#178 CosmicBrain AI: How to Train Your Robot

Josh: mimimimimi

Lisa: mama made me mash my m&ms

Josh: You ready?

Lisa: I’m ready

Josh: Welcome to The Pitch, where startup founders raise millions and listeners can invest. I’m Josh Muccio.

Lisa: And I’m Lisa Muccio. 

Josh: On the show today we have Anto with Cosmic Brain AI. He’s raising $5 million for his robot training company where you, with a pair of VR glasses, can teach a robot how to do just about anything. Also, he brought a robot to our studio.

Lisa: So, I don't know if you know this, but I don't live in San Francisco. 

Josh: Don't? Really? Where do you live, Lisa? 

Lisa: I live in Florida, but the point is I'm not used to seeing like robots walk around. Just-

Josh: The robots are coming, Lisa

Lisa: It freaked me out just a little bit to see this robot walk up to me at the pitch event

Josh: wearing an American flag, and-

Lisa: and dirty oven mitts

Josh: Dirty oven mitts.

Josh: Anto, he's obsessed with robots. He has been his entire life, now he is all about physical AI. Lisa, are we gonna have a robot in the house soon?

Lisa: Oh gosh, if only it could fold my laundry. Then I will allow one. 

Josh: Maybe the dish too, that’d be nice

Lisa: Yes, because nobody wants to load the dishwasher

Josh: The question is, is Anto finally going to make our robot dreams a reality?

Lisa: I hope so cause the kids suck at doing chores

Josh: The pitch for Cosmic Brain AI is coming up after this. And if you wanna join us for the live taping of season 16, we are gonna be in Tampa this April. Learn more about our upcoming show at Pitch.Show/Tampa. 

Lisa: And founders, listen up, if you want to pitch in Tampa, apply to Pitch at Pitch.show/apply.

BREAK

Welcome back to the pitch for Cosmic Brain AI. Let’s meet THE INVESTORS.

Cyan Banister with Long Journey Ventures
Someone of his caliber is taking this seriously? I’m just stoked about it

Charles Hudson with Precursor Ventures
Sometimes a business is what it is, not what you want it to be

Immad Akhund, CEO of Mercury and prolific angel investor
I think with one day of coaching, this would be a 20 million valuation

Monique Woodard with Cake Ventures
I'm looking for a founder that excites me. And I like weird things.

And Rohit Gupta with Future Communities Capital
Maybe I'm not charming!

Josh: Cosmic brain. [clap] We shoulda had cosmic brownies. 

Anto: Hello guys. Meet my cofounder who will be walking in in a minute.

[robot walking noises]

That stomping you’re hearing is the dirty oven mitt robot, which was supposed to look like a boxer.

Anto: Alright, Hi guys. I'm Anto, CEO of Cosmic Brain AI. This is my, uh, robot co-founder can you actually wave your hands and greet the VCs for me? 

Charles: Woah

Rohit: I feel like I should wave. 

Anto: I'll tell you how I actually got inspired in robotics, So I'm seven years old. My mom, she's a physics teacher. I opened her physics textbook and chapter by chapter, I fall in love with physics. From the age of seven, all the way to high school, I ended up building hundreds of robots. A robot dog that can actually guard our house, to a microwave generator that can generate electricity from ocean sea floors. From there I went to engineering school, for electrical engineering and computer science. I almost got kicked out of my dorm room for having a model robot in my room with a lot of equipment, from PCBs to, drillers to grinding irons and stuff, right. I ended up building a smart water bottle that can actually monitor vitals for an athlete, which I ended up selling to the military before graduating my undergrad. From there, I decided to go do my business school at Stanford. Which I dropped out to go work at NASA. I was part of building the largest LLMS in the world, and I still continued my passion for robotics because I believe this is gonna influence the next 50 years of America. So what we are building right now at Cosmic Brain is we are building the Cursor for robotics. This is the future of robot training. If you look at the robotics industry right now most of these companies have to spend millions of dollars, hundreds of expert engineers, and thousands of GPUs just to train their robots on a very specific function. What we are doing is we are making this as simple as possible. So if you're a high school student, developer, or a business owner, we are democratizing the process of robot training into the hands of the public by recording video of you doing different tasks, whether it be cooking, folding your laundry, the robot watches videos and teaches itself on what to do. And we can do all this with just a few clicks and that is what Cosmic Brain is about.

