Latest / Going Direct Conversations / #54 Saman Farid: America Doesn’t Have a Robot Problem. It Has a Deployment Problem.

Transcript

#54 Saman Farid: America Doesn’t Have a Robot Problem. It Has a Deployment Problem. · Going Direct Conversations

0:00
SPEAKER_00: Today I'm sitting down with Saman Farid, the founder and CEO of Formic.
0:03
Formic works with more than 20 robotics manufacturers and
0:05
provides robotics as a
0:06
SPEAKER_00: service to small and medium-sized manufacturers in the United States.
0:10
You typically might expect 6,000 hours is like a full,
0:14
relatively full production schedule.
0:16
SPEAKER_02: The average American factory runs about 2,000 to 3,000 hours.
0:20
SPEAKER_00: Now,
0:20
SPEAKER_00: the company has deployed robots in more than 100 factories,
0:23
so there's no one better to speak to than Saman about the
0:25
current state of affairs
0:26
SPEAKER_00: of robotics in the United States.
0:28
SPEAKER_00: No fluff, no headlines, just the reality from on the field.
0:31
We take care of all the engineering,
0:33
installation,
0:33
management,
0:34
maintenance of the robot.
0:35
SPEAKER_00: I hope you enjoyed the conversation.
0:37
SPEAKER_00: Remember to go direct.
0:45
SPEAKER_00: Hey, my friends, Jordan here.
0:46
SPEAKER_00: Just a few words before we hop into today's episode.
0:49
SPEAKER_00: First and foremost,
0:50
SPEAKER_00: a sincere thank you and appreciation for all of our early listeners,
0:54
SPEAKER_00: readers and supporters at Going Direct.
0:56
It's really what keeps me going to do this each and every week
0:58
because there's
0:59
clearly a community of folks out there looking to build a Going
1:02
Direct economy,
1:03
SPEAKER_00: which gets me super excited.
1:05
SPEAKER_00: Secondly, if you haven't subscribed, please do so.
1:07
SPEAKER_00: Our main channel and community is forming over on Substack.
1:11
SPEAKER_00: The link is in the show notes below.
1:13
And we'll also be putting out our weekly conversations at
1:15
YouTube,
1:16
Spotify,
1:17
and Apple Podcasts.
1:18
SPEAKER_00: I hope you all have a great day and remember to go direct.
1:22
SPEAKER_00: I will listen to Saman.
1:24
SPEAKER_00: I appreciate the time and I think, you know,
1:28
SPEAKER_00: I'll kind of frame this up and then we'll hop right into it.
1:31
So everyone that's watching or listening knows the whole idea
1:35
behind going direct
1:36
SPEAKER_00: is how do we build this kind of going direct economy,
1:38
SPEAKER_00: which is about decentralizing the real economy.
1:41
SPEAKER_00: How do we start making stuff closer to where we consume them?
1:45
How do we become more connected to the people that make them
1:47
and what's actually inside of them?
1:50
SPEAKER_00: And so obviously decentralized production is the core of this.
1:55
in order to do that realistically in the west we're going to
1:57
have to incorporate
2:01
much robotics and automation and create more decentralized
2:04
production capabilities
2:06
and so one of the biggest things there in kind of my my dream
2:11
for the future that i
2:12
want to help create is how do we really create a recreate a
2:17
robust network or much
2:18
SPEAKER_00: bigger network of small medium-sized manufacturers and
2:24
SPEAKER_00: You know,
2:24
SPEAKER_00: I followed your work for a while and a former's work for a while,
2:26
SPEAKER_00: which is why I reached out.
2:29
And I think the biggest thing that always drew me to what you
2:32
were doing before all
2:34
this momentum in the last four or five years in the robotics
2:37
space was you're
2:38
SPEAKER_00: really addressing this deployment gap that exists,
2:40
that has existed in the past and will continue to exist for
2:44
next generation
2:44
SPEAKER_00: robotics.
2:45
SPEAKER_00: So,
2:46
SPEAKER_00: and the last thing I'll say,
2:46
SPEAKER_00: then we'll hop into it,
2:48
SPEAKER_00: is,
2:48
SPEAKER_00: you know,
2:49
SPEAKER_00: most of folks that I've spoken with,
2:51
are largely in the lab and working on a lot of these next
2:53
generation technologies,
2:55
maybe working at pilot phase with a few customers versus you
2:58
and your team are
2:59
SPEAKER_00: literally in,
3:00
SPEAKER_00: is it a hundred factories?
3:01
SPEAKER_00: Is that right?
3:02
SPEAKER_00: That's right.
3:02
SPEAKER_00: Yeah.
3:03
SPEAKER_00: So,
3:03
SPEAKER_00: I mean,
3:03
SPEAKER_00: you,
3:03
SPEAKER_00: you have a very interesting perspective of what,
3:05
SPEAKER_00: what is the true game on the field.
3:07
SPEAKER_00: And so, yeah, so I'm excited to, to, to hop into it.
3:10
SPEAKER_00: All right.
3:11
So I want to start here with the,
3:12
with it,
3:12
with a tweet from a couple of months ago from you.
3:14
SPEAKER_00: All right.
3:14
SPEAKER_00: And then I'll let you take it from here.
3:16
SPEAKER_00: Sounds good.
3:18
SPEAKER_00: Okay.
3:18
SPEAKER_00: So I'm reading here.
3:19
So you wrote factory tours should be a mandatory part of middle
3:23
school education on up.
3:25
SPEAKER_00: Knowing where the stuff you use comes from is such an inspiring thing.
3:30
It builds an appreciation for labor,
3:31
design,
3:32
engineering,
3:32
scheduling,
3:33
and so much more.
3:35
SPEAKER_00: So that's like a very, very interesting perspective.
3:38
Love the perspective,
3:39
obviously,
3:39
but kind of tell me why,
3:40
why that's something you believe in.
3:41
SPEAKER_00: Um,
3:44
SPEAKER_02: So first of all, thanks for having me on.
3:46
SPEAKER_02: I'm excited to be here.
3:47
SPEAKER_02: I think in reference to that tweet,
3:51
there's so much complexity that goes into every single everyday
3:58
object that we
3:59
SPEAKER_02: don't appreciate,
4:00
SPEAKER_02: right?
4:00
I think every motor has magnets and windings and bearings and
4:05
all this amazing
4:05
SPEAKER_02: complexity controls.
4:07
SPEAKER_02: Every object that we use every day
4:11
is made with enormous dedication in engineering,
4:15
manufacturing,
4:16
production,
4:16
supply chain.
4:18
It's kind of scheduled for all those parts to arrive at the
4:20
right place at the
4:21
SPEAKER_02: right time,
4:21
SPEAKER_02: logistics to get it to the store near your house.
4:24
And you can go down the street and spend $5 to buy the most
4:27
complex object in the world.
4:28
SPEAKER_02: There's an interesting article.
4:29
SPEAKER_02: I can't remember the name of the guy,
4:31
SPEAKER_02: but this famous economist said basically,
4:34
there is not a single person that exists on planet Earth today
4:36
that knows end to
4:37
SPEAKER_02: end how even something as simple as a pencil is made.
4:42
SPEAKER_02: if you talk about the graphite that goes in the,
4:44
SPEAKER_02: in the lead,
4:45
SPEAKER_02: you talk about Milton Friedman,
4:47
SPEAKER_00: Milton Friedman.
4:47
SPEAKER_02: That's right.
4:48
SPEAKER_02: Yeah.
4:48
The paint,
4:49
the metal that goes on that route,
4:50
you know,
4:52
the,
4:52
the,
4:52
the erase,
4:53
right?
4:53
SPEAKER_02: Like there's so much complexity in even something as simple as a pencil.
4:58
SPEAKER_02: And obviously it,
5:00
SPEAKER_02: The amount of complexity that we have beyond that is enormous.
5:03
SPEAKER_02: So for me,
5:03
I think there's just such an important part of the education of
5:06
every person to
5:08
SPEAKER_02: think about,
5:08
SPEAKER_02: you know,
5:09
SPEAKER_02: what is this world we live in?
5:10
SPEAKER_02: Where did it come from?
5:10
SPEAKER_02: How did it get created?
5:11
What is the kind of enormous amount of cooperation that went
5:16
into the creation of
5:17
SPEAKER_02: this world that we're in,
5:18
SPEAKER_02: right?
5:18
SPEAKER_02: You know, I think it's easy to talk about competition.
5:20
SPEAKER_02: It's easy to talk about the other...
5:23
SPEAKER_02: But at the end of the day,
5:24
human civilization is this thing that is an iterative addition
5:28
to what other people
5:29
SPEAKER_02: have done before.
5:30
SPEAKER_02: And whether that's in science,
5:31
SPEAKER_02: whether that's in engineering,
5:32
SPEAKER_02: whether that's in manufacturing or production,
5:34
SPEAKER_02: you just see it so clearly in the world of production.
5:37
SPEAKER_02: So obviously a lot of passion around that.
5:40
And then the extension of that is if you actually go and see
5:44
real factories,
5:45
SPEAKER_02: the other thing that really stands out is
5:50
SPEAKER_02: they are, there's a lot of small, medium-sized factories, right?
5:55
So there's obviously the General Motors and Teslas and TSMCs of
5:58
the world.
6:00
But for any of them to be able to survive or produce the thing
6:04
that they want to
6:05
SPEAKER_02: produce,
6:05
SPEAKER_02: you know,
6:06
SPEAKER_02: for a single automotive plant,
6:08
SPEAKER_02: there's like 4,000 suppliers that make,
6:11
SPEAKER_02: you know,
6:11
the little buttons that open and close the window and the
6:14
windshield wipers and the
6:15
SPEAKER_02: lights and the
6:17
seats and every single component in that is made usually by
6:20
some third party,
6:21
SPEAKER_02: oftentimes with a smaller,
6:23
SPEAKER_02: medium-sized manufacturing business.
6:26
SPEAKER_02: And those businesses are so,
6:28
SPEAKER_02: you know,
6:30
SPEAKER_02: I think today,
6:31
SPEAKER_02: especially in the Western economy,
6:33
SPEAKER_02: really overlooked,
6:34
SPEAKER_02: right?
6:34
They seem kind of boring or stodgy or they seem outdated,
6:39
but they're like the spinal cord of
6:46
SPEAKER_02: modern civilization, right?
6:48
It's the guy who makes the nuts and bolts and the bearings and
6:50
the seals and all
6:51
SPEAKER_02: these things that end up in every complex product.
6:57
Oftentimes,
6:57
these are small family-owned businesses,
6:59
multiple generations.
7:00
In America,
7:01
it's common for them to be second,
7:03
third generation businesses.
7:05
A lot of the time,
7:06
they employ the majority of the people in the town that they're
7:10
in.
7:12
SPEAKER_02: And so they have a critical impact on the economy.
7:15
They have a critical impact on the flourishing of the people in
7:17
that city and town.
7:18
SPEAKER_02: And there's just this stack of things.
7:19
And if you start to kind of remove that stack,
7:22
the whole infrastructure starts to crumble.
7:26
SPEAKER_02: Yeah, a lot of passion around it.
7:27
SPEAKER_02: Young people should be aware of this.
7:28
SPEAKER_02: Young people should be involved in this, should be contributing to this.
7:32
SPEAKER_02: And I think it's enormously meaningful.
