In short
```markdown
Podcast Notes
Relentless - WTF is Metal Forming | Machina Labs
Episode Overview Guest: Ed Mehr, Co-Founder & CEO of Machina Theme: Discussion on metal forming technology, the journey of Machina Labs, and the challenges faced in early prototyping and capital structuring.
Key Discussions
Introduction to Machina Labs
- Conceptualization of Metal Forming:
- Traditional methods involve heavy, expensive dies (e.g., $120 million for car dies).
- Machina's innovative robots allow for the forming of sheet metal without the need for dies, using a process similar to pottery.
Development Process
- Early Prototypes:
- Initial prototypes faced significant challenges, taking up to 25 iterations to achieve acceptable results.
- Challenges included tearing, buckling, and surface quality issues with formed parts.
- Evolution of Techniques:
- Transitioned from human-guided adjustments to a more automated process where robots adjust in real-time based on sensor data.
Application and Impact
- Market Fit:
- Machina targets industries requiring rapid design iteration and low production volumes, such as aerospace and defense.
- Custom car design collaborations with companies like Toyota.
- Philosophy of Expression:
- Ed emphasizes the desire to enable freedom of expression in design and manufacturing, allowing unique car and product creation.
Challenges and Learning Moments
- Rapid Iteration and Learning:
- Early on, the team learned to adapt and iterate quickly, often making manual adjustments during long robot operation cycles.
- Major Shift Decisions:
- The decision to focus on assembly rather than selling individual systems or parts was crucial for long-term success.
Capital Structure Challenges
- Complex Funding Needs:
- Hardware companies require diverse funding, including equity for R&D and debt for equipment purchases, unlike software companies that primarily rely on equity.
- Influence of Major Clients:
- Relationships with large companies (e.g., Toyota, Lockheed Martin) who can help fund and support operations became essential.
Cultural and Philosophical Insights
- Team Dynamics:
- Emphasis on the importance of the team and shared mission over the technology itself.
- Creating a supportive environment through shared meals and open communication fosters creativity and problem-solving.
- Managing Expectations:
- Setting realistic expectations with customers regarding capabilities and timelines is crucial for maintaining trust.
Personal Insights from Ed Mehr
- Resilience and Mental Challenges:
- Discussed the internal struggles and fears encountered while navigating startup challenges.
- An approach of mentally preparing for worst-case scenarios to mitigate anxiety about potential failures.
- Gratitude for Opportunities:
- Reflection on the advantages of operating in a supportive entrepreneurial ecosystem in the U.S.
Key Takeaways
- Innovation in Manufacturing:
- Machina Labs represents a shift towards more flexible and designer-friendly metal forming technologies.
- Importance of Adaptability:
- The journey includes continuous learning, adaptation, and iteration to refine processes and products.
- Building Relationships:
- Cultivating strong partnerships with customers and stakeholders is essential for long-term success and growth.
- Mindset Matters:
- Emphasizing a positive yet realistic mindset can help navigate the uncertainties of entrepreneurship effectively.
Conclusion The episode highlights the transformative potential of innovative manufacturing processes while underscoring the importance of resilience, adaptability, and a strong team culture in overcoming the inherent challenges of launching a hardware startup. ```
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOHonoring a Design Legend
0:46 to 1:30
Discussion about the tribute piece, Franz Solo, created for Franz von Holzhausen.
“And Machina is basically building these robots that allow you to form metal in a very unique way.”
The Process of Creating a Sculpture
1:31 to 2:15
Explaining the technical process behind scanning and making sculptures from designs.
“I know he was being nice, but it's a pretty massive, basically, sculpture.”
Automating Metal Forming Techniques
2:16 to 4:42
Describing the automated process of turning designs into physical parts using robots.
“This comes down to the core thesis of Machina.”
The Evolution of Production Trials
4:43 to 5:56
Discussing early challenges in production trials and how they've evolved at Machina.
“It was a shroud for a rocket engine, and it took us 25 different trials to even get it right, even ballpark it right.”
Innovative Metal Processing Methods
5:57 to 6:52
Exploring how Machina's robots form metal parts without traditional dyes.
“And now you can program that facility to do all kinds of operations to make all kinds of different parts.”
Applications in Aerospace and Automotive
6:53 to 11:10
Overview of Machina's applications in aerospace, defense, and automotive industries.
“Two robots, they have two stylus finger-looking things that are super accurate, very highly rigid, can apply a lot of force.”
The Vision Behind Machina
11:11 to 13:43
Ed Mayer shares the inspiration and background for starting Machina and its unique approach to manufacturing.
“How did you kind of come up with this idea in particular?”
Creating More Innovators
13:44 to 14:02
Discussion on what it takes to cultivate more visionaries like Elon Musk.
“Working at SpaceX and kind of seeing Elon work firsthand, what do you think is the limiting factor to basically making more Elons?”
The Potential of Multidisciplinary Thinking
14:02 to 15:46
Exploring how a multidisciplinary approach can foster innovation.
“If society could make a billion Elons, I mean, like we're going to be in a completely different world.”
Lessons from SpaceX: Risk and Iteration
15:47 to 17:07
Discussing the importance of rapid iteration and risk-taking in engineering.
“I think those two, multidisciplinary and significant amount of courage to try things, it just became significantly useful in the areas he went into.”
Show all 41 chapters
Creative Problem-Solving in Startups
17:08 to 19:15
How to source equipment creatively under startup constraints.
“So we were like, okay, you know, we need to in a matter of month, get some parts for me.”
Building a Customer Base from Existing Networks
19:16 to 21:08
Strategies for leveraging industry connections to secure initial customers.
“And he was really excited about what we were doing.”
From Idea to Execution: The Manufacturing Journey
21:09 to 23:36
The step-by-step process of turning concepts into manufactured products.
“How did you kind of think through going from nothing in an empty warehouse to actually manufacturing parts for customers?”
Navigating Challenges in Product Development
23:37 to 27:20
Overcoming hurdles and setbacks in the early stages of product creation.
“You know, we're not even thinking about, oh, it's failure.”
Hiring for Innovation: The Right Mindset
27:21 to 28:00
Identifying and attracting self-starters with a bias for action.
“I imagine if you're creating something from scratch, you want people that are just like self-starters and they're figuring things out.”
Navigating Early Challenges in Metal Forming
28:00 to 29:05
Learn how the team's initial naivety and diverse expertise shaped early experiments.
“And so I started trying to kind of get a little bit disappointed.”
The Importance of Passion and Expertise
29:05 to 30:28
Discover how passion and relevant expertise can drive successful projects and collaborations.
“Did you have to screen out anyone that kind of had a huge amount of expertise and looked good on paper, but just didn't, it wouldn't make sense in a real world scenario?”
Choosing Applications for Growth
30:28 to 32:46
Understand how selecting specific applications can lead to significant business opportunities.
“I think our long-term goal is you have a system that can do all kinds of metal structures autonomously.”
Designing for Personal Expression
32:46 to 33:56
Explore the connection between personal passion projects and business success.
“I think, you know, the building, the truck was one of that.”
The Future of Personalized Manufacturing
33:56 to 36:30
Learn about the vision for a world where anyone can bring their designs to life.
“It's the, to me, it's the best plane ever made.”
Overcoming Constraints in Business Growth
36:30 to 38:00
Find out how identifying and addressing constraints can lead to successful scaling.
“So I think the core really inspiration was, and it still is for this company, is like, I would love to live in a world where any great idea can become real.”
Prioritizing Risks for Company Survival
38:00 to 42:07
Learn the importance of ranking and addressing company risks to ensure survival.
“And then the right facility gets programmed to make that part and ship it to you, right?”
Identifying Key Risks in Startups
42:07 to 43:36
Learn how to prioritize and tackle the most significant risks in a startup.
“That's like number one risk for the company that I put on there.”
From Machine Prototypes to Iteration
43:36 to 45:59
Understand the iterative process of developing and scaling machine prototypes.
“But yeah, I think that constant reminder, it's a good option to brutally get rid of noise, right?”
Lessons from Initial Software Design
45:59 to 47:47
Discover the importance of user-friendly software design for scalability.
“Is it modular system that can be easily kind of connected together?”
Building a Culture of Connection
47:47 to 52:04
Explore how fostering informal conversations around food can boost productivity.
“for example you know when we first started to build our software stack yep we were uh hell bent on, okay, this needs to be really automatable, almost like an API kind of software where it's all command line.”
Navigating Customer Relationships
52:04 to 55:09
Learn how to engage multiple champions within clients to ensure adoption.
“scale up to delivering a million truck bodies a year.”
Strategies for Gaining Client Buy-In
55:09 to 56:00
Understand the strategic approach to getting buy-in from clients and stakeholders.
“It's like, you're not selling to one person, you're selling to the whole organization.”
Building Stakeholder Alignment for Success
56:00 to 57:27
Learn how to align various teams and stakeholders for business success.
“Help them even develop the business case.”
Identifying the Right Partnerships
57:27 to 58:45
Discover how to identify customers and partners that fit your business model.
“And if you, you know, that sort of thing can go bad if DNA doesn't match.”
Setting Customer Expectations
58:45 to 1:00:06
Understand the importance of managing customer expectations in tech delivery.
“Bunch of supply chain issues happened and basically they got delayed massively.”