Immad: So you're making the robot or you're just making the software? 

Anto: No, we are completely the software, the brain behind the robots. So we are robot agnostic 

Immad: And you're selling this software to the robot manufacturers?

Anto: No. So we are Cursor for robotics. So just like, video format like MP4, MOV format, each robot have their own format system files. So with our software, you can convert that into any software that is adaptable to any brands, whether it be Tesla, Figure, KUKA, or any other off the shelf robotic thing that you're buying 

Immad: Okay. So the world you imagine is like, I buy a robot, this thing can do a little bit of things, like it can walk, but it can't like, fold my laundry and then I'd show it how to fold my laundry. And then I put that on there and I release it, and then it's folding my laundry. 

Anto: Yeah. There is an application for like, in your house, trying to automate your laundry folding, or in your kitchen, putting the trash away, these kind of stuff. Right. But if you look into medium and small scale businesses out there, most of the factory owners have to ship the robotic arms or large machines back to the manufacturers. Wait for two months to get it retrained, reprogrammed to adapt to the new evolving industry, and get it back. This is a huge hurdle for a lot of industries out there. We are enabling that to happen on-prem by the engineers who are present there. 

Monique: How often is it, does this retraining process typically happen? 

Anto: Yeah, anytime you need a new skill set. So let's say you're an Amazon worker, right? You actually want to introduce a new kind of packaging.

Monique: Hm

Anto: You don't have to now ship the machine out to a different manufacturer. You can do that on prem within a couple of days rather than waiting for years to happen. We also have like a marketplace of skillset. So let's say you are a developer wanting to build a skillset for the robot and you want other people to use your product, right? Just like apps in an app store. You can actually sell it to people on our marketplace. So if you're a Chipotle owner who wants to automate your kitchens, and we have Chef Robotics VLA in our system, right? So you can just plug and play and the robot will learn how to do it. 

Immad: What does VLA stand for? 

Anto: Vision language action models. It's just like LLMs, it's video, plus language understanding, plus how that can be translated into actions. So just the robot waving the arm. So each and every torque, the motor movement, is fed into the software. 

Cyan: So, if Chipotle, for example, shared their training data, do they get paid for it? Like then people can remix it? 

Anto: Yeah. So there are two parts to the platform, right? One is software as a service for the training itself, just like how you see Cursor or ChatGPT APIs or Grock APIs. So depending on usage, hours of operation training, we will charge you based on that. Then there is a marketplace where the VLAs can be sold as an API, so that every time the actual robot functions in that environment, for example, a kitchen, every time the robot actually dice a carrot or an onion, they're using our VLAs and that's money for us and money for the guy who made the VLAs as well. 

Monique: And how much? 

Anto: The base price, can be like, a hundred dollars per month or per million tokens used per day. It can go up to like $200,000 for large enterprises. For let’s say for large clients, right? Like Amazon or Walmart. we have VLAs that's priced at $500,000 per annum. And based on number of hours of usage, number of robots, the pricing increases up to 2 million, $3 million per annum. 

Rohit: How come manufacturers aren't doing this themselves? 

Anto: It's very complex. So right now it takes a lot of money to do this. You need thousands of GPUs. You need expert robotics team behind the scenes. You need hardware on premise

Rohit: Or they need you. 

Anto: Exactly

Rohit: 'cause I'm sure you don't have thousands of GPUs and- 

Anto: Yeah. 

Rohit: I guess what I'm asking is like, if I'm a robot manufacturer, it is in my best interest to shorten these training times, right?

Anto: Yes, exactly. 

Rohit: So like, I just don't see why they wouldn't do this themselves. 

Anto: Yeah. So that's what we are trying to make it happen. Right? So we are enabling, even if you're not a robotic engineer, you should be able to do that with just a few clicks. Before even thinking about spending millions of dollars into buying in specific hardware, you can collect data either through making people wear Meta RayBan glasses, or the data that we have collected over the years, and enter that into the system, and then train the robots or the VLAs that can be used for the hardware you purchase later on.

Rohit: How much video does it take to train a new skill? 

Anto: So we do one-to-one transfer, so from a single episode to a single action. The more number of videos, the more autonomous the robots are. 

Rohit: Yeah, one to one is really fast. That's crazy impressive. 

