00:18:14
averages, venture is still going to look
00:18:15
extremely biodal. a handful of winners
00:18:17
and a ton of losers.
00:18:19
>> How does that play out?
00:18:20
>> I mean, one, that's how venture is,
00:18:22
right? 75% of funds lose money. But two,
00:18:25
in order for Founders Fund or pick any
00:18:27
fund to get that 100red billion out,
00:18:29
they have to sell that stock to someone
00:18:30
else. Otherwise, it's just on paper. So,
00:18:33
who's the buyer for that? Is it is it
00:18:35
retail? Is it uh you know what? You
00:18:38
you've got to make a a business case in
00:18:40
the public market that can show that
00:18:43
this business is worth the discounted
00:18:44
value of its future cash flows. And so
00:18:47
whether it's SpaceX or Enthropic or so
00:18:49
forth like can that case be made? We'll
00:18:51
see six months after or so I know
00:18:54
they're playing with the uh uh with the
00:18:56
lockups to kind of drag that out but
00:18:58
we'll we'll see what the public market
00:19:00
thinks of that.
00:19:00
>> Okay. B So we have we have this one set
00:19:03
of companies and then there's everything
00:19:05
else. What do you like in the everything
00:19:07
else bucket as a venture investor?
00:19:09
>> So so I'm going to make an analogy to
00:19:11
the gaming industry. We all get asked
00:19:14
and we all think about well what does
00:19:15
the future look like? You know when when
00:19:17
AI is everywhere and and there's doomers
00:19:20
on one side and utopian Zork on the
00:19:22
other. That's Zork. I'm going to get to
00:19:23
that. Just give me bear with me 30
00:19:25
seconds. It's probably not as bad or as
00:19:27
great as everyone says. So let's look at
00:19:29
the gaming industry. So I used to play
00:19:31
this game Zork. There's one called
00:19:33
Planet Fall back in the 80s and it was
00:19:35
very brittle. It was turn response turn
00:19:38
response. Grab the lamp. Oh, I didn't.
00:19:41
It's a lantern. I should have said
00:19:42
lantern. Go north and and you wait for
00:19:44
the computer to respond. Let's show the
00:19:46
most sophisticated retail available AI
00:19:49
system out there today on the next
00:19:51
slide. And tell me how different it
00:19:53
looks. So, so what what's happened to
00:19:56
the gaming industry from the 80s to
00:19:58
today is going to happen in AI but in
00:20:00
the next like 5 years. So that will be
00:20:02
compressed in terms of how quickly that
00:20:04
change happens. But we would all agree
00:20:06
games are better today than they were
00:20:08
then. They're photorealistic. you can
00:20:10
like inhabit them and they're they're
00:20:12
moving very quickly. On the AI side,
00:20:15
there will be ambient computing. There
00:20:16
will be uh the problems that Zork had
00:20:19
will be solved for AI. Lack of memory,
00:20:22
lack of consistency, session resets and
00:20:24
and so forth. How did we get there? You
00:20:27
to answer your question, I don't plan on
00:20:30
investing in kind of larger models,
00:20:33
right? Just like it it wasn't uh better
00:20:35
stories that would make better games. It
00:20:37
was controllers and physics engines and
00:20:39
GPUs and and those are the parts of the
00:20:41
AI uh cycle that I'm interested in which
00:20:44
is which is all the platforms that need
00:20:46
to be built to machinery.
00:20:48
>> Correct. That is going to make this
00:20:50
reality real in the next 5 years. And
00:20:52
it's not just bigger models. I think
00:20:54
we're at the Atari command line stage of
00:20:57
of AI and we're going to get to the, you
00:21:00
know, PlayStation 10 stage in the next 5
00:21:03
years.
00:21:03
>> You uh you also used to do a lot of
00:21:05
stuff in life sciences. Yeah.
00:21:07
>> Um, not as much anymore.
00:21:10
>> My interest in life science, I founded
00:21:11
Calico and been very interested in that
00:21:14
space and we were investors in flat iron
00:21:16
uh and ver and lots of other companies.
00:21:19
Uh, I'm very interested in that space
00:21:21
because it has a dual benefit of helping
00:21:24
people and also good do well.