Scaling Operations Efficiently
1:00:06 to 1:01:27
Explore strategies for scaling manufacturing operations effectively.
“And I think we fully deployed, you know, we have, you know, 11 manufacturing systems, two of them we kind of like brought down, but 11 manufacturing systems we had by the end of 2023.”
Long-Term Vision for Business Growth
1:01:27 to 1:03:28
Learn about the long-term goals for expanding customer base and offerings.
“So we can make those decisions based on the customer commitment, right?”
The Decision-Making Process in Business
1:03:28 to 1:06:04
Gain insights into crucial decision-making moments in a startup journey.
“then we can kind of fill the gap with these smaller, smaller orders and smaller customers.”
Investing in Hardware and Efficiency
1:06:04 to 1:10:00
Understand how to structure investments for scalable hardware businesses.
“And they were deciding whether or not they could basically do a monthly subscription or an annual subscription that allowed you to get free unlimited shipping.”
Challenges in Metal Structure Supply
1:10:00 to 1:10:59
Learn about the challenges and long-term strategies in advanced metal structure supply.
“They don't have any strategic capital to employ.”
Capital Structure in Hardware Companies
1:11:00 to 1:13:07
Understand the complexities of raising capital for hardware versus software companies.
“Yeah, I think, you know, one, to your point, one main difference between this hardware company and a software company is that your capital needs going to require a different structure.”
The Importance of People in Startups
1:13:08 to 1:14:56
Explore why the success of a startup hinges on its people and culture.
“Well, I mean, the thing is now is you've actually got a product that you can show and customers.”
Overcoming Challenges in Startups
1:14:57 to 1:17:13
Discuss the internal struggles and resilience required in startup environments.
“You know, it's not really about the tasks, at least for me, the tasks that I do.”
Mental Strategies for Decision Making
1:17:14 to 1:19:50
Learn about mental strategies for handling worst-case scenarios in decision making.
“It's like something inside, There's some resistance you have a lot of time around things.”
Reflecting on Opportunity and Gratitude
1:19:51 to 1:20:53
Reflect on the opportunities available in the U.S. for entrepreneurship and innovation.
“And so we just by definition kind of have to like course correct for that.”
Transcript
Automatic transcript. May contain errors.0:00I was talking with some of my friends in Tesla, a full set of dyes for one of the first cars they made was$120 million. With our process, we realized, okay, we can get these robot sheetformers to form sheet metal without any need for dye. Two robots come from two sides, start pinching the sheet. As you expand, one of the robots starts giving it shape. I would love to live in a world where any great idea can become real. I'm a car guy, so I used to make cars panels with hand. We announced a collaborative work that we're doing with Toyota, enabling people to design a car that's uniquely how they want it to look like.
0:32If you make a company that enables freedom of expression in physical world, that's going to be a massive company. Today, I have the pleasure of sitting down with Ed Mayer, and he is the co-founder and CEO of Machina. And Machina is basically building these robots that allow you to form metal in a very unique way. We're sitting next to the, I think, third or fourth iteration of the Franz von Holzhausen. I don't know how you describe that. Franz Solo? Yeah, Franz Solo. How did that happen? Yeah, no. So we were going to a conference and the organizers of the conference now say, hey, in this conference, we want to really honor Franz for all the contribution he has had to the auto industry.
1:14He's like the chief designer of Tesla. Chief designer of Tesla. Yeah. And he's a great guy, fantastic human being. And it's like, you know, we want to come up with something that, you know, really captures him in something that's permanent and, you know, it matches his contribution into industry. obviously he's built all these amazing metal machines right um and i was talking to him at the time to the organizer i was like hey you know like we've been doing a lot of sculpture so maybe we can do something around that and then it was like yeah that's a great idea and then um we chatted a little bit and the idea of han solo and franz solo came about so uh it shows franz being uh memorized you know kind of memorized in a in a in a metal um and um there's a picture of us giving it to him as a gift.
2:00And I think he looked happy about it. I know he was being nice, but it's a pretty massive, basically, sculpture. When you decided to do this in the first place, what does that process actually look like from just scanning someone's face or body to actually having it made in metal? Yeah. This comes down to the core thesis of Machina. Can you turn ideas into physical reality as fast as possible in the most automated way. So it starts from a design. We feed the design to our software system. Our software comes up with a rough set of instructions for the robots to manufacture the part. Then we send those instructions to the robot.
2:39The robot looks at the sensor data every four millisecond and then make adjustments. So you have the rough instructions, but online, the robots are constantly making adjustments to make the accurate part in real time. So there's both like an offline component and a real-time component. And then at the end, once you're done, the robot scans it so we can look at the final result and compare it to what the design was, make any adjustments we need to do or not. The robot can come up actually with a set of adjustments to make the part more accurate. But the whole kind of tool chain from design and idea all the way to the final physical product is part of our technology stack.
3:15You mentioned when we were just walking around that it used to take like more than 25 attempts before you got a part or, you know, that was made correctly. How is that? How did that start? And how has that kind of evolved over time? Yeah. So what we're doing is something pretty chaotic, right? You know, we call these systems robot craftsmen because the idea is that they can operate like a human blacksmith or a craftsman. a set of robots, they can pick up different tools, then come up with a set of process parameters and sequence of operations to turn an idea, a design, a CAD into a final product.
3:48But a lot goes into that. As a human, we're constantly looking at the piece, making adjustments, changing our course of action. So it's a pretty chaotic, high variability kind of process. So we needed to tame that, right? So early on, we didn't have any data to figure out how we're going to tame this process, right you know that's one of the challenges going from scratch exactly robotic application is that there's not a whole lot of real data out there for you to use to be like okay how am i gonna form sheet metal like humans would with a hammer um so we first built the machine and the idea was like a human will guide it right so we come up with some kind of heuristic set of parameters let the robot do its thing look at the results we make adjustments go another trial another trial So basically it was just human in the loop, but instead of human picking up a hammer himself, the robot was picking up the hammer and doing the work.
4:40And we constantly would look at the data and make adjustments. So early on, one of our early parts we got was from SpaceX. It was a shroud for a rocket engine, and it took us 25 different trials to even get it right, even ballpark it right. And what were all the issues during those 25 trials? You know, like one trial, it comes in and the part tears. You know, we just apply too much force in one area and the part tears. Another, you know, you're going through and suddenly the part in the middle of the forming buckles because of the stresses. Another one, you know, you start getting really bad surface quality, right?
5:16Digs into the material and it just doesn't look good. And then even if you get the part roughly correct, then it's just inches off of cat, right? Because the sheet moves in all kinds of unpredictable ways as we were forming it. So a whole bunch of challenges there. So we slowly start adding more and more tools so that the human operator can kind of like, you know, make adjustments that are necessary in a more direct way. And now we are like averaging around four or five trials. The goal is, you know, with enough data, we can build accurate models that like, you know, zero shot at the first time the robot figures out how to form the part accurately without having to like do any kind of iteration.
5:56I think this is like a very novel way of forming a part. can you talk about how the process actually works from just taking a flat piece of sheet metal to actually having a form part so forming is our core operation right um actually the robots do more than form they do trimming they do uh bending hemming those are some of the things we're working on heat treatment did you just start with forming forming was a core operation right but the goal is that the system can do you know anything a human craftsman can do because once you have that, then you set up a facility, you have like 20, 30, 100 of these.
6:31And now you can program that facility to do all kinds of operations to make all kinds of different parts. But we thought really hard about, okay, this is a platform that eventually can do any parts, but where do we want to start with? Forming is one of the, I think is the largest metal processing sector. Most of the metal parts you see around you are formed sheet metal parts. um so and the the traditional way of making formed parts is that you have to make a die uh giant stamping presses these are sometimes like four-story built you know building size stamping presses you put two matching dies you put a sheet in between them you apply pressure with the with the stamping press and you get your part so and the bottleneck over there is just those dyes you know a you know i was talking with some of my friends in tesla a while back uh you know a full set of dyes for a for one of the first cars they made uh was 120 million dollars right so you're spending significant amount of dollars to to get your first batch of uh die sets to get your first batch of parts so with our process we realized okay you know we can get these these robot sheet formers to form sheet metal without any need for dye um and the way they work is like similar how to a potter form, you know, forms a clay bowl.
7:48Two robots, they have two stylus finger-looking things that are super accurate, very highly rigid, can apply a lot of force. They come and start from a flat sheet that's kind of fixtured in between them. The two robots come from two sides, start pinching the sheet and expanding the sheet as they pinch it. You know, the same way, you know, you get a little, you know, Play-Doh and you push it from two sides, it expands. And then as you expand, one of the robots starts giving it shape. So it deforms it, expands it, and gives it shape. And then incrementally over time, you get to a shape that is defined through software, just instructions sent to a robot, and there's no static die that is required.
8:27So complete gets rid of the die. A matter of hours after your design is done, you can start getting the first physical part. For something like this, when you are doing like skilled production of a single car, let's say like the Model S, it makes total sense to invest in that die and buy the four-story building thing. But for most like iterative, working on things and just iterating on a designer process, like that does not make sense to make a die for. So what kind of initial products did you guys work on? Yeah. So yeah, the natural fit is where the design constantly changing. Right. You know, everybody thinks of our application, oh, maybe it's a good prototyping, you know, kind of tool.