Anto: Yeah

Immad: I think Rohit's Point was like, Figure is selling a robot. 

Rohit: Yeah. 

Immad: Why wouldn't they go to Amazon and say, hey, you know, give us some videos of like packages and we'll train the Figure robot. Like why would, why would you be required as a software piece?

Anto: Yeah. So that's an excellent question, right? So most of the, robots, they're trained like large, large foundation models, which require all these complexity. Second thing, it requires teleoperation to train these robots, right? Humans have to wear exoskeletons. You need hardware on premise. There's no way with that methodology, you're going to collect the entire data of the world. So this is the massive data scaling opportunity for us. without even deploying humanoids, or robots into the premise, we can collect egocentric human data and transfer that into robots. So we can deploy glasses across the world from Thailand to China to Middle East, or South America, and capture carpenters, plumbers, electricians, painters or warehouse workers doing their daily task and convert this data into actionable tokens. 

Immad: And do you have your own foundation model that you're training?

Anto: Yes. So we have our small, foundation models, small VLAs that we used to train the models as well. and the beauty of that, it, it does not require a million dollars to do it, So we have the base functions like the actuator moments, the torque of the robots, the way in the, which the arms work, all this fed into our small foundation models. 

Immad: Do these robots have like a computer in them that you can like drop your software? 

Anto: Yes

Immad: Like they're smart enough that you can do what you need to do? 

Anto: Yes. So most of the industry is kind of conquered by Nvidia, so they have Jetson chips in it, which makes our task easy, most of the robots are already compatible with a lot of things that we do in the backend Yeah. 

Immad: Is your model on the robot or is it in the cloud? 

Anto: We can do it two, two ways. For large models. It can be on the cloud and it can also be on device. The beauty of small VLA’s is that it can be installed on-prem, on the robot itself, and that's one of the biggest bottlenecks with other industries who are building large foundation models. It can never be on-prem, so there's wifi connectivity issues and when building a lot of robots, they have to tackle that. So a huge amount of pain. 

Cyan: Do you have any customers today? 

Anto: So this is very early for us. So we are pre-revenue right now, but we have some, uh, signed POCs from agriculture, to, hospitality industries, to warehouses. And we are talking to big clients like Amazon, uh, Walmart and all that. 

Cyan: What keeps you up at night? 

Anto: Yeah, so this is a very, very exciting thing that I'm building and, I don't think anyone else have done this before at this, way that I'm doing it. So it requires, me, like coding almost like 10 hours, 17 hours a day, right? The scale at which it has to operate is massive, and I'm looking at how fast I can do it rather than waiting for two years, three years, like a research lab, which most of these companies are right now. We want to deploy this in the market very soon. 

Immad: Do you imagine a world where there's a few big robots, like there's Figure, there's Optimus.

Anto: Yeah. 

Immad: And and that's like how robots come, or do you imagine a world where there's like thousands of robots that have slightly different form factors? 

Anto: Yeah. 

Immad: You know, they come from China and like, I guess in the former world, I kinda imagine that they will build their own software that does this, in the latter world, which is like, probably closer to like the PC world versus the Mac world. You know, I can imagine there's so many different people, they don't know how to build great software and you deliver like this verticalized software. So is that your bet that like the latter world is the world that exists?

Anto: Yeah, so we are, our software is very horizontal in nature, right? So any brand, any robotic industry you are, you should be able to use our software. So when you look into the robotics industry. don't think about it as the next Apple iPhone or Google Pixel, right? It's just like the car industry. There's like hundreds of car brands out there and you'll see like, companies in China building robots, companies in Asia, building robots like in South America, in in US. So, uh, we are deciding the software that it can be supportive to any kind of hardware out there, even a 3D printer hardware you're building at home. 

Immad: Hm. So thousands of robots, thousands of different form factors. Thousands of companies building. 

Anto: Exactly. 

Immad: And you are the software intelligence layer.

Anto: Yeah. 

Monique: You mentioned that you have some signed POCs. 

Anto: Yeah. 

Monique: How long are those POCs? What's the trigger that turns them into full contracts and what do you think or anticipate a full average contract value might look like?

Anto: Yeah. So right now it's a massive data collection process for us. We are collecting data that's untapped, that no other industry in this world has, not even Figure, right? So we are going on-prem where actual work is happening and collecting that data and training models based on that. And that is the, work we are doing with our POCs right now. And eventually in the next couple of months, they will probably turn into full-time paying clients. 