7:37
And if I have to hear one more teenager telling me they want to
7:40
be a YouTube
7:40
SPEAKER_02: influencer,
7:41
SPEAKER_02: I'm going to kill myself.
7:47
SPEAKER_00: I was around a podcast.
7:48
SPEAKER_00: No, I'm just kidding.
7:48
SPEAKER_00: No, I completely agree.
7:50
SPEAKER_00: Well, listen, I mean, so I'm from Detroit or just outside Detroit.
7:54
SPEAKER_00: So everything that you just said,
7:57
SPEAKER_00: and I think this is probably why it's just in my blood and it's just why,
7:59
SPEAKER_00: it's just what a lot of my work and investing is about.
8:02
SPEAKER_00: And you said this, which is like, people don't talk about this.
8:07
SPEAKER_00: And it was like,
8:09
SPEAKER_00: it's exactly true,
8:10
SPEAKER_00: but you have to grow up in a place to see it,
8:12
SPEAKER_00: to understand it,
8:13
SPEAKER_00: which is like,
8:14
the company like the companies that sponsored my little the
8:17
little league baseball
8:18
teams and the soccer teams and all the all the um many of the
8:21
parents that work
8:22
there like we're part of like making these physical things and
8:26
we're like literally
8:26
the fabric of the local community and there's just a level of
8:31
like respect and
8:32
SPEAKER_00: connection and realness to folks that like make stuff
8:37
anyway so i appreciate your perspective on that because i do
8:40
share your your um i
8:42
guess perspective and pat and passion for it and hence why i
8:45
think we both do the
8:46
work that we do so yeah absolutely i think um just to echo that
8:51
right like i think
8:52
when you grow up in a place where you see that evolution i i my
8:56
childhood was was
8:57
SPEAKER_02: mostly spent in china my parents moved to china when i was a little kid
9:00
SPEAKER_02: I lived there until I finished high school.
9:02
SPEAKER_02: So I got to, you know, my dad would take me in and out of factories.
9:05
SPEAKER_02: But also I got to see China go from,
9:07
SPEAKER_02: you know,
9:07
SPEAKER_02: Beijing when I moved there was,
9:08
SPEAKER_02: you know,
9:09
SPEAKER_02: dirt roads and donkeys carrying bricks,
9:11
SPEAKER_02: you know,
9:11
SPEAKER_02: down the road.
9:12
SPEAKER_02: And by the time I left, it's this kind of modern metropolis.
9:15
SPEAKER_02: And when I saw that evolution over that specific period of time,
9:19
it's this incredible evolution of like our ability to produce
9:24
the things that we
9:25
SPEAKER_02: need is directly correlated to the quality of life of the people.
9:31
SPEAKER_02: And just being able to see that transformation happen in real time,
9:34
I think whether it's a place like Detroit or whether it's an
9:39
economy that's in the
9:39
SPEAKER_02: process of developing and modernizing,
9:42
I think being able to see that firsthand just gives you such an
9:45
appreciation for
9:47
how much human ingenuity goes into making the things that we
9:50
can walk down to
9:51
SPEAKER_02: Walmart and grab.
9:52
SPEAKER_00: Well, I think it's interesting.
9:54
SPEAKER_00: I had that down.
9:55
SPEAKER_00: I thought it was very interesting.
9:56
SPEAKER_00: You grew up in China, so we'll definitely dig into that.
9:57
SPEAKER_00: But I think that point,
9:58
I think you had an interesting view of also what you see happen
10:01
when there's
10:03
SPEAKER_00: industrial policy,
10:06
SPEAKER_00: right,
10:06
SPEAKER_00: to actually make that a priority for the nation,
10:09
SPEAKER_00: which is like,
10:11
SPEAKER_00: well,
10:12
SPEAKER_00: we'll get into it.
10:12
SPEAKER_00: But OK, so I have a stat for you.
10:17
SPEAKER_00: This is a pretty interesting one.
10:18
SPEAKER_00: Let me know if you kind of have a similar from your research knowledge,
10:21
SPEAKER_00: but I thought this was very fascinating.
10:24
So in the US,
10:25
there's somewhere between approximately 240,
10:27
250,000 manufacturing firms.
10:28
SPEAKER_00: 65 to 75,000 have between 20 and 500 employees.
10:28
So that's about 30% of companies,
10:29
of manufacturing companies in the US are between 20 to 500.
10:33
SPEAKER_00: But 98% of manufacturing sites have fewer than 500 employees.
10:50
SPEAKER_00: Is that accurate based upon my research?
10:52
SPEAKER_00: Yeah, absolutely.
10:54
SPEAKER_00: Okay, so that is super fascinating.
10:56
SPEAKER_00: And so what I wanted to ask you was,
11:00
SPEAKER_00: you know,
11:00
SPEAKER_00: everyone says,
11:01
SPEAKER_00: and automation is just around the corner,
11:03
SPEAKER_00: I'm very excited about the future,
11:05
SPEAKER_00: but most of these factories aren't automated,
11:07
SPEAKER_00: obviously.
11:08
So based upon your experience,
11:10
kind of explain to me what the core bottleneck is.
11:17
SPEAKER_02: That's a great question.
11:19
SPEAKER_02: And like any good question, it has a very compound answer.
11:22
SPEAKER_02: Like it's a bunch of things.
11:24
SPEAKER_02: But I think at the simplest level,
11:29
SPEAKER_02: there is a operational complexity with modernizing small,
11:33
SPEAKER_02: medium-sized factories.
11:34
SPEAKER_02: And then there's a cultural complexity to it.
11:38
So I'll talk about the operational complexity first because
11:40
that's easier to
11:40
SPEAKER_02: understand,
11:41
SPEAKER_02: which is
11:43
These factories are typically experts in making their specific
11:47
product,
11:48
right?
11:48
SPEAKER_02: Let's say they're making dog food or chocolate chip cookies or,
11:52
SPEAKER_02: you know,
11:52
SPEAKER_02: door locks or handles or whatever,
11:55
SPEAKER_02: right?
11:55
They know everything there is to know about that product and
11:58
that industry.
12:00
To expect them to know everything there is to know about
12:02
chocolate chip cookies and
12:03
everything there is to know about robots is just like a really
12:07
tall order,
12:08
SPEAKER_02: right?
12:08
Especially as robots become exponentially more complicated as
12:11
they have over the
12:12
SPEAKER_02: last 20,
12:12
SPEAKER_02: 30 years.
12:14
SPEAKER_02: And I think when it comes to robots, that's an obvious thing to see.
12:19
SPEAKER_02: It's even more obvious when you look at, for example, servers, right?
12:22
And I think that that's an analogous technology that I think is
12:25
worth talking
12:26
SPEAKER_02: about,
12:26
SPEAKER_02: where most businesses in the world have decided that they are not
12:34
SPEAKER_02: best equipped to own and manage their own data centers.
12:37
SPEAKER_02: They said, you know what?
12:39
I will focus on being an accounting firm and I will just let
12:42
AWS manage the servers
12:44
SPEAKER_02: or whoever else.
12:47
I think it's easy for us to accept that with something like
12:50
with servers.
12:51
SPEAKER_02: But for some reason,
12:53
SPEAKER_02: I think the historical idea is like,
12:54
SPEAKER_02: oh,
12:54
SPEAKER_02: no,
12:54
SPEAKER_02: but you should manage your own robots.
12:56
You should build and figure out how to buy and manage and
12:59
deploy your own robots.
13:00
SPEAKER_02: And I think that it's just asking a lot from that.
13:05
SPEAKER_02: There's also a lot of,
13:06
SPEAKER_02: you know,
13:08
SPEAKER_02: in the spirit of kind of going direct,
13:10
there's also a lot of really poorly incentivized intermediaries
13:14
that try to sell
13:17
SPEAKER_02: you crap,
13:17
SPEAKER_02: basically,
13:18
SPEAKER_02: right?
13:18
SPEAKER_02: So if you're a factory,
13:20
typically the way that you would automate is you would go and
13:22
hire some expensive
13:23
automation consultant and they would come in and they would
13:26
spend you a few hundred
13:27
SPEAKER_02: thousand dollars of your money doing this kind of study and analysis.
13:31
SPEAKER_02: And then they would,
13:34
pick probably the most complex thing in your factory to
13:37
automate because then they
13:39
SPEAKER_02: can charge you for engineering services,
13:41
SPEAKER_02: right?
13:41
And then they'll charge you a bunch of engineering services to
13:43
design some super
13:45
SPEAKER_02: complex,
13:45
SPEAKER_02: super custom piece of machinery that only works for you.
13:49
SPEAKER_02: And then they would charge you a bunch of money for the hardware.
13:53
SPEAKER_02: They would upsell you.
13:54
SPEAKER_02: The more hardware you buy,
13:55
the more margins they collect because it's always done on a
13:58
cost plus basis.
14:00
SPEAKER_02: And then at the end of all of that, if you're lucky, it works.
14:06
SPEAKER_02: But very often, most likely, it doesn't work.
14:09
SPEAKER_02: And then there's a change order.
14:11
SPEAKER_02: You get a change order that says,
14:12
SPEAKER_02: oh,
14:12
SPEAKER_02: by the way,
14:13
SPEAKER_02: I forgot to mention,
14:15
SPEAKER_02: this doesn't work if it's raining.
14:16
SPEAKER_02: Or, oh, it doesn't work if the sunlight is on.
14:18
SPEAKER_02: And so all of a sudden,
14:18
you've already spent $500,000 or $1 million or $2 million on
14:20
this expensive piece
14:21
SPEAKER_02: of machinery.
14:26
SPEAKER_02: What are you going to do?
14:27
SPEAKER_02: Say no?
14:28
You know,
14:28
like you just got to keep paying until finally,
14:30
hopefully it works.
14:32
SPEAKER_02: So,
14:33
SPEAKER_02: you know,
14:33
it's this really horrible experience where there's this
14:37
information asymmetry
14:38
SPEAKER_02: between the buyer and the seller.
14:40
SPEAKER_02: The seller is heavily incentivized,
14:41
SPEAKER_02: actually heavily disincentivized towards efficiency and standardization.
14:45
SPEAKER_02: And so whether that's,
14:48
SPEAKER_02: you know,
14:49
SPEAKER_02: traditional robotics or modern robotics,
14:51
you still have the same bottleneck where factories don't know
14:53
what they need.
14:54
SPEAKER_02: They don't know how to buy it.
14:55
SPEAKER_02: They don't know how to manage it.
14:56
SPEAKER_02: So that's the kind of operational complexity.
15:00
There's also this kind of cultural complexity,
15:03
which is I see more commonly in the U.S.
15:07
SPEAKER_02: than I did when I lived in China.
15:10
SPEAKER_02: And this is probably a little bit controversial to say out loud.
15:12
You know,
15:14
most American manufacturing businesses are either third
15:18
generation manufacturers
15:21
like owned businesses or their private equity owned
15:24
manufacturing business.
15:26
Very,
15:26
very few manufacturing businesses in America today are founder
15:29
owned and operated.
15:31
SPEAKER_02: And as we know from the world of startups, right?
15:33
Like when you're a founder owned and operated company,
15:35
you're willing to take a lot of risks.
15:37
SPEAKER_02: When you're a third generation manager of the business,
15:41
SPEAKER_02: oftentimes you're more concerned with preservation of what you have
15:45
SPEAKER_02: than you are with really pushing the boundaries around capability.
