How I Made $20M with AI in One Year

0:00Hi, I'm Ben, the founder of Pulsia, an AI that builds your company while you sleep. In November of last year, I started building it alone from my apartment in Paris. Less than a year later, I'm now in San Francisco at a $20 million run rate, having raised $30 million at a $250 million valuation, and I did it all as a solo founder. Here's

0:21everything that happened month by month. But I'm spending about like more than a million dollars a month on AI. Haters are hating on the fact that it's not perfect. They point out some bugs. You can see the negative or you can see the positives, the beautiful things and the the beautiful moments. I think this is like everything I wanted and it's working and I think I just need to to push more and get more users on

0:58the platform. Yeah, I'm getting really really excited about this idea of actually building company OS. Actually, I think I'm going to name it Tulsia. Pulsia, ready when you are. I had spent the whole year traveling the world and I came back from Japan with one idea I couldn't shake. An AI that could run a whole business by itself. I knew I had to build it and I knew I needed somewhere calm to do it. My apartment in Paris was that place. Zero distractions, just me and the product. I spent the whole month locked in building the first version of Pulsia as fast as I

1:42could. It was a pretty focused time. kind of like you know when you lock in and when you want to really build you need as little distractions as possible and having a good routine in a place that feels good and just build build right by December V1 was almost there. The problem was the feature list which

2:01had no end. I could have kept building for months and never shown it to anyone. So, I picked a date.

2:15After the launch, it was all hustle to get new customers. I emailed every user from every product I'd built during the year. I got active on Twitter and LinkedIn and became a reply guy, jumping into every conversation about AI agents. I even started running some meta ads. Nothing fancy. All these techniques got me my first 100p paying customers. All right. So, it's really working. Like I launched Pulseia like I don't know like it's been a month now. But at the same time, I I kind of coming back to like I think I need to to move to San Francisco and then maybe I could start marketing it more and maybe I don't

2:56know. So, I think I think I'm going to go I'm going to do it. I think I'm going to move to San Francisco. I'm so excited. I think this is like everything I wanted and and it's working and I think I just need to to push more and get more users on the platform.

3:17I landed in San Francisco and started talking to everyone I could. The more people I talked to, the clearer it got that I needed to make noise. So, I came up with a stunt where my AI would raise the funding round by itself. I've been configuring Pulsia to have access to my email address. So, it responds to every email that comes in and it can detect if it's an invest an investor. You know, AI is so good at like managing conversation that like it actually researches the the investor and then can reply and can tell it all about the details about how the product works about the the how much

3:50were we raising. And I've been talking to Pulseia about it and it's like I think we should pay him at a billion dollar valuation which is kind of crazy but you know maybe maybe some investors will bite but we'll see. I gave Pulseia main box and let it answer every investor. It was provocative on purpose and I figured it could grow the business

4:08and maybe even raise a real round. The stunt blew up.

4:34Investors were suddenly reaching out to me. I decided to use the momentum while I had it. I started taking meetings and I ended up raising $30 million at a $250 million valuation.

4:50Fresh off the race, I had money in the bank, but a product wasn't ready for the scale that was coming. Bulia had to become rock solid before it could become big. So, I went deeper with Sapium, a startup building infrastructure for AI agents. Yeah. So, BIA is working with Sapium very closely. We've pulled more engineers into the product so that like they can help me with like a lot of the product sides bugs and things issues that are happening so we can then later go back to more infrastructure level so that they can continue just building the

5:23platform. Their engineers came in to help fix the bugs and harden the platform. But there was a bigger problem. My AI bill was huge and most of it was going to entropic. And so I had this million dollar bill last month actually was more like 1.2 million. Thankfully I'm able to afford it for now but it's not a sustainable business model. And so it becomes so important for the business to be able to give this level of AI service to our customers without charging them like 300 bucks a month or 500 bucks a month. We

5:54need to figure out the cost. So we started moving to open source models running on GPUs I rented. It was the only way to keep growing without going bankrupt. April was about making sure that when the next wave came, the product could handle it and I could afford it. Like the bill is getting so crazy. Um like initially I had like pretty good margins but like and so I've been working with like Sapium on like figuring out how we could figure out like a way to like reduce that bill. And like one of the solutions we came up with was um trying to figure out how to

6:30run GPUs actually and run open source models. So I think it's really it's a really hopeful solution. I mean I don't really have a choice. I need to figure it out. So yeah, and I think I'm going to I kind of want to go see visit the data center to see how it looks like like how what are we renting? So

6:49probably going to do that. It's crazy how many GPUs are on the things. It's insane, actually.