9:10Prototyping is great. But I think, you know, we always thought that this could be even go to early production volumes. So we look at applications where, you know, they're making a few thousand of each design a year. Not a million like you would do Model Y or some, you know, Ford F-150. Few thousand a year. But then you want to constantly change your designs after, like, you know, you manufacture a few thousand. So aerospace became a very good fit. You know, the number of airplanes that, you know, totally manufactured is in a few thousand a year, you know, at best. You know, rockets, defense systems.
9:47So those became early, early customers. So Department of War become became one of our early customers to repair their current systems, their fleet. Right. You know, an aircraft gets damaged and mission. How fast can we make a replacement part for a skin of an aircraft, a landing gear door for an aircraft? Instead of waiting four years, going making dyes, you know, can we get it in a matter of hours? So that was the first application. Expanded to more defense applications, missiles, hypersonics. Those are one of our biggest applications today. But then on the commercial world, there's a whole bunch of applications as well.
10:23You know, we have Franz Solo here. So we had a lot of art, entertainment, architecture applications. but my favorite application is actually, you know, I'm a car guy. So I used to make cars panels with hand. So I always wanted to use these machines to make custom cars, to make expressive cars. We announced recently a collaborative work that we were doing with Toyota, being enabling people to design a car that's uniquely expressive of what they want and, you know, how they want it to look like, you know, being even able to go into a website and say, Hey, I want a Toyota Tacoma, but I don't want it to look like everybody else's cars.
10:57I want the door to look like this. I want my favorite sports team to be formed on the back of the tailgate. And I want this unique vehicle. And I think that's another very interesting application we recently started working on. And I'm excited for it. How did you kind of come up with this idea in particular? Because my understanding is that no one is really doing something similar to this. Yeah. So, you know, I come from aerospace background. I actually have like a unique kind of thick mix of aerospace and software. You know, academically, I was on the software and AI side, but I ended up going, you know, more than a decade ago to SpaceX.
11:32And that's where I started getting more familiar with the manufacturing. And kind of very early on, I kind of realized that, like, the biggest bottleneck for a hardware world is not that we don't have ideas. It's not that people don't want to make all kinds of great products. It's that every time you have to build something, you have to build a factory for it. Um, so once I was at SpaceX, I really saw this tangibly, you know, that, you know, every factory is custom built for the product you're trying to make. You know, you look at the SpaceX's history, you know, now they're a 20 something year old company was founded in 2001, 2002.
12:09Um, there is two rocket families that has been manufactured. Um, one of them is like halfway there, Starship. Right. And both of them are building two different facilities, two different sets of tooling, because you have to build a factory every time to try to make a product. So the seed of my career kind of started at SpaceX. It's like, hey, can we build factories that can make car parts in the morning and then rockets in the afternoon, right? And they're not custom and specific. So I got a lot of excitement about 3D printing. So after SpaceX, I spent some time at Relativity focused on 3D printing.
12:42But 3D printing was just amazing, but it could only enable certain type of parts in an agile way. And then I started Machina in 2019, really focused on, okay, what is the underlying platform we can make that can do all kinds of manufacturing processes? And to give credit to the people who have been working on this, I mean, people have been thinking about this for decades, right? There are people who have been thinking about using robots in an incremental way to do different types of manufacturing operations in universities since 70. Some of those folks give credit to like Jien Chow, Professor Jien Chow at Northwestern, they've been thinking about these since the 90s and writing papers around them.
13:21The key was, how can you do it accurately? How can you do it cheaply? Like consistently? Consistently cheap, in a cost-effective way, and accurately. I think the missing pieces was cost-effective robotics, but then artificial intelligence to come up with the right set of process parameters as a human mind would do. And I think those were the pieces that didn't exist till today. But we're certainly standing on the shoulders of a lot of researchers and folks that have been working on this for a long time. Working at SpaceX and kind of seeing Elon work firsthand, what do you think is the limiting factor to basically making more Elons?
13:56I think, you know, it's a good question. And I think it's a billion dollar question, right? If society could make a billion Elons, I mean, like we're going to be in a completely different world. From my perspective, there's two components. One is Elon is extremely multidisciplinary. Right. You know, I think if you have a common knowledge in one discipline could be extremely useful in a revolutionary in another discipline. You know, you have some idea in automotive that, you know, everybody knows about it. It's commonplace. And you go to aerospace and be like, oh, my God, why haven't we think of that?
14:39So thought of that. Same thing with software. I think he started in the software world and he captured all these ideas around agile manufacturing, rapid iteration, rapid testing. And then he brought in the hardware world, which was very risky a worse world. Right. You know, like waterfall planning for years. Nobody wants to take risks. And he kind of combined those things. I remember there were days, in one of the last few months that I was at SpaceX, they wanted to test a Rapture engine, which is the engine that became the Starship's engine. And people were like, this engine is not ready for testing.
15:17We're not going to do it. And he was like, just do it. It's going to be my birthday. Worst case, it's going to blow up fireworks, right? So he also brought this idea from software world, like rapid iteration. and do it really fast. I think that helped a lot. And then I think also it was just a significant amount of courage that probably got reinforced through successes in the past where he was just positive, optimist. He was like, okay, you know what? We're going to do this and it's going to be fine. It's going to cost millions of dollars, but it's going to be fine. We're going to figure it out.
15:46So there's that courage of trying new things and not be scared. I think those two, multidisciplinary and significant amount of courage to try things, it just became significantly useful in the areas he went into. I think some of the ex-co-founders of PayPal, when he was first starting SpaceX, they made a video, a compilation of all the rockets exploding to try to deter him from basically incinerating his$100 million that he had just made. It was like, that's fantastic. Yeah. And he just kind of decided to do it anyway. We were talking earlier and you said before you basically made your first prototype, it cost a total of roughly$300 ,000.
16:26and you basically got the parts required to build your first prototype just through very unique ways. And I'd definitely like to kind of go over what are those. How do you do that? Yeah, yeah. You know, I think that's one thing I learned from SpaceX is that make things working as soon as possible. Like don't get stuck in analysis paralysis. Get something out there. Try it and see how it works. So when we started the company, you know, we had relatively small seed round at the time. I think we raised like$2.3 million. So we certainly didn't have a lot of money to do a lot of experimentation. And, you know, a lot of like kind of thinking through things before we get started.
17:10So we were like, okay, you know, we need to in a matter of month, get some parts for me. And, you know, we went to some of our vendors that we use today. And everybody has like six to one year lead times. before they can even sell us any new equipment. So we're like, okay, this is not going to work out. So I - We're trying to move at startup speeds. This doesn't work. Exactly. So I was like, okay, you know what? I'm going to go to Detroit and try to see if we can buy some used equipment, used robots, used sensors. And I was like, okay, Detroit has all these auto companies. They use a lot of robots, scanning sensors.
17:46And I remember I went to this, it's like, I would say like, you know, kind of cemetery of robots. It was a giant robot in the outskirts of Detroit, a giant warehouse in the outskirts of Detroit. And it was actually pretty eerie. You enter and it's this like giant room of dead robots. Just failure. Sitting there. And this guy is buying these from OEMs, automotive OEMs, to just scrap them and sell them for parts. So we went through that warehouse and found a couple that was working. and I think if we bought those at the 10th of the price of a new set of robots and it was immediately available just put them on a backup flat but send it to California what would normally cost like millions of dollars cost like a couple yeah I think so I mean like you know we bought the robots itself at the time where you know new was like maybe 200k each and then we bought them for like 20k so like like 10th of that price and you know I ended up also at the time I wanted to be very cost effective.
18:46So we had a very good landlord that was very into robots and he was constantly bringing his friends to show the facility that we were working in. And I'm like, hey, you know, what would be cool to have in your facility is a couple of robots. And he was like, yeah, that would be great. I was like, hey, if you want to chip in for half of the price of the robots, I can kind of like give you more equity. And I was paying rent with equity at the time. And how did you negotiate that deal? We just had a good landlord. I think Like he was, you know, he and his family, they had a defense company as well.
19:20They had a defense industrial company. And he was really excited about what we were doing. So they kind of had some history of like understanding that this stuff, typically in the early days of like, especially a hardware company, you need space. But, you know, you're tight on cash. Yeah. And they were a huge supporter. I was like, oh, this is cool. I want to be part of this next generation of manufacturing. So they were early believers. and I think, you know, so did they bought it? You know, they were like, okay, you know, the first year you can pay with equity, you know, buy some of your equipment with equity.
19:52But then, you know, the catch was like, you know, I want to be able to bring some of the friends and people that I know to see this. I'm like, that's totally fine. It works out. Yeah, so we were pretty scrappy. You know, in a matter of three, four months, I think we put together a system together and formed part for our first customer, which at the time was Air Force in NASA. So how'd you end up getting that first customer? Yeah. So I think, well, one of the benefits of our team was that we came from aerospace. So, you know, and I spend one year after relativity space talking to a lot of people in the industry.
20:26So by the time we started the company, you know, we already knew who the customers were. You know, they were like, hey, there's no doubt we need complex formed sheet metal assemblies. Someone's hair is on fire and you can see it and they know it. It's just, it costs so fucking much and the lead times are ridiculous. Exactly. And they're like, if you can do it, if you can make it with robots and to do it faster, we'll buy it. So we had the customers even before we started the company. But so, you know, it was our own initial network. So we didn't have a huge challenge with the customers early on.