15:48
It's very hard because,
15:49
you know,
15:50
you weren't there to build this thing from scratch.
15:52
SPEAKER_02: You kind of inherited it.
15:53
SPEAKER_02: It's difficult to go and make very risky changes.
15:58
SPEAKER_02: And private equity firms,
16:00
SPEAKER_02: let's not even talk about,
16:00
SPEAKER_02: they're even less likely to make aggressive,
16:03
SPEAKER_02: creative changes,
16:04
SPEAKER_02: right?
16:05
SPEAKER_02: And so...
16:06
There's this cultural complexity for manufacturing businesses
16:09
around trying new things.
16:11
It makes it really,
16:12
really difficult for them to adopt modern technologies and to
16:15
take risks.
16:17
SPEAKER_02: And if you compound that with operational complexity,
16:20
SPEAKER_02: which is like,
16:20
SPEAKER_02: I'm not even taking a risk on something I don't understand is even worse.
16:24
SPEAKER_02: So I think historically,
16:27
American manufacturing has really suffered because it's been
16:31
difficult for people
16:32
SPEAKER_02: to do that.
16:33
You compare that to China,
16:34
which in the last 30,
16:36
40 years has really grown a lot in industrial.
16:39
SPEAKER_02: It's super common for you to go to a Chinese factory.
16:41
SPEAKER_02: And in this room, they're making toilet paper.
16:44
SPEAKER_02: In that room, they're making engine parts.
16:46
SPEAKER_02: In the other room, they're making automotive components.
16:53
SPEAKER_02: There's no rhyme or reason.
16:54
SPEAKER_02: It may not be a good idea to do all those things on one roof.
16:58
SPEAKER_02: But...
16:59
Because they're founder-owned and operated,
17:01
they look for opportunity and then they jump on it.
17:05
SPEAKER_02: Obviously, that's changing as these businesses mature and evolve.
17:08
SPEAKER_02: But I think it continues to be
17:10
a point of strength for China that these businesses are willing
17:14
to take risks and try things.
17:15
And I think it's a challenge for American manufacturers because
17:18
it's difficult to do that.
17:21
SPEAKER_00: Interesting on that point.
17:22
So four or five months ago,
17:24
I did a podcast with Dylan Patel from Semi Analysis.
17:28
I don't know if you know that firm,
17:29
but they do really good research on kind of
17:34
SPEAKER_00: both the robotic supply chain and everything around AI infrastructure.
17:39
And we spoke deeply about how far had China really is ahead of
17:44
the West when it
17:45
SPEAKER_00: comes to robotic supply chains.
17:46
SPEAKER_00: Anyways,
17:47
SPEAKER_00: what he described,
17:49
SPEAKER_00: which I thought was a really good way of putting it,
17:50
SPEAKER_00: I want to get your perspective because you spent a lot of time in China,
17:53
SPEAKER_00: was he was just basically saying,
17:54
SPEAKER_00: he goes,
17:55
SPEAKER_00: the Chinese way,
17:56
SPEAKER_00: if you think about it,
17:56
SPEAKER_00: is...
17:58
SPEAKER_00: He goes,
17:58
SPEAKER_00: basically,
17:59
you have many companies you've never heard of that are listed
18:02
on the stock exchange
18:04
over in China who might be literally world leaders in their
18:07
respective industries.
18:09
SPEAKER_00: And because of how the system works there,
18:13
you have companies that are competing with each other to
18:16
continue to drive costs
18:18
SPEAKER_00: down.
18:18
But then what they would rather do to maintain their market
18:21
share is when they
18:22
SPEAKER_00: drive that cost down,
18:25
SPEAKER_00: They don't keep that margin for themselves.
18:27
They are passing that margin on to their customers and to the
18:30
market to allow the
18:31
SPEAKER_00: whole supply chain to continue to function,
18:33
SPEAKER_00: right?
18:33
SPEAKER_00: Because everyone downstream needs their margin to make it work.
18:37
Versus the American way is like,
18:38
how can you get the highest margins possible?
18:41
Anything that you can do to bring costs down,
18:43
you're not passing it on to your customer.
18:44
SPEAKER_00: You want it all for yourself.
18:47
SPEAKER_00: And so I thought that was an interesting perspective.
18:49
I mean,
18:49
I don't think it's out of the goodness of their heart that they
18:52
pass the margin down.
18:53
SPEAKER_02: I think it's because of this extreme competition, right?
18:55
SPEAKER_02: It's like, if you don't do it, the other guy will.
18:58
And I think that it comes from a much more vibrant ecosystem
19:03
where there are
19:04
SPEAKER_02: hundreds of competitors for kind of any small product.
19:07
And so I think there's just,
19:09
you know,
19:11
in the kind of fitness function of natural evolution,
19:14
you end up with much more fit companies because they have to
19:17
fight off hordes of
19:19
SPEAKER_02: hundreds or thousands of competitors.
19:23
SPEAKER_02: I think the tech ecosystem in the US is kind of like that,
19:27
SPEAKER_02: where a lot of tech companies have historically had to really,
19:30
SPEAKER_02: really fight to survive because there's a lot of competition.
19:33
And so they ended up being quite fit from a kind of resilience
19:37
perspective.
19:38
manufacturing businesses in America,
19:41
if they're still in business in the U.S.
19:43
in 2026,
19:44
it's likely that they have some kind of monopolistic advantage,
19:49
right?
19:49
SPEAKER_02: Either geography or customer relationships or,
19:51
SPEAKER_02: you know,
19:52
SPEAKER_02: buying relationship with the government or whatever.
19:54
SPEAKER_02: Like there's something that allows them to stay in business,
19:56
SPEAKER_02: which is great for them,
19:58
SPEAKER_02: but sometimes leads to less,
20:00
SPEAKER_02: like less intense competition,
20:02
SPEAKER_02: right?
20:02
SPEAKER_02: Or there's oftentimes duopoly or, you know, there's three options.
20:05
SPEAKER_02: And so you don't get the same level of competition that you get in China.
20:11
SPEAKER_00: All right, I want to move over to the job debate.
20:14
SPEAKER_00: Not to,
20:15
SPEAKER_00: we could sit here and pontificate on,
20:18
SPEAKER_00: you know,
20:19
SPEAKER_00: whatever,
20:19
SPEAKER_00: in whatever future we think is going to happen.
20:23
I think what I wanted to get a better perspective on really
20:26
kind of the current
20:27
SPEAKER_00: state and kind of in the short medium term of what you're seeing that,
20:30
SPEAKER_00: you know,
20:30
SPEAKER_00: a lot of folks that are like robots are going to be replacing jobs.
20:34
SPEAKER_00: And the other side of the narrative is,
20:36
SPEAKER_00: well,
20:36
we can't find work enough workers to work in the factories that
20:40
they're actually
20:40
SPEAKER_00: filling jobs.
20:41
And so can you just kind of kind of tell us what you see is
20:45
kind of the game on the field?
20:47
SPEAKER_00: as it relates to jobs and yeah,
20:50
whether it's on the scarcity side and robotics are truly
20:55
filling that gap or kind
20:56
SPEAKER_00: of where that stands to be.
20:57
SPEAKER_00: It'd be interesting to get your perspective.
20:59
SPEAKER_02: So I'll give you an anecdote and then I'll give you my full answer.
21:02
SPEAKER_02: I was in a factory three weeks ago, one of our customers in New Jersey.
21:08
SPEAKER_02: And they make all kinds of products.
21:11
SPEAKER_02: I won't go into the details of what they make.
21:14
SPEAKER_02: The factory has about 165 headcount.
21:15
SPEAKER_02: So not huge.
21:16
SPEAKER_02: 165 employees there.
21:22
SPEAKER_02: In the calendar year 2025,
21:25
SPEAKER_02: just to keep those roles filled,
21:27
SPEAKER_02: they had to hire 250 people just to stay in business.
21:34
SPEAKER_02: So it's like more than a 200% headcount attrition rate of labor.
21:43
SPEAKER_02: So...
21:45
I think that gives you a little bit of a sense of how dire the
21:47
situation is.
21:49
SPEAKER_02: For a lot of factories in America,
21:51
SPEAKER_02: labor cost is obviously a bit of a problem,
21:54
SPEAKER_02: but more than that,
21:55
SPEAKER_02: labor availability is a giant,
21:57
SPEAKER_02: giant problem.
21:59
And these factories are not able to take on more business
22:03
because they don't have
22:04
SPEAKER_02: people who can run this factory for a second shift or a third shift.
22:09
SPEAKER_02: And so if you're a factory and you're running one shift a day,
22:13
SPEAKER_02: just by simple math,
22:15
SPEAKER_02: two thirds of the available hours of the year,
22:18
SPEAKER_02: your factory is sitting idle and collecting dust.
22:20
SPEAKER_02: You've invested a few hundred million dollars usually in facilities,
22:25
SPEAKER_02: running electrical,
22:26
SPEAKER_02: forklifts,
22:27
SPEAKER_02: machinery,
22:28
SPEAKER_02: conveyors,
22:29
SPEAKER_02: all this stuff.
22:30
SPEAKER_02: And then you leave it idle for two thirds of the hours.
22:34
SPEAKER_02: It's just kind of inconceivable, right?
22:36
SPEAKER_02: And so I think the natural answer to your question is like,
22:40
SPEAKER_02: if you tell any factory owner that,
22:42
SPEAKER_02: hey,
22:42
SPEAKER_02: overnight you can have three times more available headcount,
22:46
SPEAKER_02: would you use them?
22:48
SPEAKER_02: There isn't a single factory that would say no, right?
22:49
SPEAKER_02: Like they all want more people.
22:53
SPEAKER_02: You know,
22:53
SPEAKER_02: in the lack of our ability to suddenly magically create human beings,
22:57
SPEAKER_02: we have to create robots,
22:58
SPEAKER_02: right?
22:58
SPEAKER_02: That can go and fill that headcount.
23:02
SPEAKER_02: And, you know,
23:04
SPEAKER_02: It sounds self-serving coming from a robotics guy,
23:07
but there's a lot of evidence that shows factories that
23:11
automate and have robots
23:12
SPEAKER_02: actually end up having much higher paid workers,
23:16
SPEAKER_02: much happier workers,
23:18
SPEAKER_02: much...
23:19
SPEAKER_02: less workplace injuries and,
23:22
SPEAKER_02: you know,
23:23
all kinds of benefits because they're able to kind of take a
23:26
lot of those menial,
23:28
SPEAKER_02: repetitive tasks and automate.
23:31
I'll end on one last anecdote,
23:33
right,
23:33
which is I was,
23:34
again,
23:34
at one of our customer sites.
23:36
SPEAKER_02: few months ago that makes, they make like body lotion, right?
23:42
And there was literally a row of 35 people standing there and
23:49
eight hours a day,
23:49
SPEAKER_02: their job is to screw the bottle caps onto the
23:54
SPEAKER_02: these bottles of lotion.
23:56
SPEAKER_02: And,
23:56
SPEAKER_02: you know,
23:56
they would do three of them and they would shake their hand
23:58
because it's so painful
23:59
SPEAKER_02: and they would do three more.
24:01
And I would just ask anybody who asked,
24:02
you know,
24:03
wonders,
24:03
what is it all?
24:04
SPEAKER_02: Is that really like the best use of a human being?