Rohit: Will you have product in a couple months?

Anto: Yeah, we already have a functioning prototype. Yes. 

Cyan: I know this is gonna be out there question, but what about toys? 

Anto: Uh, uh, that's an interesting question. Maybe in the future

Immad: What do you want the toy to do? 

Cyan: I just see that robotic toys with AI are coming and I think everyone's gonna have one in their house.

Anto: So there is a mini robot that we actually, so like, one of the company, uh, built. Um, we can like probably, teach to write or sign your name or clean your desk or something like that. So at home I have like a cookie feeding robot, which is attached to my desk, and while I'm coding, it feeds me cookies.

Cyan: No way. 

Anto: Yeah

Cyan: That's so great. I need to see a video of this

Immad: I'm impressed you've kept your weight in check. 

Anto: Yeah. 

Immad: What's been like the most impressive thing you've seen a robot do using your technology? Like, what's the most impressive task that you're like, wow, that's cool. 

Anto: Yeah, probably cleaned my desk. So- 

Immad: Why was that impressive? 

Rohit: You haven’t seen his desk. 

Immad: There’s coffee cups everywhere.

Rohit: There’s cookie crumbs everywhere

Anto: Definitely one aspect to it. But the fact that robots are able to teach itself just from watching videos, right, which is never possible before

Immad: Hm

Anto: And the robots are able to identify, which is a pencil, a pen, and a crumble of paper on the desk and sort it where it belongs. That was thought impossible three years back and now it is possible. So it's massive, massive technology that we are dealing with. So in, for me to see that in real life, that was like, breathtaking. 

A robot cleaning my desk… would take my breath away…take my breath away [singing]

We’ll be right back.

Charles: Who is responsible for thinking up all the use cases, do those come from you? Like your imagination of like, the, the cleaning robot, the cookie robot? Are you like the idea generator of like this, basically the skills that need to be built? 

Anto: Yeah, so from this time that I started the company, three months ago, four months ago, we interviewed a lot of people, hundreds of manufacturers, professors at Stanford, Berkeley, we understood what the industry actually needs. What are these people trying to solve with robotics, with automation? And what is the lacking factor in the market. Then we got introduced to Agritech. Oh. There's a huge use case in farmlands. There's a huge use case in warehouses, even in homes. Right. So that's how we identified and solved it. 

Cyan: How much have you raised to date? 

Anto: So we have closed some SAFE checks. around 500K. And I got my two other teammates, who worked in Google with Larry and Sergey for the past 20 plus years, who have great experience with robotics and, scaling businesses across the world. So I have a great, phenomenal team. We created the product over the past three months, building every day, coding 10 hours a day. And now we have, in that pace, where this product right now, which we have in house, can be scaled like that. So that's what we needed before the fundraise, and now we are there. 

Immad: And how much are you looking to raise now? 

Anto: Uh, we are looking to raise $5 million. 

Immad: What's the valuation for the 5 million? 

Anto: So, I'm letting you guys put a value on it. Most of the SAFEs that we have raised are uncapped. So we are letting the lead VC put a valuation on us right now. 

Monique: And what do you expect the $5 million to get you to? 

Anto: Yeah, so we are looking at an annual revenue of 10 million, at least with the current, signed POCs that we have. and probably expansion of the team to, we are talking to a couple of like really, really good engineers that we need to hire, and the team, and also buy more robots and GPUs. 

Charles: How many addressable robots are there for you right now? 

Anto: So that's why we are robot agnostic, right? 

Charles: Mm-hmm. 

Anto: So if you look at the world right now, or even the US, almost 60 to 70% of factories have some kind of automation in it.

Charles: Yeah. 

Anto: Right? 

Charles: Yeah. 

Anto: So robots everywhere from your dishwasher to your robot arm in a pick and place warehouse to every, everything is robotics. So we are aiming those, right? We are not waiting for a humanoid to come out and make that flourish and waiting for them to manufacture. We are going with the legacy industry, legacy, manufacturing hubs and everything. 

Monique: What do you think you need from your investors in the next stage of growth? 

Anto: Yeah, we are looking for strategic investors who can actually introduce us to more industries, more manufacturers. Because, we believe that network is what makes the company scale faster than ever. And that's what we value more. 