21:00It was mostly about just like, can you actually deliver a product? Which was a big if at the time. When you kind of set out, you know that you have customers and you're going to start this thing, you've got the space and you've got, you know, your first or like essence of a plan. How did you kind of think through going from nothing in an empty warehouse to actually manufacturing parts for customers? It's kind of, well, now I'm thinking about it. Honestly, it was just like one step out of time, you know, it's like, okay, we need to buy robots. Okay. We need to buy robots. We need to put together and weld things together and put stuff in the ground.
21:33You really, it was, it was like, you know, a thousand steps and start taking the first a step and then do it as fast as you can. I think I was lucky that I was able to kind of work with some of the folks that I knew from previous companies and Relativity and SpaceX, you know, kind of joined me early on. My interns at Relativity joined me and, you know, we were all too naive to know that this is not going to work. We're like, okay, let's just, or it works or not going to work. And we just took the first step and then the second and the third. I just did, you know, do it as fast as you can. Um, so we, it wasn't really, it was just generally like do it as fast as you can get to the forming part.
22:14Um, and then just figure stuff out as it comes up. Um, uh, beyond that, really, it was just, just tenacity and just figuring problem solving. Right. Were there any, uh, big like hurdles, uh, before, you know, you were able to ship, like going into, you know, it's nice to say in like hindsight, it took 25 different attempts to make the first part but i imagine that at the time you know that kind of sucks you know you make the first version and you're like holy shit everything broke yeah the second 10th even 15th and you're like fuck it's still not there yeah no it's kind of funny you mentioned that but we were just having fun i remember the first part we made we gave some instructions to robot uh formed a part and we were like excited that roughly the shape is there we're like oh my God, we never thought this is going to work.
23:01But it kind of looks like the ghost of the part we wanted to make. And that's good enough for us to get all excited. So I think we're just excited to get any result. I remember in early days, we didn't even have a full-on software to control the robot based on the sensor inputs. So we will actually take turns. I would literally sit, or some of the other team members will sit at the robot. And sometimes these builds are multiple hours, right? And we will look at the sensor data and play like we are the controller. We would actually make adjustments to the location of the robots to make it work.
23:33Just manually, just vibro-botting? Manually. And it was just exciting to go and see how these parts come out. You know, we're not even thinking about, oh, it's failure. It's like, okay, I wouldn't want to do anything else. This sounds fun. We have giant robots that are forming complex shapes out of sheep, and we're here to play with it. Um, uh, yeah, to some extent it was just naivete of like not caring about the end results and just pushing through and do the cool thing with the robot until it works. Right. And then we will work as much as we could and then we'll get tired and we'll go to sleep and then come back the next day and just keep doing it.
24:09Right. Uh, we just felt kind of lucky that we have the resources and the time to do this. Uh, and we don't have to worry about a lot of other things. Right. Did you have a similar experience at SpaceX? because I think you were working there like wasn't 2012 ish a time frame um that was like before they actually landed the first rocket yeah right what kind of experience did you take from that to this company yeah I remember you know one thing was probably very formative for me at SpaceX was we had to do way more things that we had the expertise for in those days what's an example that like for example you know you know we might have to do a physics-based simulation of a certain thing certain you know manufacturing phenomenon and we don't have the phd in physics simulation that can do it and the idea was like well who's going to learn it when who's going to make a first attempt um you know and and that was a pretty complicated concept for me to wrap my head around because, you know, I came from grad school before that, you know, I left my grad school to go to the industry.
25:20And, you know, my previous thinking was like, you need to go become expert in one field. And that's what you would know. And you go try another expert to figure out their things. And to some extent, that's true. I mean, there's a value in becoming an expert in a field. But when you're working in such a multidisciplinary field, you know, the coordination between the experts is also a whole other set of challenges where a lot of time it makes sense for the first early iterations for people to gain more breath. Yeah. Right? And I think that was the unlock for me at SpaceX. It's like, hey, anybody can go much outside of their field and learn a lot of other skill sets and start putting the early version of things together.
26:01At some point, you need experts to come in, but you cannot start with all the experts present in the room. Right? And that was actually not something I learned in school, not something that was taught to me when I was like a kid. It was like, I thought like, yeah, you go through school, you become an expert in one thing, and that's going to be your contribution. But SpaceX kind of broke that for me. I was like, no, no, no. Like you can go learn those other things. Yeah, there's definitely an advantage and kind of, you know, on purpose, keeping the team small and giving like latitude to people to just go own an entire part or process and figure it out.
Read the full transcript
26:36Because then, you know, it's not like playing telephone with three different parties. It's just one person focused on that thing 24-7. Yeah. And then you're just thinking about, you know, I mean, not to, I think, again, there's a huge value in being an expert. But when you're setting something early on up, you know, what is, who is an expert? Somebody who spent four years maybe of their PhD figuring something out. Okay, if I can fast track that, can I learn that in two years if I'm really excited about it? I think that was the mentality I developed a little bit at SpaceX. And I think it's hugely important for these like multidisciplinary startups where you need to know enough to get things going.
27:12At some point, you need the experts and they need to come into play. But, you know, getting over that fear of, right, you know, like you could learn these other very advanced fields enough to get started. I imagine if you're creating something from scratch, you want people that are just like self-starters and they're figuring things out. Did you look for kind of track record of having built things in the past or what were you looking for? Yeah, I think, you know, bias for action is huge. To your point, people who have done things in the past are not scared of trying things and building the whole stack.
27:47To some extent, also, I remember in early days, I would go talk to experts in robotics. And they would be like, no, this is not possible. You know, you're applying a lot of force with these robots. They're going to deflect. They're not going to be accurate. Robots are known not to be accurate. You know, this is not going to work. And so I started trying to kind of get a little bit disappointed. Okay, these experts maybe are not the good fit to start with. So I think early days, you know, I went back to my interns even, like in previous companies. I was like, hey, you guys are about to graduate from school.
28:21Do you do something? And these guys are smart, high bias for action. but almost all of us were a little bit naive and um you know to criticize it too much we're like okay this would be cool let's try it and figure out what's going on and then that early team came in um and then they tried it they didn't know what the outcome is going to be and it kind of worked and then afterwards like you know some of the experts looked at it was like oh i guess i guess it could work yeah you know and now we can you know you know now we can develop it further But yeah, I think those early days, you want really biased reaction.
28:56You want really smart people that have a lot of biased reaction. And again, to some extent, naive enough to try things and not immediately dismisses. Oh, it's not possible. Did you have to screen out anyone that kind of had a huge amount of expertise and looked good on paper, but just didn't, it wouldn't make sense in a real world scenario? Oh, yeah. I mean, we sometimes screened them and sometimes didn't screen them and they came and, you know, and then later on realized maybe they're not, it's not a good fit for them. Right. You know, it's a double edged sword. You know, you know a lot about certain topic.
29:32You know, obviously you can do a lot, but also like it just kind of cripples you. You put you in the analysis paralysis mode, right? Where you constantly double, you know, checking yourself, you know, putting doubts. So we certainly had to screen for it. And sometimes, you know, we hired some folks that end up realizing, okay, maybe this person was, you know, smart, smartest guy ever, but maybe not a best fit for where we are as a company. I really love Travis Kalanick. And he had this thing where whenever they, you know, Uber entered a new city, he would try to like celebrate the city in some way.
30:04And I think a lot of what you guys have done is kind of similar to that, where you find some customer and you kind of celebrate the customer. Even with Franz, you know, he's this like amazing designer and has designed cars that millions of people use every single day. And you kind of celebrated him by creating this. What type of philosophy are you taking to that when you're just making something like a Toyota truck or, you know, a Franz statue? Yeah. No, no, you're absolutely right. I think our long-term goal is you have a system that can do all kinds of metal structures autonomously. And that's a huge market.
30:39It's a trillion-dollar market. But almost that can become the impediment to actually grow fast because now you have so many markets that you can go after. It's just too many applications. Like, okay, am I going to work in heavy machinery, in agricultural machinery, in automotive, in aerospace, in defense, in architecture, in entertainment? Everywhere needs metal. So to some extent, we were like, okay, we need to make few applications very happy. And what we chose in those few applications were things that we were also excited about ourselves. You know, I loved custom cars. I used to make custom cars.
31:17I was like, okay, this is an application we're going to go. I know enough where I can have a very good intimate conversation with our customers. You know, one of the things we did with Toyota, I remember early on in the conversations, the Toyota manager that we worked with, you know, we went into a meeting with other Toyota execs. And he was like, like, Ed is one of us. Like, you know, he knows cars and he's putting his money where his mouth is. And he just wants to love to build cars. And he already built a car for, you know, themselves without any customer. or they already have a car that's amazing.
31:51It's fantastic, right? So, you know, and the same thing when we came from aerospace, you know, we were in that industry. We were excited about the application. We could become thought partners. We're not just a supplier, right? We're not like you call us and then we'll build whatever you want to build. I'm going to give you opinions. I'm going to be like, hey, this is how you should build it. This is how it's going to get better. You know, I've been thinking about it over the weekend and here's my ideas and my thoughts. And the customer is like, okay, you're not just a supplier. and you're excited about what I'm doing and you want to be a thought partner into it.