24:07
Like all the intelligence and creativity and emotion and
24:11
they're going to stand
24:12
SPEAKER_02: there twisting bottle caps.
24:13
SPEAKER_02: Like it's just, we can do better as a society.
24:19
SPEAKER_00: Well,
24:20
SPEAKER_00: you know,
24:20
SPEAKER_00: it's interesting that on that,
24:21
SPEAKER_00: I mean,
24:22
SPEAKER_00: I'm clearly on one side of that debate,
24:24
SPEAKER_00: which is because it's just so obvious for me.
24:27
But I also find it fascinating when people have this other
24:29
perspective around this
24:30
SPEAKER_00: idea of replacing the jobs.
24:32
SPEAKER_00: Many people I speak to about that,
24:33
that are very strong on that have never been in a factory or
24:39
actually understand.
24:40
SPEAKER_02: If you've seen these jobs, you know, nobody wants them.
24:42
SPEAKER_02: Yeah.
24:43
Yeah.
24:47
SPEAKER_00: Well,
24:47
this other idea of actually understanding when you have
24:51
customers and you need to
24:52
SPEAKER_00: have as fast kind of production cycle times as possible and then dealing,
24:57
SPEAKER_00: like you said,
24:58
SPEAKER_00: with people not showing up to work,
24:59
SPEAKER_00: inclement weather,
25:00
SPEAKER_00: all these things that it's just really burdened.
25:03
SPEAKER_00: And the reality of it is your fixed costs are your fixed costs.
25:06
SPEAKER_00: You're paying for everything you have in there no matter what.
25:09
SPEAKER_00: And I just don't think people understand that.
25:11
SPEAKER_00: I don't know.
25:12
SPEAKER_00: For some reason, they don't get it.
25:12
SPEAKER_00: Yeah.
25:14
SPEAKER_02: Yeah.
25:14
SPEAKER_02: And I mean,
25:15
SPEAKER_02: maybe just to extend that point a little bit more,
25:17
SPEAKER_02: you know,
25:18
there's a lot of talk today in the West and kind of in the US
25:24
in particular around,
25:26
SPEAKER_02: hey,
25:26
SPEAKER_02: we need to build more production capacity.
25:28
SPEAKER_02: Let's go build more factories, right?
25:30
SPEAKER_02: Like we need to reshore manufacturing.
25:31
SPEAKER_02: We need to build more factories.
25:33
And,
25:34
you know,
25:35
there are obviously contexts where that makes a ton of sense,
25:37
right?
25:37
SPEAKER_02: We need to build chip fabs.
25:39
SPEAKER_02: We need to build new battery production lines and things like that.
25:43
But the easiest way to get more production capacity is to
25:47
increase the utilization
25:48
SPEAKER_02: of the 250,000 factories that we already have.
25:52
How many factories can you build in a year in a country versus
25:55
how easy would it be
25:56
SPEAKER_02: to increase your utilization of existing factories by 30%?
25:59
You'd get much more benefit out of doing that from an overall
26:03
capacity perspective.
26:06
SPEAKER_00: Do you have any data?
26:07
SPEAKER_00: Actually,
26:08
it'd be an interesting thing that I haven't seen data on the
26:11
overall utilization
26:12
SPEAKER_00: rate of US manufacturing facilities.
26:15
SPEAKER_00: Do you have any idea what that kind of is broadly?
26:18
SPEAKER_00: It averages about 3,000 hours a year.
26:22
SPEAKER_02: out of 8,600 possible production hours.
26:25
SPEAKER_02: So 8,600 is if you used every minute of the year, right?
26:29
SPEAKER_02: If you take away weekends and holidays and things like that,
26:32
SPEAKER_02: you typically might expect 6,000 hours is like a full,
26:37
SPEAKER_02: relatively full production schedule.
26:40
SPEAKER_02: The average American factory runs about 2,000 to 3,000 hours.
26:46
SPEAKER_00: Wow.
26:46
SPEAKER_00: So basically half, if you're being, that's even being like,
26:52
SPEAKER_00: Gracious, okay, interesting.
26:56
SPEAKER_00: Okay, so let's get over to robotics as a service.
27:00
SPEAKER_00: So I'm just gonna give the formal description from the website,
27:04
SPEAKER_00: then I'll let you take it here and we'll dig into it,
27:06
SPEAKER_00: so I'll dig into it.
27:07
So Formic delivers fully supported robotic automation at a low
27:11
hourly rate,
27:12
including deploying the system and providing continuous
27:15
monitoring and maintenance
27:16
SPEAKER_00: throughout the engagement to ensure success.
27:19
SPEAKER_00: So the concept of as a service,
27:22
SPEAKER_00: obviously most folks are familiar with it,
27:23
SPEAKER_00: but it's kind of over,
27:24
SPEAKER_00: it's used a lot.
27:25
SPEAKER_00: So can you just break down at a high level,
27:27
SPEAKER_00: obviously that explains it,
27:28
SPEAKER_00: but a little bit more about actually what Formic does?
27:31
SPEAKER_02: Yeah.
27:32
SPEAKER_02: So what you described is the product that we offer, right?
27:36
SPEAKER_02: Which is the easiest way to get robots into your factory, right?
27:42
SPEAKER_02: So they call Formic, we show up with the robot.
27:45
SPEAKER_02: And that's literally all they have to do.
27:47
We take care of all the engineering,
27:48
installation,
27:49
management,
27:49
maintenance of the robot.
27:52
SPEAKER_02: And what we believe is that
27:56
That is also the best way for factories to automate because
27:59
number one,
28:00
it aligns incentives.
28:01
SPEAKER_02: We talked a little bit about incentives earlier.
28:03
Number two,
28:04
it's the only way for factories to get a robot that has a SLA
28:08
attached to it,
28:09
right?
28:09
SPEAKER_02: Like there's some performance commitment associated with that system,
28:14
SPEAKER_02: which is typically not the case with traditional robots and automation.
28:18
SPEAKER_02: So...
28:19
SPEAKER_02: But to answer the question more fully around what does Formic do,
28:23
SPEAKER_02: in order to deliver on that promise of easy automation,
28:28
SPEAKER_02: easy robots,
28:29
SPEAKER_02: we had to build a ton of software to evaluate the factory,
28:33
SPEAKER_02: figure out which robot is the right one for the job,
28:35
SPEAKER_02: spec out the hardware,
28:36
SPEAKER_02: use the software to automatically program the robot,
28:39
use machine learning models to make it resistant to all the
28:42
different deviations
28:44
SPEAKER_02: that might occur in that manufacturing environment.
28:45
SPEAKER_02: build software for monitoring and maintenance,
28:48
SPEAKER_02: remote teleoperation,
28:49
SPEAKER_02: error handling,
28:51
SPEAKER_02: error recovery,
28:53
component monitoring for failures and dynamic redundancies and
28:59
things like that.
29:00
So we've had to build a ton of software infrastructure as well
29:05
as a ton of
29:05
SPEAKER_02: operational infrastructure.
29:06
We've had to build a nationwide network of installation crews,
29:09
maintenance crews,
29:12
SPEAKER_02: supply chain of hardware.
29:14
And then we also have to build a bunch of financial
29:16
infrastructure to kind of fund
29:17
SPEAKER_02: and build these things without digging a giant hole for ourselves.
29:21
It takes a lot of behind the scenes work to be able to offer
29:25
this relatively simple
29:26
SPEAKER_02: product,
29:26
SPEAKER_02: which is,
29:27
SPEAKER_02: hey,
29:27
SPEAKER_02: show up and we'll drop in a robot and it'll work.
29:31
SPEAKER_02: But that is kind of the point, right?
29:32
SPEAKER_02: Which is like the end customer is not going to do that on their own.
29:35
SPEAKER_02: The chocolate chip cookie factory in an example.
29:38
SPEAKER_02: And so you need somebody to basically say,
29:41
SPEAKER_02: I'm going to take all that complexity,
29:42
SPEAKER_02: I'm going to abstract it away for you and make it very easy to consume.
29:46
SPEAKER_02: And that is what's allowed us to get into,
29:48
SPEAKER_02: to your point,
29:49
100 plus factories in the last three and a half years that
29:52
we've been in operation.
29:55
SPEAKER_00: Great.
29:55
SPEAKER_00: And then, so who owns the robot?
29:57
SPEAKER_00: How does that work?
29:59
SPEAKER_02: We own the robot.
30:00
SPEAKER_02: Yeah.
30:00
SPEAKER_02: So we are the owner operator of the fleet of robots.
30:05
SPEAKER_00: Okay, great.
30:05
SPEAKER_00: So if I have a bottle, I'm just making this up, right?
30:07
SPEAKER_00: I have a bottling facility to keep it simple.
30:09
SPEAKER_00: I have a bottling facility.
30:10
SPEAKER_00: I work with you to,
30:11
SPEAKER_00: I'm looking for some automation around some functions around,
30:15
SPEAKER_00: around different,
30:16
SPEAKER_00: different bottling.
30:17
SPEAKER_00: You come in and you basically do,
30:20
SPEAKER_00: you do an assessment and you analyze and then,
30:22
SPEAKER_00: and then you say to me what,
30:23
SPEAKER_00: like I'll deliver you this robot,
30:25
SPEAKER_00: take care of everything.
30:26
SPEAKER_00: And I'm charging you X number of dollars per hour that it's,
30:30
SPEAKER_00: you're using it or like per,
30:32
SPEAKER_00: is it a per month charge or how does that actually work?
30:35
SPEAKER_02: Yeah.
30:35
SPEAKER_02: I mean,
30:36
the details of the contract vary a little bit depending on the
30:38
type of product,
30:38
SPEAKER_02: but sometimes it's a monthly rate,
30:40
SPEAKER_02: sometimes it's hourly rate.
30:40
SPEAKER_02: Okay.
30:42
But yes,
30:42
I mean,
30:42
basically what you described is correct,
30:44
which is our sales team shows up.
30:47
SPEAKER_02: We use a LiDAR scanner to scan the entire factory.
30:51
SPEAKER_02: We generate a 3D model of that factory.
30:54
SPEAKER_02: We also have,
30:55
SPEAKER_02: during that scan,
30:55
we have a normal camera that observes all the human work that's
30:58
being done.
30:59
SPEAKER_02: We quantify all those tasks and
31:03
say,
31:04
okay,
31:04
hey,
31:04
this task is being done today by a human at a rate of five
31:07
items a minute.
31:08
SPEAKER_02: That task is being done at 25 items a minute.
31:11
SPEAKER_02: The boxes look like they're roughly this size and this dimension.
31:14
SPEAKER_02: The bottles are roughly this weight.
31:17
SPEAKER_02: We have to kind of guesstimate a lot of things,
31:20
but information collection about the factory and about the task
31:25
is step zero for
31:29
SPEAKER_02: building and deploying a robot.
31:33
That's something that I think most robotics companies totally
31:36
miss,
31:36
right?
31:37
SPEAKER_02: They say,
31:37
SPEAKER_02: oh,
31:37
SPEAKER_02: if somebody comes to me and tells me exactly what they need,
31:40
SPEAKER_02: then I can deliver my humanoid or I can deliver my forklift and whatever,
31:44
SPEAKER_02: and it'll work.
31:45
SPEAKER_02: The problem is the end user doesn't know what they need.
31:48
SPEAKER_02: They don't have usually most of that data that I just described.