00:21:25
>> Correct. However, uh the uh the
00:21:30
therapeutic space that requires human
00:21:31
clinical trials is a specialist
00:21:33
investment area that uh we're not uh
00:21:37
spending a lot of time on. I'm very
00:21:39
interested in computational biology and
00:21:41
in in those areas which is
00:21:42
>> it seems if you just look on X that
00:21:44
there's a renaissance happening in human
00:21:45
health. I don't know if that's true
00:21:47
whether it's cures for pancreatic
00:21:49
cancer, cancer vaccines, peptides,
00:21:52
obviously there's just an explosion and
00:21:54
a lot of it seems to come back to
00:21:55
computation. Um, but this class of
00:21:58
winners so far is not really
00:21:59
computationally driven. It was just
00:22:01
really good science 10 years ago.
00:22:03
>> Yeah.
00:22:03
>> And so do you think that we're about to
00:22:05
see this massive
00:22:06
>> I hope so. So I started Calico and and
00:22:09
again it was like fringe science
00:22:11
longevity at the time and now we're
00:22:13
investors in uh um New Limit which is
00:22:15
Blake Buyers and uh and Brian
00:22:17
Armstrong's company and a number of
00:22:18
other companies in that space which
00:22:20
doesn't seem so crazy anymore. However,
00:22:23
because of the human biology and the
00:22:27
FDA, if you find a compound and you
00:22:30
think you've got something, that's like
00:22:31
5% of the work. Well, there's still all
00:22:34
kinds of titrating and safety testing
00:22:36
that needs to go on. And so, I don't
00:22:38
think it's going to go quite as
00:22:40
exponential as we would all like it to.
00:22:41
However, if we can achieve a realistic
00:22:45
stimulation of a human cell in silicone,
00:22:47
then you will see that accelerate as
00:22:49
well. We're not quite there yet,
00:22:51
>> but generally we're seeing some might
00:22:53
say a flight of capital to India and
00:22:55
China right now. Are you seeing that
00:22:56
that the biotech path to market is
00:23:00
faster if you invest in firms that are
00:23:03
based off show?
00:23:04
>> I think the US has always um uh indexed
00:23:08
on human safety o over speed to market
00:23:11
and that has cost us in some ways.
00:23:14
However,
00:23:16
some other countries are indexed in the
00:23:18
opposite direction which costs lives and
00:23:20
that so there's a balance there. Uh but
00:23:23
there are certainly there's research
00:23:25
going on in China and other places
00:23:27
experiments and cloning and all sorts of
00:23:29
things that that as far as I know aren't
00:23:31
happening here. Uh, so yes, and I think
00:23:34
the gutting of the CDC and the NIH and
00:23:38
the an anti-science
00:23:41
vibe that has now pervades this country
00:23:44
has driven a lot of mind share elsewhere
00:23:48
as funding is drying up for basic
00:23:50
research.
00:23:50
>> I mean, China's got their own paperclip
00:23:52
model now. They're recruiting some of
00:23:54
the best scientists from Europe and
00:23:56
India, and they're all immigrating to
00:23:58
China. Yeah.
00:23:59
>> To go do work. And that used to be a
00:24:01
scientific pool that we used to access
00:24:02
and we used to recruit
00:24:04
>> and we're losing.
00:24:05
>> We really need the the the neurological
00:24:09
reserves here uh and this business with
00:24:13
>> or brain trust would be another way to
00:24:14
say that as well. But the the h the the
00:24:17
the pushing out of H1B holder like
00:24:21
there's so much happening now that it's
00:24:23
causing it's just easier to go
00:24:24
elsewhere. That's not good for science.
00:24:26
What's your view on what's been called
00:24:28
deep tech for the last decade these
00:24:31
traditionally long investment cycle
00:24:33
capital inensive high-risk like Elon is
00:24:37
one of the few entrepreneurs that has
00:24:39
successfully tackled uh deep tech
00:24:41
business model with SpaceX and Tesla. Is
00:24:45
this becoming a more tractable area for
00:24:48
entrepreneurs to activate and for
00:24:50
investors to invest in because of AI
00:24:52
enablement and physics engines and
00:24:54
>> absolutely because things are moving so
00:24:56
much faster.
00:24:56
>> What areas like that are you focused on
00:24:58
investing in?