32:22And that was a key principle for us to choosing applications we're going into, to have a significant amount of knowledge and interest in that application. And then slowly, and each of those applications could become hundreds of million dollars in revenue for us, right? So they're not small applications. But instead of going broad and just selling to everybody and being kind of this dumb, uninterested customer that just want to sell you a machine or sell you a product, we really wanted to partner um did you do anything that was just purely for your own curiosity and excitement like uh making any part that wasn't for a customer but just because you wanted it to exist and wanted to see if if the machine could actually make it happen um i think if you if i asked me i would always be like no there's always that rationale behind everything we did you ask our team they're probably going to be a lot of times like and was like why are we making this and it's probably what, what he wanted to do.
33:15I think, you know, the building, the truck was one of that. Right. And then, and then later on end up becoming, you know, help us close deals. But I think that's, that's part of the situation, right? Part of the, you know, if you're not excited about it, then you can't sell it either. Right. You know, if I just purely making decisions based on logic, then I'm out there and nobody connects with that, you know? So I think a lot, a big part of this is you kind of have to, you have to create a story that other people want to believe in and the best, the best stories are something that they haven't heard before.
33:48And you kind of come up with it in your own head and you just show it in the real world. And then they're surprised by it and it resets their own expectations of what's possible. Yeah, we made it exactly. We made this show piece. Uh, it's SR 71. It's the, to me, it's the best plane ever made. Uh, and it's very complex shape, you know, um, it's one of the fat is, it is at the time, the fastest plane made. And, you know, I just purely, we purely did it because we were just a fan of that plane. The moment we made it, it was just, Mark, we wanted to take to a show. Half of my customers, a lot of them would start calling me.
34:22I was like, hey, can I get a version of that and have it in my facility? Because yeah, you do something authentic. I think people just connect with it and be like, okay, these guys are not here to just make money. These are not here to just make machines. They're excited about this industry, right? And we are. I know there's different kind of schools of thought where some founders want to basically build a company in order to integrate with some other business and then sell that. And I think most of the best founders are basically just trying to build something that will just keep on evolving over time and grow into a massive company.
34:50How do you think about that when you're first deciding to start it? I don't think I ever joined any company where my end goal was to sell it. You know, at Relativity Space, you know, I joined part of the founding team. That was not the end goal. Same thing with Mokita. You know, I honestly, we even, we didn't even start this company to be like, oh, let's make a profitable business. Um, it's just, it wasn't really that. Right. I mean, I had this kind of remote trust that like, if you build something that's cool and useful, you will make money. Right. Um, while we started, I really, what I wanted to do was like, you know, I was really interested in building physical things.
35:33And I was like, you know, as a college graduated, in order for me to build physical things, I have to join giant companies, SpaceX's of the world, Honda's of the world, Toyota's of the world. And they're great companies, but I need to be part of this larger whole that has enough resources, can build factories, can turn my ideas into reality. So really at the core of this was like, can we enable expression for people with great ideas? starting with ourselves, with the people who started the company. If I want to build a Franz Solo, how easy can I make it? And I think that was... Something that you can do at your own company, but you can't necessarily just go inside of a big company and say, I want to make this cool thing and I want to use one of our machines to do it.
36:15Yes. And I think, and even if you want to do it in those large companies, that means huge investment in tooling and a lot of custom specific machinery and equipment that you have to build to make it possible. And just economically not feasible to just do whatever you need to do or you want to do. So I think the core really inspiration was, and it still is for this company, is like, I would love to live in a world where any great idea can become real. You know, I walk around. I mean, you walk around in the streets, all cars look the same. All buildings roughly look the same. you know there's not really design diversity in our physical world so really the core thought was if we and we make a company that enables um a freedom of expression in physical world that's going to be a massive company that's a company that's you know that's going to define the future you know if tomorrow i mean any sci-fi movie that i watch there's bajillion cars and hovercrafts and rockets and satellites and bajillion designs of buildings on multiple planets.
37:29I'm like, okay, the factories we have today are not going to make that happen, right? So we need something that's flexible and can do all of these things. So that was really the initial motivation. And, you know, and we love building things. And then our customers are great. You know, they come in. I was like, yeah, we're on the same page. Let's make this happen. And then the rest becomes, you know, what is your path? you know in technical terms we call it go to market okay who's your initial customer and then your second customer and your third customer that sees this benefit until you build that world so our long-term goal is i think you know machina could be a very large company a company where you know you can you have some ideas you go on a portal you start turning your intention into a design by even speaking to it or working with it through neural network, sorry, neural lace, transferring your intent into a design and be like, hey, I want 300 of that in Hawthorne, California, in Los Angeles, California.
38:31And then the right facility gets programmed to make that part and ship it to you, right? And anybody in the world can do that. I think that's a trillion dollar company. John Collison has this line, I think he tweeted it a while back where he said, if you just look around in the world around you, everything is so extraordinarily difficult to actually make reality, even like a bench, like a park bench is so complex to get, you know, deal with all the regulation, all these, these hurdles that is effectively just like a universe. You look around and it's just a universe of passion projects. And I think it's very, very positive for any company if they basically lower the barrier to enabling people to make those passion projects.
39:12Yeah. Yeah. I think, you know, I mean, if you think about, you know, humans, I always think about what does it mean to be human, especially now it's becoming, you know, kind of like a more pertinent question with AI and everything else. I think for me, being a human is, you know, we're here to express ourselves in each of us in our unique ways. You know, you're expressing yourself with this amazing podcast that you're making, you know, and that's how you want to define the world. And that's your contribution to a wall. That's your expression. That's my leverage point. Yeah, exactly. And I think, you know, if you can make expression for people easier and easier, it's very easy to put your thoughts now in words, in computer screens, in apps, in a physical world.
40:00It's still very hard to be expressive in it. It's very hard to have a car that uniquely expresses you. It's very hard to have a home that uniquely expresses you. It's very expensive, very time consuming to do that. So I think that's a frontier for me. I mean, like if you think about philosophical frontier is expansion of the boundary of expression. And I think that's what I'm hoping we're doing here at Market. When you first start a company, there's a million constraints. Like you said, you raised a little bit of money, like two million bucks. And that might be enough money to start, but it's not necessarily enough money to scale to massive facilities that cost hundreds of millions of dollars and deploying huge amounts of robots.
40:36So how have you kind of thought about at any given point in time, what are my constraints? And then what can I do with maximum leverage on the resources that I have available in order to get to the next stage where you have less constraints or different constraints? Yeah, it's like, you know, almost it's like a running list of what is the highest risk to the company. You know, you might have 200 risks to the company. And when we say risk to the company, risk to the company's revenue and survival, right? Like how long you can do this and how reliably you can do this. What are the risks to that?
41:09What are the dangers to that? And then you kind of create a list. And then your job, I think, as a CEO or founder is like rank that list from the most imminent danger to the least imminent danger and then go after that list. Like, OK, how can I resolve number one? And like you literally have to be like almost dogmatic about it, not care about anything until number one is resolved. Or usually like number one to three, because like, you know, there's some fluidity between which one is number one. You stack rank the list of fires and you say that one will burn down the building. Yes, exactly. And actually, I have like, you know, it's on my phone.
41:43Like, you know, if you look at like it's like they have there, there's like three, four things on the on the background of my phone that are just the highest risks for the company. What are those right now? Number one is we need to learn a new forming policy in sim using AI. You know, I want to make sure if we finally be able to fully just in simulation, learn a policy to form a sheet of metal. That's like number one risk for the company that I put on there. Number two is actually talent. Right. Like, you know, there's a lot of stuff around hiring, hiring the right people. And then and then with number three is our largest customer that's coming up.
42:22That deal needs to close. So those are my top three today. today and I think yeah I think that's just what it comes down to right you know and be very realistic about you know what are the solutions to those top three and constantly iterate until they're resolved I kind of love that because people like spend a huge amount of time on their phones and having it's really useful to know exactly what the biggest problem is at any given time and be like basically have a constant reminder of this is the thing I need to be focused on when you when you open your phone and you're like gonna check Twitter or LinkedIn, the first thing that you see is, holy shit, this customer needs to fucking close or else we're dead.
43:00Exactly. And I think it's kind of funny because, you know, in a startup, you're constantly being pulled in multiple directions, especially like, you know, as a CEO or as a founder or even executives or even like, you know, high performing engineer in a startup. There are a million things you need to work on and people are going to tap on your shoulder, right and left and be like, can I do this? And everyone's in a while, I have this, I look at it and be like, okay, I can just not respond to that right now. I don't need to respond to that. I shouldn't. If I'm responding to that, I'm wasting time and it's not fair.
43:29I'm actually not doing the right thing for the people in the company and the company, even though they want me to respond to that, I need to let it go. But yeah, I think that constant reminder, it's a good option to brutally get rid of noise, right? And focus on the top three most important things. So going from you, you build the first version of the machine and you start making parts with it. How did you kind of go from deciding to just use that first machine versus like making a v2 and starting to scale up and make new iterations on that like what was that process like yeah i think it's a combination of you know some of this stuff was some of the stuff that just doesn't work yeah right you look at him like okay this did not work so we need to change that and then and the vision that you have for the machine long term right so like you're making compromises between the two you know for example you know again it goes back to that list of risks so you'd be like you with the first machine you build your de-risk the most important things yeah just straight up will we even build the first thing can i even form a part with it right and and if you got a question mark in the back of your mind when you first started the first one yeah i mean it's like can we put this together and form a part that kind of does it even look like a design of a part that we want to form um so that was the first and if you go to the we went to the other building that was the first cells right And then once you remove those risks, then becomes risk number two, number three, number four, and start kind of going through that.