Immad: I really like the idea, the bit that I'm struggling with it, a little bit is like, I, I kind of had assumed a robot future where there's like a bunch of humanoid robots and they can basically do anything like a completely horizontal humanoid robot. Whereas your pitch is very much like, oh, there'll be lots of robots and they're like mostly stupid, or many of them are already stupid, and we are gonna go train all these existing robots. Which like, I'm trying to think like 10 years from now, which of the futures is like more realistic and maybe it is amalgamation of all the future, both of these. But that, that's the bit I'm struggling with personally too, like okay. You know, what is the future and like, how does this fit in it? 

Anto: Yeah, that, that's a good question. So, if you look at like, LLMs, right? A generalist model is good for a generalist purpose, or it can like, vaguely understand different niches.

Immad: But that's true now, right? 

Anto: Yeah. 

Immad: Like don't you think in 10 years the generalist model would do everything?

Anto: So when it, especially when it comes to the physics, physics world, there's a lot of very precise, highly skilled data that you need to feed to it. So that's why it, it has to be very task specific. Skillset specific VLAs, rather than building a large model that can understand just generalist stuff, right? So if you want to replace a mechanic or a surgeon. Or a painter, right? These are all highly skilled labor, which needs a lot of learning for the robots to do. So the small VLA model is the way to go in this industry.

Charles: Yeah. I think for me, I feel like we're on the cusp of all of this stuff happening. It still feels a click too early for me. So I'm out for that reason. But if I were gonna bet on anybody, it'd be somebody with your background. I mean, you've got the cookie robot, you've got all of the primitives you would need. So this for me is more a statement of like, my limited understanding of the market and where it feels like we are, not anything about your capabilities or ability to execute. 

Monique: I'm similar. I just feel like I don't understand enough about robotics in order to be super helpful to you as an investor. I love robots though. And so I hope you win. 

Anto: Yeah, absolutely. Hopefully you, you can buy a robot from us. 

Monique: I will buy a robot from you for sure. 

Cyan: Will you be a reseller of robots? 

Anto: Because our platform, we can list VLAs and robotic, brands, we can get commission out of every sale of any robot out there. So not a reseller, but an enabler of the market. 

Cyan: That makes sense. Well. I've been thinking about atoms versus bits right now, especially with, um, just trying to figure out like, where I want to put my attention for the next few years. And I've been thinking about robots and I actually do think about robotics and manufacturing as being a defense aspect of our country, because I think it's really important that we bring manufacturing and, and this stuff back on shore. So I would like to put 100K in. And then bring you in for a larger check to meet the whole team. 

Anto: Yes, absolutely. Sounds good. 

Immad: The bit that I'm kind of struggling with a little bit as well, and I feel like I just need to research it a little bit, is like, it seems obvious. Which is also a good thing, but I'm just surprised there isn't other people doing this. Robotics and AI has had like, it was like 2 billion in investment last year or something. So, my main worry would be like, yeah, maybe there's a bunch of people doing this. 

Anto: Yeah. 

Immad: I've done a bunch of robotics investments. most of them are hardware, but they've always taken like way, way longer than anyone anticipated

Cyan: Oh yeah

Rohit: Oh yeah

Immad: and it's been much, much harder than anticipated

Anto: Software is the missing piece.

Immad: Yeah

Anto: So hardware, in the last three years, hardware has got super good, from Chinese robots doing backflips to dancing to anything you can imagine, kung fu to karate. The software is the missing piece, and that's what we are doing. 

Immad: Yeah. I feel like your model of applying it to existing hardware, like that will happen way faster than like someone doing a hardware startup. Uh, but yeah, so I guess out for now, but, you know, I reserve the right to go think about this more. 

Josh: No. Once you make a decision in this room you’re forever tied to it

Immad: Just never allowed to speak to Anto again

Cyan: Never allowed. 

Immad: He's blacklisted. 

Rohit: Yeah. I, I like the space. I've, I've looked at a ton of stuff here. I just, it's hard for me to say what's a winner. Like at this point I, I think I'd just be betting on you, which is kind of where I like to be. But I'd have to understand the industry a lot better to say like, my one bet is this person. 

Anto: Yeah. 

Rohit: So I'm out for this one. 

Anto: Sounds good. 