31:51
SPEAKER_02: The reason we have to collect it is when we started,
31:53
SPEAKER_02: we would just ask the customer,
31:54
SPEAKER_02: hey,
31:54
SPEAKER_02: what tasks do you do?
31:55
SPEAKER_02: What rate do you do it at?
31:56
SPEAKER_02: What's the size and weight of your product?
31:58
SPEAKER_02: And they don't know.
32:00
SPEAKER_02: You know, that's great.
32:01
So like,
32:02
because they don't know,
32:02
like,
32:03
you know,
32:03
what are you going to do?
32:05
So we realized we basically had to build a bunch of software
32:08
for analyzing the
32:09
SPEAKER_02: product,
32:09
SPEAKER_02: analyzing the facility,
32:11
SPEAKER_02: analyzing the tasks.
32:13
SPEAKER_02: And then to your point, we have a matching algorithm.
32:16
We have a bunch of robots in our library that are kind of
32:19
standard hardware solutions.
32:21
SPEAKER_02: The software will pick one of them.
32:22
It'll run that robot in simulation in that physical footprint
32:25
with that specific
32:26
product with that specific rate and validate that it can meet
32:29
those performance
32:30
SPEAKER_02: demands.
32:32
That allows our sales guy who's on site to generate a quote
32:35
within five minutes and
32:35
SPEAKER_02: says,
32:36
SPEAKER_02: hey,
32:36
SPEAKER_02: I have a robot.
32:36
SPEAKER_02: It's roughly this size.
32:38
SPEAKER_02: It's roughly this footprint.
32:39
SPEAKER_02: It needs this kind of power line dropped.
32:41
SPEAKER_02: And it's going to be $5,000 a month or $25,000 a month or $500 a month,
32:42
SPEAKER_02: depending on the type of thing.
32:48
SPEAKER_02: And if the customer says,
32:50
SPEAKER_02: hey,
32:51
SPEAKER_02: I want one,
32:51
SPEAKER_02: you know,
32:52
SPEAKER_02: a few weeks later,
32:52
SPEAKER_02: we show up with that robot.
32:53
SPEAKER_02: We typically within four hours, it's fully installed and operational.
32:58
We then stay usually for a day or two and do training of the
33:01
customer to ensure
33:02
SPEAKER_02: that they can know how to run it.
33:04
SPEAKER_02: And they're off to the races.
33:07
And then we take care of all the maintenance and spare parts
33:09
and things like that.
33:10
SPEAKER_00: And that's all part of that monthly or hourly fee?
33:14
SPEAKER_00: Okay.
33:14
SPEAKER_00: That's right.
33:15
SPEAKER_00: How many manufacturers are you working with?
33:17
Like what does that catalog kind of generally look like from a
33:19
quantity standpoint?
33:20
SPEAKER_02: Yeah, it's 100 plus manufacturers.
33:23
SPEAKER_00: Oh, wow.
33:23
SPEAKER_00: Okay.
33:24
And so when you think about kind of building all that kind of
33:26
middleware or kind of
33:28
SPEAKER_00: software to kind of make everything work,
33:31
SPEAKER_00: Are all of these,
33:33
SPEAKER_00: well,
33:34
SPEAKER_00: all of them are already kind of open for that anyway,
33:36
SPEAKER_00: so that's how they have to do it today,
33:37
SPEAKER_00: right?
33:37
SPEAKER_00: Working with systems integrators to integrate into their core,
33:39
SPEAKER_00: the core basically operating system of the factory,
33:42
SPEAKER_00: correct?
33:43
Yeah,
33:44
sorry,
33:45
you were talking about how many manufacturers of robotics do we
33:47
work with?
33:48
SPEAKER_02: That's a smaller number, so I thought you meant customers.
33:51
SPEAKER_02: On the robotics side,
33:51
SPEAKER_02: you know,
33:52
SPEAKER_02: Ralph,
33:52
SPEAKER_02: today we have like maybe 20 or so types of robots in our catalog.
33:57
SPEAKER_02: We're adding new tasks kind of on a constant basis.
34:01
SPEAKER_02: To answer your question,
34:02
SPEAKER_02: yes,
34:02
partially there are some open protocols that we can use for
34:06
communication with the
34:07
SPEAKER_02: different kinds of hardware.
34:10
there's a lot of work that needs to be done to make them
34:15
performant at the rates
34:17
SPEAKER_02: and capabilities that we need.
34:18
SPEAKER_02: So for example,
34:19
SPEAKER_02: like you can buy a robot arm from KUKA,
34:22
SPEAKER_02: but the arm by itself doesn't do anything.
34:25
SPEAKER_02: So you need like a work cell that's built and designed around it.
34:29
SPEAKER_02: You need end-of-arm tooling,
34:31
SPEAKER_02: you need sensing,
34:31
SPEAKER_02: you need cameras,
34:32
SPEAKER_02: you need safety scanners,
34:33
SPEAKER_02: you need emergency stop buttons,
34:35
SPEAKER_02: a whole bunch of different things like that.
34:36
SPEAKER_02: You also need communication.
34:37
So you need to kind of build that full work cell and have that
34:40
be a standard.
34:41
SPEAKER_02: On the software side,
34:42
SPEAKER_02: the KUKA robot may come with its own path planning capabilities,
34:48
but generally it's very limited in terms of the types of
34:50
dynamic tasks you can ask
34:52
SPEAKER_02: it to do.
34:53
SPEAKER_02: And so you still have to integrate it with your cameras,
34:56
SPEAKER_02: your vision system,
34:58
SPEAKER_02: whatever machine learning models you have,
34:59
SPEAKER_02: whatever task management system you have.
35:02
SPEAKER_02: So for example,
35:04
SPEAKER_02: this factory might want me to use that robot to do one type of task,
35:07
a different factory might need me to use the same robot to do a
35:09
different task.
35:09
So we have to build kind of the operating system layer that
35:14
runs on top of the
35:14
SPEAKER_02: robot and manages all the kind of components to get the job done.
35:18
SPEAKER_02: So there's quite a bit of development work.
35:21
involved in you know adding a new category of task but each
35:26
incremental task gets
35:27
easier and easier because relatively standard library of things
35:31
got it and is there
35:32
kind of a core when you think is there kind of a core task or
35:35
core type of customer
35:36
that that you work with or is it pretty wide ranging yeah so
35:40
today uh the five main
35:42
SPEAKER_02: tasks that we do uh are um
35:46
We do a lot of palletizing,
35:48
which is kind of the end of the production line.
35:50
SPEAKER_02: We do case packing, which is putting items into the production line.
35:53
We do autonomous forklifts and AMRs for kind of moving product
35:57
around the factory.
35:59
SPEAKER_02: We do a lot of machine tending for mills and lathes.
36:02
SPEAKER_02: So there's a lot of...
36:04
um uh metal fabrication facilities that that that need there's
36:10
kind of cnc machines
36:11
managed so we have robots load and unload those things and the
36:15
last one that we do
36:15
is plastic injection mold pending again so lots of plastic
36:19
injection molding
36:20
machines are underutilized you put a robot in front of it and
36:24
you can get triple of
36:25
SPEAKER_02: throughput out of that machine
36:26
cool all right let's move over to the financing side because i
36:29
thought this was
36:30
like um i think it's really important because i was like
36:32
talking to folks um i mean
36:34
you essentially have kind of three businesses in one which is
36:36
pretty interesting
36:38
SPEAKER_00: i'm sure nice and uh nice and challenging
36:42
So on the,
36:43
actually,
36:43
sorry,
36:44
before I hop to the financing,
36:45
I had one question,
36:46
actually.
36:47
As you guys thought about building out your actual integration
36:52
crews and support
36:53
SPEAKER_00: crews and maintenance crews,
36:54
SPEAKER_00: are they all Formic employees?
36:57
And did you kind of do it from a geographic,
36:58
like initially from a specific geographical area?
37:00
How did you think about building that fleet out of,
37:03
you know,
37:05
technicians and maintenance folks?
37:07
yeah um so we did take a geographical approach right what we
37:12
tried to do is build
37:12
geographical density region by region once we built a certain
37:16
amount of density we
37:18
were able to kind of fund with the kind of robots that are in a
37:21
region the field
37:23
service team the deployment team and engineering teams that you
37:25
need for that
37:25
system as we've gotten bigger that's become less and less of a
37:30
concern because now
37:30
we basically have coverage over the whole country and we can
37:32
basically add new
37:33
robots to the fleet without necessarily incrementally adding a
37:36
lot more headcount
37:37
A lot of the work there is also building tooling to give those
37:40
people superpowers.
37:41
So how do you allow a field service engineer to manage 100
37:44
robots by themselves instead of 10?
37:46
And that requires obviously a bunch of software infrastructure
37:49
and tooling.
37:52
SPEAKER_02: But to your point,
37:52
SPEAKER_02: you know,
37:52
we are building kind of three companies in one and managing
37:56
like an operational
37:57
field service team is not the kind of thing that a typical
37:58
software company would
37:59
SPEAKER_02: have had to do.
38:01
SPEAKER_02: But,
38:02
SPEAKER_02: you know,
38:03
SPEAKER_02: again,
38:03
going back to our philosophy around like this is the this is
38:06
what's necessary.
38:08
SPEAKER_02: to make robots consumable by the end user, right?
38:11
SPEAKER_02: Like that,
38:12
SPEAKER_02: the biggest problem around robotics is adoption,
38:15
SPEAKER_02: not around making the robot do a new trick.
38:17
And so like,
38:17
to your point,
38:18
you know,
38:18
we need,
38:18
we needed to solve a bunch of these kinds.
38:20
SPEAKER_02: So yeah, geographical density is a big part of it.
38:23
SPEAKER_02: Partnerships is part of it.
38:24
SPEAKER_02: Although,
38:25
SPEAKER_02: you know,
38:26
SPEAKER_02: partnerships is difficult to do when you're a young startup,
38:28
because you can't get the performance quality from a third
38:31
party that you can from
38:32
SPEAKER_02: yourself.
38:34
You also can't get adoption of tooling and things like that as
38:36
well by third parties.
38:38
SPEAKER_02: But as we've gotten bigger and bigger,
38:40
we've actually been able to force much better kind of
38:43
performance across third
38:45
SPEAKER_02: parties.
38:47
SPEAKER_02: And then there's also other kinds of logistical things, right?
38:50
SPEAKER_02: Like where do you store the spare parts?
38:52
SPEAKER_02: How do you ship things around?
38:53
SPEAKER_02: Do you have trucks or not?
38:54
SPEAKER_02: You have to kind of build all that stuff in mind.
38:57
SPEAKER_00: Got it.
38:57
SPEAKER_00: Great.
38:58
SPEAKER_00: All right.
38:58
SPEAKER_00: So let's move over to financing.
38:59
SPEAKER_00: So obviously the whole as a service model, it removes the
39:04
SPEAKER_00: was the upfront cost for the customer.
39:07
SPEAKER_00: And you kind of,
39:10
SPEAKER_00: I heard some quotes kind of,
39:11
is your kind of whole perspective was this was the biggest
39:14
barrier to automation
39:15
SPEAKER_00: adoption is the financing component.
39:17
You know,
39:17
now that you're five or six years in,
39:19
do you still believe that to be true?
39:21
SPEAKER_00: Or is it a mix?