44:53So our first iterations was all about feasibility. Then I think at our second iteration, we started thinking about, okay, if the system works and you can actually form parts and make metal parts in a fast way, the next biggest risk is how fast can we set up factories that can do these things? Deploy the system at scale. Second iteration was like, it needs to be deployable. So we went from grounded, connected to the concrete, rooted in concrete systems to build out of a container platform that can easily be deployed because that was a second risk. You can make parts fast, but if the factories take forever to build, then that's the biggest risk.
45:33So let's make a system that's portable. You've basically got the system now that it can basically fold up and fit on the back of a semi truck. Did you kind of come in with that idea and say, if we're able to put it on the back of a semi truck from day one, like it will make it very easy to move around? Or did you come up with that later on? Yeah, no. So I think we always knew that the system needs to be easily set up. Yeah. But what was that form factor? Was it like, oh, you can put it in the back of a truck? Is it a container? Is it modular system that can be easily kind of connected together?
46:06I think that was not fully fleshed out. What were your other initial ideas of what that looks like? I think the other idea was basically the stuff that I said. I think we did one iteration that was not a container. You could just put it on a flatbed. And then we saw some challenges with it. And it was like, we dropped that design. The other iteration was like, can we make a modular design that can easily be assembled? You put it in ship sets, but then it's kind of like Ikea. You go to the other side and in a matter of day, you can put it together. And then we realized the system is so complex that's not even possible.
46:39So we landed on this container idea. And to some extent, you know, it's, you know, philosophically, that was always in the back of our mind is that any technology that stays with us got to move with us. Right. You know, we are no match species. Right. You know, we have been constantly on the move. You know, even on Earth, we've been on the move. And now we're going to other planets. So any technology that's too static is never going to be able to survive long term. And container was this like unit of mobility. So it kind of made sense afterwards. It's like, oh, you just got to be a container. In hindsight, it made sense.
47:21But obviously, you can't just know looking forward. We're trying a whole bunch of different things. Going from the first prototypes and starting to make parts, were there any like directions that you went down where you later realized that just makes no fucking sense and we got to kind of go backwards and retry it's kind of funny i think it's hard for me to fully pinpoint that because i feel like you know we made a lot of mistakes but i think we tried to immediately fix them uh like this in design of things you know uh to begin with for example you know when we first started to build our software stack yep we were uh hell bent on, okay, this needs to be really automatable, almost like an API kind of software where it's all command line.
48:05You know, it doesn't require to have any kind of an interface. You just feed in geometry into it and the robots get going. And then we soon realize that it's like, oh no, like, you know, everybody in this space needs to have visual feedback. Right. And, you know, we build our first iteration of software to be fully command line based. And then we soon realized, okay, this doesn't scale among other than, you know, a few of us who are nerds about command line, like it's just not going to, not going to easily scale beyond us. So with the next version, we went more toward like a visual, visual paradigm.
48:45So same thing happened in our business development, right? You know, we first, I think, remember in 2023, we, you know, now we're talking about three applications that we were focused on. But 2023, we worked with 30 different brands, 30 different brands. And out of that work, you know, and each contract was like 50, 100K value. So very small values, but we just didn't know who would be the right customer to begin with. so we tried a whole bunch of things and a lot of it just painful and we wanted to deliver on those contracts painful industries that were not a good fit for our application you could make middle structures for them but we're not touching the real enough pain point that was like okay this is an industry that we can grow into and the key was like you know do it deliver it make true on your promises but then you know make an adjustment immediately so there's a lot of those painful moments but I think I would like to think of him as learning moments right um less so as like you know kind of full-on failures yeah Jim Belosick uh from Singa Send which we interviewed we both know him he didn't read a bunch of business books he hasn't watched a bunch of lectures or talks but he kind of intuited the right way to run businesses or at least for him like his model of running a business he's got all these interesting things like uh when he recycles aluminum and cheap metal, he gets paid in cash by one of his, you know, the person that picks it up.
50:14And then he basically like lubricates his business by whenever someone, you know, a FedEx driver shows up a little bit earlier, stays a little bit late or someone on this, you know, in his factory is working a little extra hard and he notices it. He just gives him a hundred bucks. And there's these interesting little ways to make the business run better and make people happier day to day. Do you have anything interesting like that? It's hard to be Jim, Maybe Jim and this kind of ideas. I think, you know, food is pretty important. Food. Yeah, good food. I think, you know, one of the very early things that I soon realized is that, you know, people have intimate conversations usually around food, around dinner or lunch.
51:00um and you know where one thing we started very early on and still kind of stayed on and at marketing is like we started with like once a week early dinner that we kind of cater to the company and bring everyone together and yeah just every come in you know just the pressure the work is off let's sit down there on table chat a lot of good points come out of those conversations that are just people are casually bringing it up. And now we have like, you know, lunches and, you know, we do happy hours often. But, you know, people think of those kind of like little perks as like perks of work. But I think it's actually around, you know, it's a huge productivity lift.
51:41Like, you know, getting people to just like, without the stress of work, just sit down and have conversations. Again, probably not as good as what Jim has. That's a pretty interesting way of like, okay, let me sell my scrap metal and reward people. Not as good at that, but I think I realized just lunch, dinner, conversations around food is always very authentic and helpful. Going straight to Toyota might not even be the right model because you can't immediately just scale up to delivering a million truck bodies a year. You can't just snap your fingers and make that happen. How did you think through which customers at what stage of the company would make sense to have the highest leverage and learnings just internally?
52:22Yeah, I think early on, it comes down to a pain point, right? Who has the biggest pain point? Because the earlier you are, the more risk you have. So the pain point of the customer needs to be so big, it's willing to take the risk, right? I think, you know, even though, you know, our company is not necessarily just a defense company. We were excited a lot, a lot of other applications, but defense and aerospace, they had a huge pain point because they have a high variety of design and huge costs to change things over. So they're looking for agile technology, technology that can easily go from design and one design to another, one material to another.
53:03So they become a very good fit. The other piece was, and then, you know, later on we went to the auto, auto became a secondary application for us. The other piece really is about, you know, manufacturing is so tightly coupled with everything our customers do. I mean, think of Toyota, think of an aerospace prime. They are a hardware making company. So manufacturing is just at the core of it. We're not like another supplier, like for example, Microsoft who sells them like this tool, productivity tool. We are the core of what they do. So any changes in that core, they take it pretty seriously, right?
53:44And almost nobody in the company is willing to take a risk to make a fundamental change in the business just by themselves. So what I soon realized was in order for our customer applications to become successful, we need to get a lot of champions involved. The technical guy might come in and be like, yeah, yeah, you guys are a good fit for application. But you got to have someone internal at those other companies that want you to be a part of their business and want them to win. You've got 10 guys. 10. Yes. In those companies. Because it's such a fundamental change. So you get the technical guy and they're saying, okay, yeah, I like it.
54:21That become one champion. You need to get the business guy to come in and be like, oh yeah, this business-wise makes sense too. So I'm willing to bet my career on this. And then you need to get the investment team at the company coming over like, yeah, yeah, We would love to make sure we see the value enough that you actually want to invest in it and get some of the financial benefit as this company grows. You want the CEO and the executive team to come in and be like, yeah, this is something I want to talk about in media and actually advertise our engagement with this. We want this associated with our brand.
54:49Yes. And like I said, it's such a fundamental change that you don't need just one champion like you would do with a lot of other products. you need to get five, six, sometimes 10 different champions across the organization to kind of make the change happen. Um, um, so that's, that's one of the learnings we had early on. It's like, you're not selling to one person, you're selling to the whole organization. Yeah. Do you have a, like kind of a systematic approach? I imagine you don't have this day one, but through this like iterative process where you're talking with lots of companies, what's been the approach that you've kind of adopted over time to make this work?
55:24If you have have 10 champions before something works how do you get those 10 champions yeah uh i don't think it's yet a science uh it's a lot of art you're putting a brushstroke on see what happens no i think um it requires you know it requires really just talking to the companies almost like you know you know full court press almost a strategy right reach out to investment team Use your connections to work with the executives. Make sure engineering teams are super happy in backing you. Make sure the business case is solid. Help them even develop the business case. So it's really about having a strategy to just make movement on all four or five fronts at the same time with all the team members.
56:11You need to have alignment across your technical team, your business development team, your investment team, your legal team to really push. And it requires a lot of also like external, like our board members, our investors to be involved and kind of make these connections happen, make these conversations happen. You know, have a good marketing strategy. Just make sure the voices hear it out there in the right way. So, you know, I would say it's a kind of full court press kind of strategy, right? And in the sports term. If you look at most of our customers, they're investors in us. Toyota's an investor.
56:46You know, Lockheed Martin is an investor. You know, Yamaha is an investor. So most of our customers, they come in and also invest in us. And that's what we found works in these fundamental changes in these organizations. They need to have a stake in your success. Yeah, they need a stake in your success. Interesting. And the beauty is that, like, you know, a single... Manufacturing is such a large industry that a single customer can make or break, right? A single customer can become hundreds of millions a year for you, right? If it's fully adopts your technology. When you started to kind of build these partnerships, and I don't know, building those relationships inside of those companies, like you said, you basically screen for talent super early and figuring out which which people are going to be a fit of the company.