Josh: Alright 

Cyan: Thank you so much. 

Anto: Thank you for the time

Cyan: Yeah, thank you.

[clap]

Josh: Cyan, you're silent, but deadly over there. 

Cyan: Yeah

Josh: Like, I don't know what you're gonna do. And then [clap] you’re in with the decision

Cyan: I'm always like that. Yeah. I'm always like that. Whenever the industry is sort of investing in a space, I try to look six years out and try to think of like, what's coming next. And, you know, I think robotics and manufacturing is gonna be a, a huge wave. I really like the idea of the marketplace of recipes and that the idea that, you know, I can train a robot and then share it and maybe even make revenue and it may make, make this whole thing more complex. I don't know. 

Josh: Yeah. 

Cyan: It's so early, but I like him a lot. And I like his experience and his background. I'm always drawn to an inception story that starts from like age seven. because that means it's their life mission. I love entrepreneurs like this and I would back them all day long. , for me, I'm glad to be in and, and hopefully we can put a larger check in here.

Josh: Yeah. So the 100k commitment, this was from what you call, what'd you call, Wayfarer?

Cyan: Wayfinder

Josh: Wayfinder 

Cyan: as you know, I used to do angel investments on this show too. What we decided was we needed to move with speed, with high conviction bets to get a toe hold into companies, if you hem and haw and you think too long about something, you don't get in and these things become oversubscribed fast. you know, and then we're also looking at, instead of multi-billion dollar outcomes, some of these outcomes end up being much larger than that. So I think it just kind of makes sense to have some exposure in some of these. 

Josh: I know you guys have quite a bit in assets under management in your new fund. Like how much of this is adjusting to the new size of your fund or is that-

Cyan: The fund, this fund's the same size as the last one. It's only slightly bigger. So it's not about fund size. It's more about, every time we raise a fund, we think about where the world is going and how are we going to get the exposure that we want for the deals that we want to concentrate on? 'cause we're also a concentration fund. and any deal that any of our venture partners have done, anybody in our angel network, as long as they have proprietary information and they're close to that company, we can write a check into the series B. And so we have a very different strategy than a lot of funds in that regard. and that's why I'm doing this, is because I just want to be able to get to know the company and move fast, and founders love that. Like when you have a high conviction, I'm in. That's what they want. They want someone to believe them , and I, I'm comfortable with a company that doesn't have product market fit and hasn't figured it out yet.

Josh: Yeah. Any other thoughts? 

Charles: Robots are cool, man.

Josh: Robots are cool. 

Cyan: Robots are cool. 

Monique: Robots are cool 

Cyan: Robots are super, super cool. 

Anto walked out of the pitch room with 100k from Cyan. And immediately celebrated with his robot cofounder. And actual cofounders.

When we come back, six months have passed. 

A lot can change in six months.

Yeah it can. What did, after this.

Break

Welcome back. Six months after The Pitch we jumped on a call with Anto to catch up.

Josh: Okay. What happened after the show with Cyan? You had the first call with Cyan and the team, and you basically re-pitched and then they said, oh, you gotta talk to Scott Banister. This is Cyan's husband. What happened after that call? 

Anto: Scott was like, hey, your go to market strategy is not good enough. How are you gonna make money right away? 'cause if you see the trend right now in the industry, most of the robotics companies are stuck as research labs. Everyone is building, building, building, and no, no traction, right? So, that's kind of the problem that we needed to solve. Like, where can we actually find those customers? Who can actually buy this product from us right away. 

Josh: Okay. And so when Scott dug into the go-to-market with you, the big question was, hey, not a lot of people have robots yet for them to even be able to use the Cosmic Brain software. So how are you going to get this to market? Right? 

Anto: Exactly. Yeah, like scalability and all that. Did a lot of study around go to market and we actually got some really good clients and partners. Now we have like automotive industry, like who manufacture cars as one of our clients. so Scott actually helped us to rethink on our strategy and everything. And he was convinced Then Cyan in order for her colleagues to get convinced fully, I had to take her to one of our fight clubs. 

Lisa: Oh!

Josh: Now these fight clubs, is it what it sounds like, a robot fight club where you just have two robots in a boxing ring, like duking it out 

Anto: very much. Yes. So yeah, it's two robots operated by humans. So I took her, her kids to this, the fight club that I was doing in SF. 