00:24:59
>> Uh I mean human biology and healthcare
00:25:02
that's probably the largest TAM in the
00:25:04
world. So super interested in that. And
00:25:06
then all of the others I I mentioned
00:25:08
that kind of underlay uh the AI
00:25:11
revolution which are the the physics
00:25:14
engines and the controllers and the GPUs
00:25:16
and the everything that is going to take
00:25:17
to to get us there.
00:25:19
>> Has Google want to bring I want to bring
00:25:21
Sax in Freeberg before we run out of
00:25:22
time if it's possible. Sax,
00:25:24
>> I'm curious your thoughts on the venture
00:25:26
capital business. I think you've did
00:25:28
five craft funds or four.
00:25:30
>> Well, we've done four venture and two
00:25:32
growth. I'm assuming you're going to be
00:25:34
going back into the venture business,
00:25:35
but I'm curious your take on when you
00:25:37
started in venture and when we started
00:25:39
as entrepreneurs 2530 years ago, this
00:25:42
was a much different playing field. What
00:25:44
are your plans based on, you know, sort
00:25:47
of Bill's
00:25:48
um look at this and do you believe in
00:25:50
the $500 million fund sweet spot or do
00:25:53
you think you need to become Andre and
00:25:55
Harowitz when you go back to the private
00:25:56
sector?
00:25:57
>> Well, I I don't think we need to become
00:25:59
Andre and Harowitz. Um but um you know I
00:26:03
I look I think fund size determines fund
00:26:06
strategy and the size of your fund cuz
00:26:09
you're going to divide your fund size by
00:26:11
20 to 25 names to achieve some portfolio
00:26:16
diversification and construction.
00:26:18
That'll determine your check size and
00:26:19
that sort of determines where you play
00:26:20
in the market. The thing that's spinning
00:26:23
through my head after Thomas's
00:26:25
presentation is, you know, are you
00:26:28
better off just focusing on, you know,
00:26:31
let's call it what used to be called, I
00:26:34
don't know, late venture early growth.
00:26:36
You know, you're writing $50 million
00:26:38
checks. You just kind of wait for the
00:26:40
breakouts as opposed to playing in this
00:26:42
really noisy super early stage game.
00:26:45
Well, I think the problem with that,
00:26:48
>> yeah,
00:26:48
>> is we have to look at the incentive
00:26:50
structure of venture. So, a $5 billion
00:26:54
venture fund that returns 1.01x gets to
00:26:57
say that they are in the 75th percentile
00:26:59
and can raise their next fund and no one
00:27:01
at the Stanford endowment is going to
00:27:02
get in trouble for writing that check.
00:27:04
They need to put two or 500 million into
00:27:07
a fund multiple times. So, so I
00:27:10
understand that dynamic. So, now let's
00:27:12
look at the GP dynamic. Well, if I have
00:27:15
a $5 billion fund, I return 1.01x, I'm
00:27:19
going to make more money than Bill with
00:27:21
his $500 million fund that returns 3x.
00:27:23
Okay, so that's also a strange
00:27:25
incentive. So now let's look at the
00:27:27
entrepreneur side. I am researcher X
00:27:31
from OpenAI. I'm going to start a
00:27:32
company. Bill says, I'll give you $20
00:27:35
million at $100 million valuation. I
00:27:37
want to buy 20% of your company.
00:27:41
Giant fund. Why? We're friends. It's a
00:27:44
different model, but Giant Fund Y says,
00:27:46
"Well, we have this giant fund. We need
00:27:48
to put 250 million in." And then
00:27:51
entrepreneur says, "Well, but my
00:27:52
company's valuation is 100." No, your
00:27:54
valuation is now 4 billion and we'll
00:27:56
give you 250 million for a percent of
00:27:58
your company. They're going to take that
00:28:00
deal every day unless you're a seasoned
00:28:04
entrepreneur who has kind of been down
00:28:06
the road and knows the pitfalls of that.
00:28:08
And so, the incentives are broken in all
00:28:10
those ways. and the pendulum will swing
00:28:12
back. So I don't think just staying late
00:28:15
stage and waiting to sniper at larger
00:28:18
companies will be a long term. The data
00:28:20
would suggest that's not going to work
00:28:21
in the long term.
00:28:22
>> Okay, let's thank Bill. Amazing job.
00:28:26
Thanks, guys.