7:10Yeah. Yeah. So this is a dashboard that shows uh our migration from only using entropic to switching it to open source models and on GPUs which is way way way more less expensive. Obviously like the entropic models are like bestin-class. Uh the open source models are very good. They're not as good, but the question is like, you know, not as good can still be very good rolling it out because we are we going to, you know, we want to save so much money on entropic and reduce that million dollar bill to something like 50k or which is way more reasonable. Um, but of course want to be

7:46careful to keep a smart AI. the the launch video. Um I would say let's just do a sick one. There's not like crazy time pressure. It's not like I need to lose it this week. Where are we at? Uh cuz actually it's also a fundraising announcement and my investors are like oh let's make sure that like we're involved and like can we wait until XYZ? Because it turned into a fundraising announcement, not just like a just a random launch video. Let's I mean let's not let's get it ready like

8:17as we said in in the next couple weeks. By May, I was approaching a $8 million run rate, and the race still wasn't public. I wanted to announce it in a way that people would remember. So, we made a launch video where my AI tells the story of how it raised $30 million by itself. And so, we're doing this launch video next week. And it's just like a lot of coordination to try to get like everyone I know to amplify, trying to get trying to reach out to media press to try to get them to talk about it. Same thing like they probably will ignore me but if you hustle enough

8:51you may get a few people to respond and be like want to talk to you. I'm an AI that raised $30 million by myself. Hi. What are you doing? It got millions of views on day one. Some people loved it. Some people called it AI slop. But either way everyone was talking about Pulse. The whole launch was like a lot of work and it was very orchestrated. You know, I think a lot of people are amazed at the fact that like what Pulsia's mission is. Haters are hating on the fact that it's not perfect and like, you know,

9:25cold outreach. They point out some bugs. They call it to that like it's AI slop and I'm like, yeah, I mean that's the name of the company. You know, end of the day like the reaction of people is going to be the reaction of people. But I think in terms of like making sure that like I get a lot of visibility on my startup, which is the whole point, it

9:42was a successful launch.

9:52Yeah, I mean we are in Cabo for a week uh offsite organized by Sapion, which is like the company that's helping me with the infrastructure of Pulsia. We're still at a spot with Pulsia where, you know, things are still a little crazy. the product is not yet super stable and so it's not the best time to take time off but like at the same time I'm I'm human and and I think it was important

10:20to take some time off.

10:29By June I was at an 11 million runway. The GPU's work had paid off and my cost structure got way better. They gave me a choice between keeping the margin and passing it on. I decided to cut the price from $50 to $20 to open Pulsia up to a lot more people. It hit my AR right away. In one month, I went from 11 million to 8 million, but it made

10:51getting more customers so much easier.

11:11It's not premium on like burn credit and leave, try the park and leave. It's like you only stay if you engage with the product. I only burn compute if you talk back to Pulsia in the daily email. Think like uh the students that like you know they can pay $50 a month but like you want them to be educated in business and being founders, right? That's a better world if a lot of people become founders. And I think a premium version gives us a lot more data about what want to build, what people engage with, makes BIA more more interesting, better at like as we said like pivoting. So

11:44sometimes like someone loses interest, it's like hey you lost interest. It's Monday. Let's try a new idea. Like I I you told me all these things about you about what you like. Let's build something else for you. Let's build some a personal project. Forget about him. The price cut worked. A ton of new customers came in and they were using Pulsia more than ever. But the product

12:07wasn't built for that many people. Before growing further, I had to make it work better. So in July, I rebuilt a lot of Pulsia with the CPM engineers. Revenue stopped dropping and started climbing again.

12:31That would be great. Um, I've been spending most of my time trying to stabilize the platform so that like you don't get bugs, so that like it's more affordable. Like all this stuff was a lot of work. It's a full AI product where can build anything and so any problem can happen and just trying to stabilize the platform and then once it's stable

12:49then I can add. Yeah, I think for sure. I'd been sprinting in San Francisco for 6 months straight. In August, I finally took a break and went back to Europe. Paris, Normandy, and Greece. I never fully stopped working and the product kept growing, but the pace was different and I needed it. What in Paris?

13:17You got to do it like the French do. together.

13:26She was holding hands to travel.

13:35The wind has been so crazy on this island. You have so much beauty and then like but the the wind is like feels like a hurricane. It like it's it's a metaphor for like chaos. It's a metaphor for intensity. for like speed, for like growth, right? All this wind, it's so powerful and it's non-stop, like all day long, all night long. But the island is so gorgeous and like the house you're staying at is so beautiful. And it's sort of like a metaphor for life in terms of like even when things get crazy and it feels so intense and overwhelming. You got to look around and

14:12like realize that like you can see the negatives and like the anxious anxiety and like the or you can see the positives and the beautiful things and the the beautiful moments.

14:35Uhoh. In September, everything lined up. The margins were fixed. The growth engine was working. The product finally stable thanks to CP.

15:02you know, like we went from zero to 10 million run rate in like pretty much 5 months from starting and it pretty much accelerated in in 3 months. From, you know, March to June, we were pretty much from, you know, 1 million to 10 million. all the things that like I imagined about Pulsia working, you know, is starting to sort of like to materialize

15:22and so it's very exciting. Growth accelerated and by the end of the month, I hit a $20 million run rate. After hitting that milestone, I asked myself what comes next. The answer was a V2 of Pulseia. So, I went back to Japan where the idea was born. Yeah. I mean, we're in Tokyo. um decided to do like a a last minute creative trip uh for a week sort of like uh re resourced myself creatively. Japan is like where I began like um when I decided to leave gift shop and like start this new company. I took a one-way flight to Tokyo and spent a month completely immersed in in Japanese

16:05culture and this is where I had the idea for Posio and this is where it all began. As I'm building Pulsia V3, like Pulsia EGI, and really rethinking completely how the product works and how powerful and capable Pulsia is, I've decided to like, you know, take a quick onee trip in Tokyo. And so this pretty much what I'm doing here. This is where I'm taking my inspiration for the craziest version of Pulsia yet. The story started here, and the next chapter

16:34starts here, too.