57:27But it also makes sense on the investor side as well to figure out which companies not only are we going to create products for, but also are going to be like DNA aligned because every investor, you know, at least that sort of thing is almost like a organ transplant. And if you, you know, that sort of thing can go bad if DNA doesn't match. So how did you kind of figure out which partners made sense? It's a good question. I mean, we don't always get it right. Have you had to fire any customers? And if so, why? Yeah. I mean, like we went, like I said, like, you know, in 2023, we worked with 30 brands.
57:59Now we work with less than 10, right? But much deeper partnerships. Almost you have to go out there and make it attempt, put your message out there, work with people. And then also be honest and be like both sides and be like, hey. This is not a good business case for you. This is not a good business case for me. You know, I'm sorry. Let's move on. Yeah, and it's tough because, like, you want to make everybody happy. But, you know, I guess the lazy answer to a question is got to try it and got to be honest about how it's going. Is it a good fit? And then cut it off if it doesn't work. Our last interview was with Mihal from Matic, and he talked about basically when he announced in late 2023 three that they were going to start shipping, you know, robots three months in the future.
58:47Bunch of supply chain issues happened and basically they got delayed massively. And then by the time they ended up actually shipping the first units, what they did was they couldn't deliver the product that they thought they could. And so what they did was they just sent emails to everyone and said, this is what our product can do and this is what it can't do. And if you want this set of solutions, then we'll ship it to you right now. And if you don't, wait three months and we'll deliver something that actually works for you. Did Did you do anything like that with customers on figuring out and setting expectations?
59:17Absolutely. I think, you know, a lot of these things for us, you know, I wish we could be as preemptive about it, like, you know, like you described. A lot of cases was like, you know, we went in, we delivered, and then discussion of the next steps. Okay, can we do this? And we're like, okay, maybe we can do that in two years, but not today, right? And that kind of started filtering out a lot of customers. Again, I think it needs to be, you always want, you know, customers to be having enough pain point that they don't want you to walk away. Right. You know, and those are the customers you want to keep.
59:53Right. So, yeah, we have to do that. But a lot of times the conversations like, you know, it's phase two or phase three or phase four of the engagement. So you're going to be scaling up massively right now. We're in factory two, factory one. How long did you use just factory one? So we moved into factory one in 2020. And I think we fully deployed, you know, we have, you know, 11 manufacturing systems, two of them we kind of like brought down, but 11 manufacturing systems we had by the end of 2023. So we're at full capacity. In 2024, we came into the second facility. I think we're going to be fully operational in this facility by 2027.
1:00:36So like looking at three. the everything's running air three year cycles for the facility to start and get to full full capacity and then we're going to our next facility this year how are you thinking about scaling up you know suddenly it goes from we're just tinkering on how this thing works in the first place and now suddenly you're gonna be shipping actual units to your next facility I think starting almost immediately and then you'll just scale up over the course of the next 12 or 24 months how is of business going to evolve and how are you figuring out how to scale? You know, for us, you know, it became, you know, like I said, we started a lot of customers, narrowed down to few, and then we start seeing some of those few out, even within that list, two or three are now getting to the point where it's like, okay, let's scale, let's start making production.
1:01:23Like I said, we're lucky enough, we're an industry that a single customer can get you to a few, you know, tens of million dollars in revenue a year. So we can make those decisions based on the customer commitment, right? If this aerospace prime comes in and says, hey, I want you to make that UAV. And now, you know, a lot of that conversation is also like, gives you like five, six years ahead, kind of like, this is what the volumes will look like. You know where things are going to be going and you can kind of plan for it. And then we can plan for it. But that also only happens by us keep narrowing and being more intimately involved with each of the customers.
1:01:57So it reduces the risk of expansion. But once we deploy the facilities with our initial customers, then that facility, that's the beauty of our business, is that that same facility can eventually service multiple customers. You can start making with this prime first and then switch to automotive application the next day. So we almost kind of seed our applications with the first application. And it's the first start factories with the first application, but easily can switch to other applications down the road and kind of like, you know, serve other customers. Interesting. At what stage does it make sense?
1:02:29So you've gone from like 30 trials to three really strong customers. At what point does it make sense to kind of expand back outwards and go the other direction and start doing that again? That's something we always talk about ourselves. Like, you know, I would love to someday kind of be like what Jim is for, you know, for some of the other metal structures that we do today, right? He does a lot of good machining, bending. We do complex metal curvatures. I think at some point I would like to be in a place where anybody can log in into our platform, put an order the same way they would do at Send Consent, and get their parts at any volume.
1:03:05So that's a long-term goal. I would say probably we are like four or five years away from it. Um, you know, we start from, we started from very dedicated customers that have, you know, very transparent needs. Um, and I think at the late stage of the business would be where, you know, we go back wide and now anybody can put any order on the portal and then we can kind of fulfill it. Once we start having three, four or five facilities that are serving our anchor customers, and then we can kind of fill the gap with these smaller, smaller orders and smaller customers. Yeah. Yeah. Jim talked about a lot of the things that he does are like non-traditional.
1:03:43He constantly talks about like, I don't want to have a CFO because a CFO will tell me not to do things. But I think the way that he's built his company is he's created an environment where he's going to be really happy day to day. And he's just going to be happy waking up in the morning and going back to work and doing that for a very, very long time. How are you kind of thinking like from the start, there's different points in the business. And I think over time, if you do it right, you are having more fun every single day because you're basically eliminating all the distractions that aren't fun and figuring out which are the fires that like I most enjoy putting out.
1:04:15How are you? How are you designing that for your life? I think it's an interesting, interesting approach Jim has. I think, honestly, I'm already there. You know, I think I love problem solving and I think I'm excited enough about our mission that doesn't matter what problem it is. You know, I'll figure it out. Right. And I'll dive into it. So. So, yeah, I don't I don't think, you know, I'm necessarily doing things that I don't like today. You know, even fundraising. I mean, people would say I'm being prosperous, not not not necessarily being honest here, but I enjoy it. You know, like, you know, I think it's part of the part of going out there, talking to people, you know, getting feedback, you know, understanding in what ways what I'm saying doesn't make sense.
1:05:05So I know maybe I'm fooling myself, but I think I've gotten to a point where I think I can enjoy every piece of the work that I'm doing. I've learned to enjoy the pieces of work that I'm doing because it's a funny transition. you know i'm an engineer by the background you know and you know used to used to make cars by hand car panels by hand and it was like okay i'm gonna make robots that make car panels so i can do the engineering work to some extent sometimes i miss that but i'm as excited to enable so that our engineers can do it enable our engineers to be able to do that they kind of vicariously live through them uh and learn from them and and and so so you know i think you know not to give you a fake answer, but I do enjoy my day today, even right now.
1:05:50I think it's just every problem is an interesting problem. Jeff Bezos has this idea of one-way doors and two-way doors. And I think the biggest example of that for him was when they were first deciding to do super saver shipping with Amazon. And they were deciding whether or not they could basically do a monthly subscription or an annual subscription that allowed you to get free unlimited shipping. And the idea, the risk there was basically that if you give free shipping, then the heaviest customers, the people that use the product all the time are going to immediately adopt it. And so at the beginning, it's going to cost an insane amount of money.
1:06:25And then as you kind of scale down the curve of normal customers, people are just going to use the product more or, you know, use the service more. But it's not going to cost as much as those early customers. Has there been any moments during the journey where there has been this like one way door that you had to decide, like, do we go through? Do we not go through? You know, it's like Rulof Botha at Sequoia has a good name for it. It's kind of crucible moments. He calls it crucible moments. Yeah, we had had few. I think, you know, when we started the company, one of the main things we had to decide on was, are we going to sell systems?
1:07:03Are we going to sell parts? Are we going to sell full-on assemblies of metal structures? and the easiest route was to just not make a decision and be like hey we're open to all three let's figure out what's going on and that's basically to some extent what we did you know sold parts for a little bit we sold systems for a little bit and then came the moment where we like decided that was like okay you know we need to we cannot be doing all three right and it's a pretty tough decision because you know you have customers who are like okay if i stop selling sells, those customers are not going to buy from sales for me anymore.
1:07:38And if I don't stop selling parts, those customers will not buy. If I start selling assemblies. So we had to start kind of thinking about, okay, you know, one of the most important things for us to be clear about what our business model is. And we need to make that decision. Otherwise, we're going to be spread too thin across too many ways of serving our customers. Because you sell sales, you need to have a support system. You need to have engineers out in the field supporting your customers. You sell parts, You need to make a lot of them before you can kind of make, you know, a significant amount of dent in your revenue.
1:08:08Right? So that was one of those moments, right? You know, one of those moments was like deciding, okay, what is our business model? We landed on, we sell assemblies, right? We actually use our systems to manufacture complex products. We sell those products. We're actually operations companies as well as technology company that builds the robotic system. Now, for some of our customers in military, we might deploy and operate these cells in the edge. But for the most part, there was a day where we're like, you know, and it's a tough decision because we're selling like, you know, three, three and a half million dollar cells.