39:22
SPEAKER_00: It's really a mix of things like, yeah, an equal footing.
39:25
SPEAKER_02: Yeah.
39:26
SPEAKER_02: So I want to clarify that I don't think financing by itself is
39:32
I think most factories can go and get a lease on a forklift or
39:37
they can get a lease
39:37
SPEAKER_02: on a CNC machine.
39:40
SPEAKER_02: That part is actually quite easy.
39:42
SPEAKER_02: The hard thing is,
39:46
balancing the risk between what the end user is responsible for
39:52
and what are you
39:53
responsible for and and financing is a way to relatively
39:58
cleanly bundle together a
39:59
bunch of complexity into a simple monthly payment right so now
40:03
the customer pays me
40:04
a few thousand dollars a month that includes hardware software
40:07
maintenance spare
40:08
SPEAKER_02: parts engineering installation and all these other things
40:10
SPEAKER_02: And so in order for us to be able to do that sustainably as a startup,
40:15
we had to solve a bunch of complexity behind the scenes or how
40:18
do you effectively
40:20
SPEAKER_02: finance large scale assets.
40:24
SPEAKER_02: Luckily, these are all revenue generating assets.
40:27
SPEAKER_02: So with enough performance history,
40:29
SPEAKER_02: we were able to prove to all of these different financing parties,
40:33
SPEAKER_02: hey,
40:33
SPEAKER_02: look,
40:34
SPEAKER_02: a robot typically has a return profile of X.
40:37
SPEAKER_02: It typically has a retention rate of Y.
40:40
SPEAKER_02: It typically has a renewal probability of Z.
40:43
SPEAKER_02: And with all of this,
40:44
SPEAKER_02: it has a redeployment and residual value of whatever.
40:48
SPEAKER_02: And you have to kind of build that...
40:51
case and build that performance history to be able to get
40:56
access to institutional
41:00
SPEAKER_02: private credit,
41:00
SPEAKER_02: basically,
41:01
SPEAKER_02: right?
41:02
And that is the way that,
41:04
you know,
41:05
historically,
41:05
you know,
41:06
data centers were financed.
41:07
I mean,
41:08
today even still,
41:08
like,
41:09
all kinds of infrastructure projects are financed this way.
41:13
SPEAKER_02: but you have to have a certain level of,
41:17
SPEAKER_02: like basically all of your kind of paper has to be finance grade,
41:20
SPEAKER_02: right?
41:21
SPEAKER_02: And so like that,
41:21
there's a lot of complexity around how do you do that and how
41:23
do you have backup
41:24
SPEAKER_02: servicing and how you manage taxes and how you manage depreciation,
41:27
how you manage recourse versus non-recourse and do
41:30
securitization of the assets or
41:31
SPEAKER_02: not.
41:32
SPEAKER_02: Like there's a bunch of things you have to solve
41:34
SPEAKER_02: to get access to this type of capital at scale.
41:37
SPEAKER_02: And once we solve that,
41:38
that allows our business to operate in a really kind of cash
41:41
efficient way.
41:43
And we're still able to kind of abstract all that complexity
41:46
away from the end user.
41:49
SPEAKER_00: Right,
41:49
SPEAKER_00: because for my recollection,
41:51
you actually have raised kind of institutional capital on the
41:54
debt side to handle
41:55
SPEAKER_00: that,
41:55
SPEAKER_00: like you said,
41:55
SPEAKER_00: capital expense to buying the equipment,
41:57
SPEAKER_00: and there's a spread there for the company.
41:58
SPEAKER_00: Is that right?
42:00
SPEAKER_00: That's right.
42:00
SPEAKER_00: Okay, cool.
42:02
SPEAKER_00: All right,
42:02
SPEAKER_00: so I have another tweet that I pulled here,
42:06
SPEAKER_00: which is highly relevant for going direct,
42:08
and I thought it was actually going to be the kind of subject
42:11
of my next written
42:13
SPEAKER_00: piece next month,
42:14
SPEAKER_00: which is this concept of...
42:18
SPEAKER_00: scale through
42:21
coordination versus consolidation right and so this idea of
42:26
like inherently
42:27
decentralized systems they actually scale through through
42:31
coordination not through
42:34
concentration right which like the traditional pe model is
42:37
through concentration
42:38
you centralize everything you drive costs down as much as
42:41
possible versus a
42:43
decentralized system you actually have to do that which is a
42:44
much harder thing to
42:45
do and much more expensive sometimes up front um which is the
42:48
coordination and so
42:49
SPEAKER_00: this is this is the um
42:51
SPEAKER_00: This is the quote that you had that was interesting.
42:53
SPEAKER_00: I want to dig in, dig into it with you.
42:55
SPEAKER_00: Okay.
42:56
SPEAKER_00: Um,
42:56
you said software scale because distribution went to zero
42:59
manufacturing stall
43:00
SPEAKER_00: because coordination never did.
43:03
SPEAKER_00: So kind of,
43:04
SPEAKER_00: as you think about going from a hundred thousand factories,
43:08
SPEAKER_00: what breaks when it,
43:09
SPEAKER_00: when it relates to coordination in the systems that you're building,
43:12
SPEAKER_00: like what keeps you up at night from,
43:14
SPEAKER_00: to allow,
43:15
SPEAKER_00: to actually allow this to scale up to that thousand factory mark.
43:19
SPEAKER_02: Yeah.
43:20
SPEAKER_02: You know,
43:21
I think that the thing that I was calling out in particular
43:24
with that tweet is
43:27
SPEAKER_02: manufacturing already is a heavily decentralized activity,
43:33
SPEAKER_02: right?
43:34
SPEAKER_02: And because of that decentralization,
43:43
SPEAKER_02: there is an enormous amount of coordination cost.
43:46
SPEAKER_02: And you can imagine,
43:47
you'll hear examples from any company that's tried to work with
43:52
manufacturing in
43:52
SPEAKER_02: the US,
43:52
which is even getting a quote from another factory on how much
43:58
it costs to make
43:58
SPEAKER_02: some product is like a multi-week long adventure.
44:03
SPEAKER_02: And sometimes you never hear back.
44:08
SPEAKER_02: That is the simplest possible thing.
44:10
SPEAKER_02: You're saying, hey, I might want to give you some money.
44:12
SPEAKER_02: What is it going to take?
44:16
SPEAKER_00: Try to do that in France.
44:17
SPEAKER_02: I'm sure that's worse.
44:20
SPEAKER_02: You want to give me money?
44:22
SPEAKER_02: What?
44:24
SPEAKER_02: It's like people will take offense.
44:26
SPEAKER_02: I mean,
44:26
SPEAKER_02: literally,
44:27
there are categories of manufacturing in America today where
44:29
you cannot get a
44:30
SPEAKER_02: quote.
44:32
SPEAKER_02: And so...
44:35
SPEAKER_02: So that's pre-production,
44:36
SPEAKER_02: let alone when you actually have to make something,
44:38
SPEAKER_02: organize,
44:38
SPEAKER_02: okay,
44:38
SPEAKER_02: how many do you need?
44:39
SPEAKER_02: Where do you want it?
44:40
SPEAKER_02: What color does it need to be?
44:42
SPEAKER_02: What surface finish do you need on the metal?
44:44
SPEAKER_02: What is the bore size of the holes?
44:47
SPEAKER_02: Then you manage the logistics and managing through a bunch of different
44:50
SPEAKER_02: intermediaries to get that thing to you.
44:52
SPEAKER_02: Is it a distribution warehouse?
44:53
SPEAKER_02: Is it directly to your facility?
44:56
SPEAKER_02: There's so much complexity and coordination that's necessary.
45:04
SPEAKER_02: And again,
45:07
SPEAKER_02: going back to the example of the pencil from Milton Friedman,
45:10
even a simple product like a pencil has probably hundreds of
45:16
suppliers in that
45:17
SPEAKER_02: supply chain,
45:18
if you go all the way down to the people who mined the graphite
45:21
and made the paint
45:21
SPEAKER_02: and chopped the wood.
45:25
SPEAKER_02: And,
45:27
SPEAKER_02: you know,
45:28
like manufacturing is this complex pyramid where like the base
45:32
of the pyramid is
45:33
SPEAKER_02: just so full of complexity.
45:34
SPEAKER_02: And then,
45:35
SPEAKER_02: you know,
45:35
SPEAKER_02: to get to any product at the top of the pyramid,
45:37
SPEAKER_02: let's say semiconductors,
45:38
SPEAKER_02: like there's just like such a complex web underneath it.
45:42
So just like that was the thing I think I was trying to allude
45:45
to in this tweet,
45:47
which is...
45:52
SPEAKER_02: Without a ton of software infrastructure,
45:56
SPEAKER_02: without a ton of artificial intelligence,
45:58
SPEAKER_02: without lots more robots,
46:00
SPEAKER_02: like that coordination cost is never going to get to zero.
46:04
SPEAKER_02: And so manufacturing has struggled to scale because of that.
46:09
SPEAKER_02: So, you know, I think obviously I can, you know,
46:12
SPEAKER_02: turn the mirror on myself and say,
46:14
SPEAKER_02: hey,
46:14
SPEAKER_02: well,
46:14
SPEAKER_02: with Formic,
46:15
SPEAKER_02: we are going to have scaling challenges,
46:16
SPEAKER_02: to your point,
46:18
SPEAKER_02: as the complexity grows and how are we going to manage that?
46:20
SPEAKER_02: I think that is absolutely valid.
46:22
SPEAKER_02: But I think even at full production scale,
46:25
SPEAKER_02: Formic is still at a much smaller scale than the economy as a whole.
46:29
And I think what we found is with a lot of our customers is
46:32
when you have robots in
46:33
SPEAKER_02: the production line,
46:35
the data that's collected from the robots and your ability to
46:37
coordinate production
46:38
needs in those factories goes up quite a lot because you have
46:43
this kind of layer of
46:44
SPEAKER_02: direct control over your production.
46:46
SPEAKER_02: You can reallocate resources much more dynamically and efficiently.
46:51
SPEAKER_00: So you guys, I want to kind of step back here at the beginning stages.
46:55
SPEAKER_00: You started in 2020.
46:56
SPEAKER_00: So you're six years in.
47:00
SPEAKER_00: And I saw another quote here that,
47:03
SPEAKER_00: you know,
47:03
the mission to give American factory superpowers via robots has
47:06
proven harder and
47:08
SPEAKER_00: more meaningful than any of us ever imagined when we got started.
47:11
SPEAKER_00: So kind of when you were going into this journey,
47:15
what was kind of is there one thing that that kind of surprised
47:18
you of how hard
47:19
SPEAKER_00: it's been or challenge that you just did not foresee um
47:31
SPEAKER_02: Oh, so many things.
47:32
SPEAKER_02: I don't know where to start.
47:35
SPEAKER_02: I'm sure.
47:36
One of my favorite pictures that you can see it floating around
47:38
the internet is
47:38
SPEAKER_02: like there's this room full of people who are like disheveled and tired.
47:42
SPEAKER_02: And there's this banner behind them that says,
47:45
SPEAKER_02: we do this work not because it's easy,
47:47
SPEAKER_02: but because we thought it was going to be easy.
47:52
SPEAKER_02: That probably describes every startup in the world.
47:56
But I think when it comes to robotics,
47:59
there are thousands of details like that.
48:03
SPEAKER_02: So I can start with the complexity around the robots themselves.
