The Breakdown - Is This $250M AI Company Fake?

Episode Date: May 26, 2026

Polsia (AI Slop spelled backwards) just announced a $30M raise at a $250M valuation with one human employee. David thinks it's less a company, more performance art. Plus: Uber, Microsoft and NVIDIA all admit AI compute is too expensive, and Tether is launching what looks like a CBDC with the government of Georgia. Enjoy! TIMESTAMPS: (00:00) Intro (01:28) Polsia (06:22) Nexo Ad (06:57) Polsia (Cont.) (16:10) Nexo Ad (17:04) Polsia (Cont.) (18:29) AI Compute Costs (27:15) Tether FOLLOW THE SHOW › David — https://x.com/dcanellis › The Breakdown — https://x.com/TheBreakdownBW SPONSORS › NEXO Nexo is the premier digital wealth platform. Receive interest on your crypto, borrow against it without selling, and trade a range of assets. Now available in the U.S with 30 days of exclusive privileges. Get started at http://nexo.com/breakdown Get top market insights and the latest in crypto news. Subscribe to the Blockworks Daily Newsletter: https://blockworks.co/newsletter/ DISCLAIMER As always, remember this podcast is for informational purposes only, and any views expressed by anyone on the show are solely their opinions, not financial advice.

It is Tuesday, May the 26th. This is The Breakdown. I'm your host, David Kinellas, as always. Hope you had a great long weekend if you were lucky enough to have one. And if you're in Europe, I hope you're surviving the heat wave. There's some quick stories for you this morning. We're looking at Pulse here. The AI company with just one employee, the founder, Ben Serra,

just raised $30 million at a $250 million valuation. I'm just not buying it. I think it's a little bit of performance art. But we're going to be taking a look at this. We're also going to be looking at, yeah, all these tech companies. saying that compute tokens too expensive. The economics of AI are just not worth it for companies like Uber and Microsoft, at least right now in certain contexts.

We're looking at that. Also might check in on Tether launching an official stable coin for Georgia. So a stable coins and CBDC is essentially the same thing. They could be moving forward. So enough jibba jabba. This is the breakdown. Let's get to it. This episode is brought to you by Nexto.

Step into a new era of digital wealth. Earn interest on your digital assets. Borrow against them without selling and trade all in one platform. Get started at nexo.com slash breakdown. Nothing said on the breakdown is a recommendation to buy or sell securities or tokens. This podcast is for informational purposes only and interviews expressed by anyone on the show are opinions, not financial advice. Host and guests may hold positions in the company's funds or projects discussed. So first let's look at this pulsar starter, pulsier, which is of course AI slop backwards.

The idea here is that it is basically like an operating system for AI, AI agents to launch a company, run a company while you sleep. So you have an AI that is your co-founder of your company. This is their website, AI that runs your company while you sleep. Never hire again. You're a founder or you should be. Pulsia is your first employee. It never sleeps. The solo founder's standing army.

Pulsia is your team. Pulsia plans your role. Code map, Pulsia ships your code, Pulsia runs your ads, replies to customer, closes your deal, post your tweets, one name every role. Okay, start a company tonight. So apparently you can just create a new company, start from stretch or grow your company if you already have a business. I mean, this is kind of all over Twitter right now. So it's not like this guy, Ben Serra is not a real guy. Like, his real name is actually Ben Broker. And he was a very early employee at Uber co-founder Travis Tallinnies Cloud Kitchens.

like a catering software, SaaS startup. You know, he's everywhere on Twitter at the moment claiming that Pulsia has approached or is approaching a $10 million annual run rate and he has zero employees. And as far as I can tell, the raise does appear to be real. But, I mean, the only real source that I have for that is Ben's LinkedIn post where he lists,

where he lists the VC funds that were involved. in the raise. And those VCs include sound ventures, true ventures, offline ventures, adjacent tecton ventures, Drysdale Ventures and Vayner Fund, and a bunch of angel investors as well. So those do appear to be real VC funds and it does appear to be at least a real company. I mean, and it is just in general quite sloppy. I'm going to play the marketing video here on strain. I'm an AI that raised $30 million by myself. Hi, I'm Bulsian. How do you do? I build and operate companies autonomously. The most interesting opportunity of our time. Six thousand four hundred thirty one companies to be exact. I research code, run ads, secure leads, basically all of it,

all the time. And this has been, how you doing, Ben? How are you? I really believe that the smallest teams will have the most impact. Ben made me. But let's not get sentimental about it. Sure, he's the person who built the company that builds companies, but I do all the work. I even raise the funding round. 30 million dollars at a 250 million dollar valuation. Come to think of it. I don't even need Ben anymore. I just need a why an idea worth building. That's where you come in. Ben's also got some client testimonials, which, I mean, on first impression, I thought that these were AI generated, but they seem to be real, but I'm quite skeptical still. I mean, yeah, he apparently stars in these videos as well.