Lisa: You take VCs to your robot fight club to get them to understand what you're doing, and then they invest.

Anto: Yes, it's not like I don't take them intentionally, but when they come there, they actually see the potential in robotics and robotics founders People are seeing all robotics is like research labs or just trying to fold a T-shirt for the past five years. These kinds of stuff, right? No, robots are actually exciting. Robots are much beyond that. So you got to see it to believe in it. 

Lisa: Did you get a 100k like in the pitch room or did you get a bigger check from Long journey? 

Anto: No, we just got the 100k from Long Journey

Lisa: Okay

Anto: but talking about long Journey got us more checks. So we closed our pre-seed in December and we’re just opening up our, another seed round. On a proper valuation. 

Lisa: So how much did you raise in total for your pre-seed? 

Anto: Around like a million dollars. 

Lisa: Okay. 

Josh: Cyan's great. I mean, their whole firm is built around the idea of backing weird mission aligned founders, so congratulations. You're weird. 

Anto: Thank you. 

Josh: We at the Pitch Fund also invested 150,000 alongside Cyan, and now you're going out to raise 5 million for your seed. Is that still the target? 

Anto: Yes. How do you know? 

Josh: Well, that was your target in the pitch room. 

Anto: Oh yeah. Yes.

Josh: No, I read your mind. 

Lisa: how does your pitch differ for this new round versus how you were pitching it previously?

Anto: Right now, we the tagline is AI data translation infrastructure layer. It’s hard to say. But yeah, So we are the platform which can translate a massive amount of data you give us any kind of messy unstructured data, we can actually convert them into intelligent models that can be used by robotics companies to make their hardware much more smarter.

Lisa: Okay

Josh: That’s great. 

Anto: In the next two years, there will be like humanoid robots getting released in the commercial space and let's say your robot needs to understand to learn to cook or work in your workshop. All they need to do is talk to us and we'll give you the VLAs which is trained on that particular skill set and you just need to deploy to the robot and the robot will start functioning like that. And we are also building a marketplace around it. So each of these skill sets and each of these clean data will be uploaded into the marketplace. So thousands of people can come into this one stop experience for all their robotics needs. 

Lisa: I like this. This is a marketplace I'm not scared of in the traditional sense. 

Josh: I mean, during your pitch, was like the marketplace was your pitch. It was we are Cursor for robotics. 

Anto: Mhm

Josh: But now it's almost like it's flipped. You still have that marketplace behind the scenes. But instead of like, selling it as a marketplace, you're selling directly into different verticals themselves and saying, Hey, we're going to be that software partner. We're actually going to help you develop these skills and train your own robots. And then you're kind of backloading essentially the skills marketplace over time by focusing on these specific verticals now

Anto: Yeah

Josh: is that, am I understanding correctly?

Lisa: Yeah I love that.

Anto: Exactly. Because when I talk about the marketplace upfront, people will be like, hey, there's not enough customers in robotics. Not many people have hardware, right? So who's going to be buyers and sellers here, right? But if I concentrate on the underlying technology we're building that makes more sense. Okay, now this company has built some infrastructure that can actually convert messy data into actionable tokens that can be used by a lot of customers out there

Josh: Interesting. Any like cool use cases that you've developed for the technology since the pitch or that customers have asked for since the pitch?

Anto: Yeah, so we were approached by large airlines, in SFO, for using humanoids to load and unload luggages. So everybody has this problem of like losing their luggage or the luggage is damaged by humans throwing it around. So they actually need robots to put it the conveyor belt or put it inside the luggage space in the aircraft. So that was a very, very interesting opportunity. So yeah.

Lisa: Interesting

Josh: Yeah. Well, we have one more question for you before you go. Where's the cookie robot? 

Lisa: we need to see the cookie robot. 

Anto: Oh no. My actual humanoid is hanging out there. If you want, I can show you that 

Josh: You don't have the cookie robot anymore. 

Anto: I disassembled it because I got a better robot

Both: Noooo!

Lisa: Anto, we heard all about this cookie robot and we never got to see it. 

Anto: Yeah. I disassembled the cookie robot because I upgraded my robot 

Lisa: Oh my gosh. What do you use your humanoid robot for? 

Anto: I trained it to vacuum my house. 

Lisa: To vacuum

Anto: Yeah. And now I'm teaching it to use my coffee machine. 