48:07
SPEAKER_02: Getting a robot to work in a lab is very easy.
48:10
SPEAKER_02: And I think you were talking about this earlier.
48:14
Lots of cool companies have demos of robots doing backflips on
48:17
the internet.
48:18
But getting that same robot to perform in a manufacturing
48:22
environment is probably,
48:23
SPEAKER_02: you know,
48:25
SPEAKER_02: 100 times harder.
48:26
And then getting that same robot to perform in a manufacturing
48:28
environment
48:29
SPEAKER_02: unsupervised for 24 hours at a time with no human intervention,
48:33
SPEAKER_02: you know,
48:34
SPEAKER_02: for months and months is another,
48:36
SPEAKER_02: you know,
48:37
SPEAKER_02: 1000 times harder.
48:37
SPEAKER_02: Yeah.
48:40
SPEAKER_02: And so I think that's really where the complexity lies.
48:44
SPEAKER_02: How do you go from,
48:46
SPEAKER_02: hey,
48:46
the thing works to it works reliably to the thing works
48:50
reliably and I don't have
48:51
SPEAKER_02: to be involved.
48:53
And I think a lot of robotics companies today are really stuck
48:55
in phase one.
48:56
SPEAKER_02: So that's one piece.
48:59
The other piece is I think a little bit related to the cultural
49:01
thing that I
49:01
SPEAKER_02: mentioned earlier,
49:02
SPEAKER_02: which is I was surprised that
49:05
SPEAKER_02: how many factories we would go to and say,
49:07
SPEAKER_02: hey,
49:08
SPEAKER_02: look,
49:08
SPEAKER_02: we can solve such a huge variety of your problems for you.
49:13
SPEAKER_02: And they would say, no, I'm not interested.
49:16
SPEAKER_02: It's new.
49:16
SPEAKER_02: I don't trust you.
49:18
What is this kind of Silicon Valley newfangled thing that
49:20
you're bringing to me?
49:22
SPEAKER_02: I only work with people I know.
49:24
SPEAKER_02: There's a lot of these kinds of cultural baggages that make it hard.
49:28
SPEAKER_02: And if you look at our website,
49:28
it's packed to the chock full of customer case studies because
49:32
that is the only
49:33
SPEAKER_02: thing that's helped.
49:34
SPEAKER_02: Now we can go to any customer and we can say,
49:36
SPEAKER_02: we'll pull up a bunch of videos and we get instant trust.
49:39
But you don't get that until you...
49:41
So our sales cycle now is like a week or two oftentimes.
49:45
SPEAKER_02: When we started, our sales cycle was probably a year, a year and a half.
49:50
SPEAKER_02: They didn't believe anything we said.
49:53
SPEAKER_02: I think that that's another big one.
49:56
SPEAKER_02: You alluded to some of the other things around supply chain complexity,
50:00
SPEAKER_02: managing a fleet of assets at scale is complex.
50:05
Incorporating new tasks continues to be a point of complexity
50:10
because it's easy to
50:12
look at two tasks that look very similar in two different
50:15
factories and say,
50:16
SPEAKER_02: well,
50:16
SPEAKER_02: the same robot can do both of those things.
50:19
SPEAKER_02: In reality, there's a bunch of nuance that often gets missed.
50:22
SPEAKER_02: For example,
50:22
SPEAKER_02: the human might be inspecting something while they do some tasks,
50:26
and you have to be able to make sure that you can do that if
50:28
you want to actually
50:29
SPEAKER_02: meet that production need.
50:30
SPEAKER_02: Or a round object may be much...
50:34
SPEAKER_02: easier to pick than a square object.
50:37
SPEAKER_02: Or in the case of CPG,
50:41
SPEAKER_02: palletizing looks all the same,
50:43
SPEAKER_02: but if you're palletizing boxes that are taped shut,
50:45
that's much easier because you can use suction cups versus if
50:49
you look at what they
50:50
SPEAKER_02: sell in Costco,
50:51
SPEAKER_02: it's typically these open top trays.
50:54
SPEAKER_02: And so palletizing open top trays,
50:55
SPEAKER_02: you just require a bunch of different tooling and infrastructure.
50:58
There are so many nuanced details that you have to know about
51:01
and get right before
51:03
SPEAKER_02: you can solve these incremental tasks.
51:07
And I think my contention is that the only way to get that
51:10
information and learn
51:11
SPEAKER_02: about those things is to be in the field and in production environments.
51:14
And robotics today has suffered enormously from just having a
51:19
lack of access to
51:21
SPEAKER_02: really good training data.
51:23
SPEAKER_02: And I think the analog that I look at a lot,
51:25
SPEAKER_02: you know,
51:25
like the most successful example of robotics deployed at scale
51:28
today is I think
51:29
SPEAKER_02: self-driving cars,
51:30
SPEAKER_02: right?
51:31
SPEAKER_02: You have on the one hand Waymo's approach,
51:33
SPEAKER_02: which is you spend two to $5 billion to train,
51:38
SPEAKER_02: you know,
51:38
SPEAKER_02: collect a bunch of data and train on a new city.
51:41
And then hopefully after you roll out in service,
51:43
you can make that money back.
51:44
SPEAKER_02: Or you have the Tesla approach,
51:46
SPEAKER_02: which is,
51:46
SPEAKER_02: hey,
51:47
SPEAKER_02: I'm going to make a product that people will buy regardless of the AI.
51:51
SPEAKER_02: And then as I'm getting paid for delivering this product and service,
51:55
SPEAKER_02: I can collect training data.
51:58
And it's much more sustainable and you can get a much larger
52:01
order of magnitude of
52:02
data and much more generalizable data if you're getting paid to
52:06
collect the data
52:07
SPEAKER_02: instead of if you're paying to collect the data.
52:10
SPEAKER_02: And so in our belief,
52:11
SPEAKER_02: I think for robotics,
52:12
SPEAKER_02: that's really,
52:12
SPEAKER_02: really critical is you have to be in the position of,
52:15
SPEAKER_02: getting paid while you're producing useful training data.
52:19
And I think that's,
52:20
again,
52:20
where a lot of the robotics AI labs today are really suffering.
52:24
SPEAKER_02: There's a lot of hype.
52:25
So there's a lot of people willing to fund people sitting in
52:29
the middle of San
52:29
SPEAKER_02: Francisco folding laundry.
52:34
SPEAKER_02: I think you're very dependent on the whims of investors in that context.
52:37
SPEAKER_02: I don't want to build a business that way.
52:38
SPEAKER_00: Sure.
52:40
SPEAKER_00: And how many people is Formic today?
52:43
SPEAKER_02: We're about 70 employees.
52:44
SPEAKER_02: Cool.
52:46
SPEAKER_00: I just want to talk just a little bit,
52:49
SPEAKER_00: just for a minute here,
52:50
SPEAKER_00: because a lot of folks that are listening or watching are,
52:53
SPEAKER_00: in fact,
52:54
SPEAKER_00: founders or unfounding teams of folks building,
52:56
SPEAKER_00: obviously,
52:57
SPEAKER_00: that are touching the real economy.
52:59
SPEAKER_00: And I thought...
53:01
SPEAKER_00: especially because I know you have an investing background as well,
53:04
but can you just talk a little bit about your strategic
53:07
decisions of how to
53:08
SPEAKER_00: actually,
53:10
SPEAKER_00: your financing capital strategy for Formic,
53:13
SPEAKER_00: both just on the equity and the debt side?
53:17
SPEAKER_02: Yeah,
53:17
SPEAKER_02: I mean,
53:18
I'll skip all the cliche answers around like find investors
53:21
that you like,
53:22
SPEAKER_02: people who understand what you're doing,
53:23
SPEAKER_02: et cetera.
53:25
But I will say one of the big things I made a very conscious
53:29
effort around with Formic is
53:32
SPEAKER_02: that I learned after being an investor for a long time,
53:34
SPEAKER_02: especially in deep tech,
53:36
SPEAKER_02: is that there's a very common trajectory in deep tech companies,
53:39
SPEAKER_02: which is you bury your head down for three or five years,
53:43
SPEAKER_02: building some really,
53:44
SPEAKER_02: really complicated,
53:45
SPEAKER_02: really complex piece of technology.
53:47
And then you poke your head up out of a hole and say,
53:50
who wants to buy this thing?
53:52
SPEAKER_02: And that usually doesn't go very well because
53:57
SPEAKER_02: You often,
53:58
SPEAKER_02: even maybe in a very small way,
54:00
SPEAKER_02: have misread the market need or the way that they evaluate that product.
54:06
And you end up with unfundable business or a very low priced
54:12
acquihire or something
54:13
SPEAKER_02: like that as a business.
54:16
The deliberate effort that we made at Formic was to do the
54:19
inverse,
54:19
right?
54:19
Which is start by not building anything and just going and
54:24
validating market demand
54:25
SPEAKER_02: and using everything off the shelf as much as humanly possible.
54:28
SPEAKER_02: Off the shelf hardware,
54:29
SPEAKER_02: off the shelf software,
54:30
SPEAKER_02: off the shelf tooling,
54:31
SPEAKER_02: you know,
54:31
SPEAKER_02: even if we have to overpay for it,
54:32
SPEAKER_02: that's okay.
54:34
SPEAKER_02: We need to prove that customers want this.
54:36
SPEAKER_02: And then once we did that,
54:38
then we felt much more comfortable and confident investing R&D
54:41
dollars to make it
54:42
better and make it more efficient and use software to automate
54:44
a lot of those
54:44
SPEAKER_02: processes and tasks.
54:46
SPEAKER_02: That's obviously not feasible in every aspect of deep tech,
54:49
SPEAKER_02: but I think there's a nugget of truth in that in every context,
54:52
which is especially very technical founding teams have a
54:59
tendency to romanticize
55:02
SPEAKER_02: the solution and not truly,
55:05
SPEAKER_02: truly understand the problem.
55:08
SPEAKER_02: So again,
55:09
SPEAKER_02: that's probably still a little bit cliche of an answer,
55:11
SPEAKER_02: but I've seen a lot of that in firsthand experience.
55:15
At what point in Formic did the debt market open up for you
55:19
guys in getting that
55:20
SPEAKER_02: first debt facility?
55:26
SPEAKER_02: In the specific type of business that we were building at Formic,
55:29
we knew from the beginning that asset financing was a critical
55:34
enabler of robot as
55:35
SPEAKER_02: a service.
55:38
One of the things that we're selling is access to hardware and
55:45
that costs a ton of money.
55:47
So I think we started with the idea that this will have a more
55:50
complex capital
55:51
SPEAKER_02: stack than your typical software company.
55:54
SPEAKER_02: It doesn't necessarily apply to every business.
55:56
SPEAKER_02: I don't think everybody should do what we did.
55:57
SPEAKER_02: But my second hire was my CFO.
56:01
SPEAKER_02: And he was the head of asset finance at EY.
56:05
He built asset financing facilities for every type of asset you
56:08
can imagine from
56:09
SPEAKER_02: rail cars to billboards,
56:10
SPEAKER_02: shipping containers.
56:12
We were able to take some of those best practices and build
56:14
that into the DNA of the company.
56:18
And I do think that it is important for founders to at least be
56:22
educated on the
56:23
SPEAKER_02: various types of capital available to them.