1:08:39We're like, oh, some of the customers want that. We're not going to do that anymore. We're not going to sell three and a half million dollar cells. And that's a big amount of revenue is scary for our board, scary for the customers, scary for our employees. It's like, OK, you're never going to sell through it. It was like, well, you know, I got to imagine you have to you have to like figure out which incentives you want. And if you're just, you know, it's kind of like if the aircraft manufacturer also had to operate the airliner, then they probably wouldn't design it so that they make all the revenue on on basically the plane fucking up and failing.
1:09:08It's the same thing with cars. I think for a long time, the car model was basically the car would be sold at cost and then all the profit would be made up on the parts when it failed. Yeah. Yep. And so with this sort of thing, you through deciding to just sell the finished assembly, you're saying we're just going to make really great machines that do something very, very well. And we're going to own the full stack of understanding is it doing it well? And like the iteration loop of we just need to output things that people use. Yes. Versus just selling the machine. Yeah. And it's uniquely the right path in the United States because to some extent, you know, we don't have an industrial base anymore.
1:09:44You know, you don't talk to, you know, Aerospace and Defense Primes. They're like, hey, there are these companies that I go buy my assemblies from. They're 100-year-old companies. These guys, you know, they don't have any margin to invest in new technology. They don't have any strategic capital to employ. So to some extent, they were asking us to step in and become new version of, you know, tier one advanced metal structure supplier. because if you want to sell into that ecosystem, and we could have sold it, but long-term, they're not in a place to buy advanced technology and invest in it. So it was almost like in short term, we can sell to these guys, but long-term, we got to fix this problem fundamentally.
1:10:24We got to actually create a new version of tier one metal structure supplier. And specifically in the United States, that has been eroded. If maybe it was we're a company in China, we would have made a different decision, but in the United States, it may just make sense for us to sell as a voice. When you're doing a hardware company, a huge part of the business is basically the factory space and the machines that you're actually doing. And you have to make this huge capital investment. And at the start, you just have to upfront the cost because no one will give you the money. But then over time, you have revenue and you can kind of share that revenue and raise debt off of that revenue.
1:10:56How are you kind of structuring so that you can scale most efficiently? Yeah, I think, you know, one, to your point, one main difference between this hardware company and a software company is that your capital needs going to require a different structure. You know, in a software company, simple, you go to VCs, you get equity dollars, and that's it. You know, and you build until you build a high margin product, you know, debatably high margin product. And then maybe you can kind of like, you know, you can become a very successful company, go public, and that's it. One thing you learn very fast with hardware companies is that you need a very complex capital structure.
1:11:35You need equity dollars for R &D investments and engineering, but then you need debt to buy hardware, finance your equipment, finance your purchase orders. You need government potentially grants early on to do some early development work that's high in CapEx. you need to work with your customers sometimes because these customers have huge balance sheets. If you're working with a Toyota, it really fundamentally changes. Yes. A million dollars or$10 million to them is nothing. In some ways, they can finance certain things better than a private equity firm or a venture capital firm or even a bank can finance it.
1:12:20So you need to bring money from your customers. you need to work with foreign governments to do foreign expansion so you almost get this crash course in 10 million ways to kind of finance your business and all of them are relevant versus in a software company or service company it's just not relevant at all when you first started the company did you realize that you were going to have to spend a lot of your like a significant chunk of your time thinking about this stuff or was that learning along the way No, it was stuff I learned along the way. I think I was open to it. I was ready to it. But, you know, one of the things you said, you know, like, you know, I never thought I'm going to start.
1:12:58I'm just doing mostly fundraise. I mean, I enjoy it now. But at the time, if you were to ask me, engineer me five years ago, I'd be like, no, that doesn't sound fun. Well, I mean, the thing is now is you've actually got a product that you can show and customers. and so it's not necessarily the same slog as like, here's the technical risk and we haven't even done it yet. Yeah, yeah. Yeah, no, absolutely. I mean, it's a much better position in terms of it's much more straightforward to finance it. But still, you need to build that chops, right? You need to build that chops around what are all the financial products you can use to build your business.
1:13:33What stories would I just never know to ask but are really interesting? I think maybe a lot of people talk about it but I don't think people tangibly understand this unless there are people in the startups. Companies at the end of the day is people. Company is the people. I truly believe this. People think, oh, it's company, the technology. It's a company, the customers. Elon literally has a line where he says, an organization is just the vector sum of all the... Each human is a vector and it's just, they need to be all pointed in the same line. Exactly. It's one of the most... Us and humans are highly variable.
1:14:15Like you have a lot of interest in, you know, things that we care about and, you know, the way we do things. So I think one thing that's maybe underappreciated in most startups is like, for the most part, is a human organization endeavor, right? And I don't think people talk about it as much, you know. You know, technology gets developed. You know, findings eventually get to happen. Customers come in if you have a useful product. But what makes company great is the people and how excited about it, how mission aligned they are. You know, the feeling that you get every day, you come to the work.
1:14:57You know, it's not really about the tasks, at least for me, the tasks that I do. It's about who I'm doing it with. You know, who are the people that are going to sit down around the table? Am I going to get challenged by them? Am I going to have fun working with them? I think those are the stories that are often not talked about, but I think probably the most important stories in the company. Were there any examples or do you have any examples of your team, because you maybe did this right over time, where your team was able to rise to an occasion or make something that you believed wouldn't be possible possible?
1:15:31Yeah, exactly. We made that truck in three months, and we had like two or three systems we can allocate to it. You know, initially it was like 50 panels. We ended up narrowing down to 13 panels we needed in. The designs need to complete. People were working weekends, you know, last minute making, you know, decisions around, okay, what portion of a technology, like, you know, okay, this portion of the feature I'm just going to turn off and not work on and do a manual adjustment as opposed to relying on the tech stack. Yeah, I think it was impossible to get a vehicle from design to full body panels manufactured and assembled in three months.
1:16:16And we all did it because we had a party coming up and we wanted to showcase the video, showcase the vehicle. No better reason than a party coming up. We need to make this happen for the party. And those are the moments that I think you're all going to remember. you know all the and they were not all pleasant yeah you know some of it was pleasant some of it was fights some of it was you know people you know um being frustrated but then at the end it all comes together right yeah okay final question yeah what's the hardest thing you've overcome you know maybe i'll give an indirect answer and maybe we can kind of dive deeper a little bit and I'll give you any more examples of it.
1:16:55You know, I think startups are probably, you know, building a startup is definitely one of the hardest things. I mean, I've been involved with three startups and just the third one, first one I'm a CEO in. And it's definitely a hard thing to do. And I think the hard piece is around, it's always internal, right? It's like something inside, There's some resistance you have a lot of time around things. When you're wrong, accepting you're wrong. When something doesn't go your way, start pushing against it and figuring out what you got to do next. I think, I guess long story short is I think the hardest things are those internal journeys that you go through.
1:17:39At the end of it, you become a better human as a result, but you feel a lot of resistance. What were the biggest moments where you felt that resistance? There are times where, you know, a business still doesn't come through, regardless of how hard you tried. You know, you have five, six months of runway and you really make sure that you need to do the next fundraise. Otherwise, you need to let a big portion of the team go and you really don't want to hit that deadline. And kind of you start going through the scenarios in your head, right, to be ready for it, to make all those decisions. I think those are the hard moments.
1:18:16A lot of them don't even happen, but like in your mind, you know, you're there. You're like, you're making that decision and you're like, okay, I need to be comfortable with it because it might happen in a month. Right. You're running this like a whole bunch of universes may unfold and trying to figure out which one you're okay with. You know, it's funny. I had this professor at university that was teaching science of decision making. and one of the things you know he was teaching was um you need to be always be positive like you need to be always thinking about um uh the best outcome and be positive and send the positivity out there and then you'll receive it i end up i think that didn't really work for me so what i do which is kind of the opposite of that like i'd be like i'm assume the worst and live through the worst situation.
1:19:10You've just mentally experienced it. And then afterwards you're like, well, it's already experienced it. Yeah. And anything that happens is going to be a better situation. But those moments of mentally going through the worst case, I think is usually end up being the hardest moment. And nobody else also knows that's what's going on. It's kind of funny. Like I was thinking about this last night before I went to bed and I'm like the number of things that I think about where I'm like afraid or worried. I don't even know if afraid is the right word, but it's just like worried and thinking about it.
1:19:40And I never actually tell anyone else. And then I just figure it out the next morning or a day or two later. It's ridiculous for it. Yeah. Yeah. It's just ridiculous. And I think, you know, people have this natural inclination to experience a loss twice as much as a gain. And so we just by definition kind of have to like course correct for that. Yeah. Because obviously, you know, we're making progress. Like things are moving forward and, you know, things aren't actually going to shit. Yeah, yeah, yeah, yeah, exactly. And also at the end of the day, I think, you know, you know, we're so lucky to be living in the United States, you know, gone to, you know, gone through great education, you know, have the opportunity.
1:20:23I mean, like where else you have like venture capital, debt financing, amazing talent and the culture that, you know, appreciates. You get to own your work and own the reward or part of the reward. Yeah, a culture that appreciates entrepreneurship and taking risks. At the end of the day, we're pretty lucky. And keeping that in mind, even the worst case is probably better than most people in the world. It's a pretty empowering thing to think about.
From the publisher
My first interview with Ed Mehr, Co-Founder & CEO of Machina.
What went wrong with early prototypes, rapid design iteration, why hardware companies require complex capital structures.