Lisa: Oh. 

Anto: It can like pour the grounded coffee into the filter.

Lisa: Yeah. 

Anto: And take the filter and put it in, in the coffee machine and press the button. 

Lisa: Will it pour you a cup when it's ready and bring it to you? 

Anto: Yeah. And you know you can take it out and pour it

Lisa: Ohhhh 

Anto: So the entire task of like operating a coffee machine, 

Lisa: I want that too. 

Josh: You don't need that. It's good to get up and get yourself a cup of coffee

Lisa: Because I have you to get it for me. That's why. 

Josh: Okay, fine. You can have a robot. 

Lisa: Someone bringing me a cup of coffee in bed is like my love language. 

Josh: I wouldn't want you to love a robot. So I guess I'll keep getting you coffee. Thank you for helping us unpack the robotics industry. Like I'm fully bought into the idea that we're gonna see massive transformation in the world economy because of robotics. So getting to work with you and hear your opinions as someone who's been steeped in this your entire life has been a lot of fun and like a really good learning experience for us. So excited to be invested. Excited to be on this journey with you. 

Lisa: Yeah. 

Josh: And excited to learn with you along the way. 

Anto: Thank you. Absolutely appreciate this whole thing. Thank you for believing in us.

Lisa: What you didn't see in this pitch is that after the robot leaves The Pitch Room, he comes and stands by me during the entire pitch. And it was a little creepy 'cause it just stood there watching me the whole time. 

Josh: You know, if you're gonna have a robot that does your dishes, you're gonna have to get comfortable with this.

Lisa: Well, I think we've determined that's not coming anytime soon. Did you see the robots at CES?

Josh: Trying to fold laundry. Yeah it was painful. 

Lisa: Yes. 

Josh: Painful to watch. 

Lisa: And I think that 

Josh: my kids could do it better. 

Lisa: No, actually they couldn't. 

Josh: yeah, we're not there yet, which is why it makes a ton more sense for Anto to be selling to airlines or 

Lisa: Yeah. 

Josh: Manufacturing facilities like the B2B use cases where it's very task specific. Makes a ton of sense. I don't think we're gonna see a general purpose robot walking around our house for a while. 

Lisa: Yeah. Not in our house. The only place I wanna see a robot is in the boxing ring. 

Josh: Duking it out.

Lisa: Yes. Okay. Also, how funny is it that he is closing VCs by taking them to a robot fight club 

Josh: Only in SF, I don't. 

Lisa: Apparently it worked though

Josh: Apparently. 

Josh: This is one of our last investments out of fund one.

Lisa: Yes, it is. 

Josh: We've just got a couple more coming soon. 

Lisa: dun dun dun

Josh: Stay tuned. But to wrap it all up, robots are cool. 

Lisa: Robots are cool and a little creepy. 

No offer to invest in Cosmic Brain AI is being made to the listening audience on today’s show. But you can invest with us by becoming an LP in The Pitch Fund. We’re raising fund 2 right now! To learn more, go to the thepitch.fund 

Next week on The Pitch… 

Joe: These are fully standalone smart sunglasses everything you need is there. Can we show you the future fitness? 

Elizabeth: Yes

Dawn: Yes, definitely. 

Elizabeth: We'd love to see this

Charles: Could’ve used it this morning when I went running 

Elizabeth: Are you running fast, Charles? 

Charles: Not fast enough.

That’s next week! Subscribe to The Pitch on your favorite podcast player so you don’t miss future episodes. You can watch full length versions of every pitch over on our Patreon at The Pitch Uncut.

And if you’re a founder raising a pre-seed or seed round you can apply to pitch on season 16! We’ll be taping in Tampa, Florida this April. Send us your deck at pitch.show/apply DON’T WAIT.

We’ll see you next Wednesday, in the PITCH ROOM.

–

This episode was made by me, Josh Muccio, Lisa Muccio, Anna Ladd, and Enoch Kim. With deal sourcing by Peter Liu, John Alvarez, and Phoebe Sun.

Music in this episode is by The Muse Maker, Breakmaster Cylinder, The Firmware Rebels, Nightdrive 2086, Imagined Nostalgia, Ønders, Amico Mio, and Peter Jean & The Runaway Queen.

The Pitch is made in partnership with the Vox Media Podcast Network.