56:26
SPEAKER_02: I think too many founders rely on VC dollars only,
56:29
SPEAKER_02: or in the slightly better case,
56:32
SPEAKER_02: maybe VC dollars and venture debt only.
56:36
come with a ton of trade-offs and are definitely not the
56:39
entirety of the capital stack.
56:40
SPEAKER_02: So depending on the kinds of things that you need to spend money on,
56:43
you should really match up the right type of capital for the
56:47
specific type of task.
56:49
SPEAKER_00: Great.
56:50
SPEAKER_00: So when was that first debt facility then that you guys brought on?
56:53
SPEAKER_00: I forgot when you announced, I don't have it.
56:55
SPEAKER_00: I'm just curious.
56:56
SPEAKER_02: We built our first asset financing facilities in 2021.
56:59
SPEAKER_02: And then we did our first...
57:05
SPEAKER_02: full-scale asset securitization deal for mid-25, middle of 2025.
57:11
But we had access to a variety of different kinds of forward
57:16
flow programs and
57:19
SPEAKER_02: non-recourse to asset financing facilities prior to that.
57:22
Securitization really takes it a step further because it gives
57:25
you a much larger
57:27
SPEAKER_02: scale of access and much more flexible capital.
57:30
SPEAKER_00: Great.
57:32
SPEAKER_00: Why do this work?
57:33
SPEAKER_02: Why are you doing it?
57:36
Um,
57:37
I think,
57:38
you know,
57:38
part of it is,
57:39
I think just like a boyhood fascination with robotics.
57:41
SPEAKER_02: You know,
57:41
SPEAKER_02: I think as a,
57:42
SPEAKER_02: as a kid,
57:42
I've loved seeing things move and going in and out of factories
57:45
and seeing how
57:45
SPEAKER_02: things are made and watching those,
57:47
SPEAKER_02: how it's made videos.
57:47
So that's like,
57:48
I think on the,
57:49
on the,
57:49
like,
57:49
why do I personally want to do this thing?
57:51
SPEAKER_02: But I think,
57:52
SPEAKER_02: uh,
57:53
SPEAKER_02: what we're doing at Formic,
57:54
SPEAKER_02: uh,
57:55
SPEAKER_02: I think is really critically important to,
57:58
SPEAKER_02: uh,
57:58
SPEAKER_02: like building a world of abundance,
58:00
SPEAKER_02: right?
58:00
SPEAKER_02: It sounds...
58:01
SPEAKER_02: very theoretical.
58:03
But when you look at it in practice,
58:04
I go to these factories every single day.
58:06
SPEAKER_02: A lot of the products that our customers make, I use in my house.
58:12
SPEAKER_02: I have that specific thing and knowing where it came from,
58:18
gives me just so much more motivation to think about how do I
58:21
make this better?
58:22
SPEAKER_02: How do I make this more accessible to more people?
58:23
SPEAKER_02: When you look at the poverty around the world,
58:26
SPEAKER_02: when you look at the people who don't have access to basic necessities,
58:29
when you look at how brittle the supply chain is during
58:32
something like COVID where
58:35
thousands of people suddenly didn't have access to something as
58:36
basic as toilet
58:37
SPEAKER_02: paper,
58:37
SPEAKER_02: right?
58:38
SPEAKER_02: Like when you talk about aerospace parts,
58:40
SPEAKER_02: you know,
58:40
SPEAKER_02: submarine,
58:40
like any kind of product that really we rely on as a human
58:44
civilization is built in
58:46
SPEAKER_02: some factory somewhere.
58:47
SPEAKER_02: And
58:49
SPEAKER_02: In my mind,
58:50
SPEAKER_02: if that whole set of machinery dwindles and fails,
58:57
SPEAKER_02: human civilization really,
58:58
SPEAKER_02: really starts to suffer.
59:00
And there are still billions of people on the planet that don't
59:03
have access to the
59:04
SPEAKER_02: basic necessities of life.
59:06
SPEAKER_02: And I don't think that that's going to magically resolve itself unless we
59:10
SPEAKER_02: drastically increase productivity in the means of production,
59:16
SPEAKER_02: basically.
59:17
cool all right you ready for a little rapid fire then we'll
59:20
wrap up sure all right
59:24
SPEAKER_00: who resists you more factory owners or systems integrators i think system
59:31
SPEAKER_02: integrators for sure i mean system integrators hate us because we're just
59:34
intermediating that's what i'm looking for um yeah all right
59:39
what best what what
59:40
SPEAKER_00: best describes formic robotics company a bank or something else
59:50
SPEAKER_02: I would say something else.
59:52
SPEAKER_02: Something else is probably like the largest robot owner and operator.
59:59
SPEAKER_02: So we are developing robots.
1:00:01
We're definitely not a bank,
1:00:03
but we have characteristics that look like a bank as well.
1:00:08
Is the current robotics wave better at building robots or
1:00:11
making launch videos?
1:00:14
SPEAKER_02: Great question.
1:00:16
SPEAKER_02: I'm going to get some hate mail for this one.
1:00:20
I will say many investors have told me,
1:00:23
you need to get better at making launch videos.
1:00:25
SPEAKER_02: So maybe that's a deficiency on my part.
1:00:29
SPEAKER_00: Good answer.
1:00:32
SPEAKER_00: You have to pick one.
1:00:33
SPEAKER_00: I know it's kind of hard to do, but you have to pick one of these.
1:00:37
If reshoring fails in the U.S.,
1:00:39
well,
1:00:39
because of policy,
1:00:40
labor,
1:00:40
or slow automation adoption.
1:00:48
SPEAKER_02: I think all of the above.
1:00:51
SPEAKER_02: I think that there really are,
1:00:52
SPEAKER_02: when I look at the manufacturing economy,
1:00:54
SPEAKER_02: there's a lot of risks.
1:00:56
SPEAKER_02: Automation adoption is a critical ingredient,
1:00:59
SPEAKER_02: but automation adoption is not going to happen by itself.
1:01:02
SPEAKER_02: You need companies like ours.
1:01:04
SPEAKER_02: You need also a lot of policy changes, I think, in the long run.
1:01:07
SPEAKER_02: And I think when you talk to lawmakers today,
1:01:11
SPEAKER_02: there's really a mix of people who are seduced by the hype.
1:01:16
SPEAKER_02: Like, oh, humanoids are going to be everywhere.
1:01:18
SPEAKER_02: We just need to
1:01:19
fund a bunch of humanoid companies to the other extreme of
1:01:22
people who are saying,
1:01:23
SPEAKER_02: oh,
1:01:23
SPEAKER_02: robots are going to take all the jobs and we don't want them.
1:01:28
SPEAKER_02: And then there's another kind of camp,
1:01:30
SPEAKER_02: I think,
1:01:31
SPEAKER_02: in government,
1:01:31
SPEAKER_02: which is saying,
1:01:32
SPEAKER_02: oh,
1:01:32
SPEAKER_02: we only want high value manufacturing.
1:01:34
SPEAKER_02: We only want to manufacture chips.
1:01:36
SPEAKER_02: We want to manufacture EVs, everything else, like let other people do it.
1:01:42
And I think it completely missed the point that you can't
1:01:45
manufacture those complex
1:01:46
SPEAKER_02: things if you can't manufacture the simple things.
1:01:49
SPEAKER_02: So, yeah, I mean, I think there's a lot of ways that this could fit.
1:01:52
SPEAKER_00: Who else should I talk to?
1:01:54
SPEAKER_00: I'm going direct.
1:01:55
SPEAKER_02: You want robotics companies or do you want other sectors?
1:01:59
Anyone you think is interesting that's kind of relevant to this
1:02:02
theme of kind of
1:02:03
SPEAKER_00: decentralization,
1:02:05
SPEAKER_00: scaling through coordination?
1:02:08
SPEAKER_00: Whose work has influenced you?
1:02:09
SPEAKER_00: Yeah.
1:02:17
SPEAKER_02: I am a big fan of Machina Labs.
1:02:23
SPEAKER_02: I don't know if you are familiar with those guys.
1:02:25
SPEAKER_02: I've spoken to Ed.
1:02:26
SPEAKER_02: You've spoken to Ed.
1:02:27
SPEAKER_00: Okay.
1:02:27
SPEAKER_00: All right.
1:02:27
SPEAKER_00: Well, I love Ed.
1:02:29
SPEAKER_00: I'm going to do another one because I saw his funding announcements.
1:02:32
SPEAKER_02: Yeah, yeah, he did.
1:02:34
SPEAKER_02: He did do a funding announcement.
1:02:35
SPEAKER_02: There's another guy that I'm a big fan of,
1:02:38
SPEAKER_02: Sunny Yu,
1:02:38
SPEAKER_02: who's the founder of a company called Fulcrum ERP.
1:02:41
SPEAKER_02: They build ERP software for manufacturing businesses.
1:02:45
And they're approaching it in a really,
1:02:48
really interesting and unique way.
1:02:52
SPEAKER_02: Cool.
1:02:54
SPEAKER_02: Yeah, lots of ideas come to mind.
1:02:55
SPEAKER_02: Let me think about it.
1:02:56
SPEAKER_02: I'll send you some names.
1:02:57
SPEAKER_02: Great.
1:02:59
SPEAKER_00: How can people reach you?
1:03:03
SPEAKER_02: I'm on Twitter just with my name, Salman Farid, S-A-M-A-N-F-A-R-I-D.
1:03:07
SPEAKER_02: You can just send me a DM or you can find me on LinkedIn.
1:03:13
SPEAKER_02: You can find my email on the internet somewhere.
1:03:17
SPEAKER_02: Yeah, please reach out.
1:03:19
SPEAKER_00: Great, man.
1:03:19
SPEAKER_00: Well, we'll make sure to put those in the show notes.
1:03:21
SPEAKER_00: Salman, thank you, my man.
1:03:24
SPEAKER_00: Anyone out there looking for...
1:03:27
to release some robots,
1:03:28
you know where,
1:03:29
you know where to find the Formic team.
1:03:30
SPEAKER_00: So I appreciate it.
1:03:32
And yeah,
1:03:33
look forward to keeping in touch and look forward to seeing
1:03:36
Formic progress.
1:03:37
SPEAKER_02: Amazing.
1:03:38
SPEAKER_02: Thank you so much for having me on.
1:03:38
SPEAKER_02: I really enjoyed it.
1:03:39
SPEAKER_02: Thanks, man.
1:03:40
SPEAKER_00: Hey, my friends, Jordan here.
1:03:42
SPEAKER_00: First and foremost,
1:03:43
SPEAKER_00: a sincere thank you and appreciation for all of our early listeners,
1:03:47
SPEAKER_00: readers,
1:03:47
SPEAKER_00: and supporters at Going Direct.
1:03:49
It's really what keeps me going to do this each and every week
1:03:51
because there's
1:03:52
clearly a community of folks out there looking to build a Going
1:03:55
Direct economy,
1:03:56
SPEAKER_00: which gets me super excited.
1:03:58
SPEAKER_00: Secondly, if you haven't subscribed, please do so.
1:04:00
SPEAKER_00: Our main channel and community is forming over on Substack.
1:04:03
SPEAKER_00: The link is in the show notes below.
1:04:05
And we'll also be putting out our weekly conversations at
1:04:08
YouTube,
1:04:09
Spotify,
1:04:10
and Apple Podcasts.
1:04:11
SPEAKER_00: I hope you all have a great day and remember to go direct.