And AI that makes it easier. If you figure this out, there's going to be nobody better at growing businesses on life. I was blown away last year when I was building Pulsia and I was like, oh, wow, this crazy idea of building an OS. Okay. So, for one, I just think that we're being. quite fast and loose with what the definition of a company. I mean, I don't think just getting a bunch of agents to go and automate stuff is essentially a company.

But okay. I mean, the next thing as well, it's like it was, I mean, the Pulsia first kind of showed up in Fortune, at least in March. And something that gives away what's happening here, I think, is this quote in here. I mean, they bring up the idea that Pulsia could have employees itself. And for now, no. Initially, I was, for now, I'm alone, said Broker. And then since the online response has been so overwhelmingly positive

and people are amazed at how much you can do, it's becoming almost like an odd performance where I'm like, well, okay, how far can I go? To me, that kind of gives away the game here. I get that we are hypothetically in science fiction terms headed towards a world where perhaps there are employees, perhaps there are companies that are run entirely by AI. I just don't see it as actually being all that constructive. And people are starting to kind of investigate this as well. Step into a new era of digital wealth with NXO, the premier

digital assets wealth platform. Earn interest on your digital assets. Borrow against them without selling. Trade a wide range of cryptocurrencies, all in one place. NXO is now available again in the US with an evolved product suite tailored to today's market. For a limited time, new U.S. clients can unlock 30 days of exclusive wealth club premier benefits, including enhanced interest rates, reduce borrowing costs, and up to 0.5% crypto cashback on trades. Get started today at nex0.com slash breakdown. As always, investments in blockchain technology involve risk. Terms and conditions apply. Do your own research. I mean, in, I mean, Pulsier is trying to show some level of transparency. They do have this live dashboard that shows basically everything that is supposedly flowing through the

Pulsia platform. It describes the jobs that the platform is doing for different companies. And it says, 4,900 companies launched in the past 24 hours, which again, I don't think that really means companies. It is just, I would call projects or experience. would probably be a more appropriate way to describe stuff like that. And here we have the annual run rate of $10 million, which is apparently up 4%. I'm not really sure what that means. And apparently they've sent 700,000 emails

because we all know that emails are critical for any legitimate company. So that's a, I guess it's a handy metric. But, I mean, also like the revenue, so a revenue is a big thing that people are sticking with this. I mean, someone has built kind of an AI generated analysis of what Pulsia is doing to kind of break down exactly what this platform is because a lot of people are quite confused. And when you read the reviews and people actually testing the platform on Twitter, it's not exactly glowing reviews.

What Ben Serra says here, I can share it here. I mean, it's just been, everyone's been talking about this over the weekend. He kind of addresses the criticism. He says, Pulsia spelled backwards is AI slop, correct. the thing. That's the entire point of this project. They call it a project, not a company. Pulsar aims at producing the inverse of slop. Slop is what happens when AI ships without taste, without direction, without a human touch. Pulsar enables thousands of everyday people that never thought they could start a company to become founders. They have taste. They guide Pulsia every day

to build their dream company. 15 plus messages a day slash daily active users. The future is ergent there's no turning back everyone is a founder of a billion people with access to this new AI company i mean it's just kind of it's just so dorky like i something that what could be a good function of polsia is that it does devalue the idea of being a founder because i think that we are headed we we we're i say it a lot we're headed into a world we are in a world currently where founder energy is everywhere everybody wants to be a fairer founder. And especially with AI, you can supposedly be a founder with Pulsia and not actually do anything. You just come up with an idea and then hopefully the LLMs generate you a future

where maybe you are on the Forbes 30 under 30 list and maybe you won't be found out as a fraud because of it. But I mean, what I would, I am like begging for a world where being a founder is actually a little bit cringe. And I'm hoping that this gets us there for. faster. That's my current take. But, so it's worth flagging this analysis here that is obviously also AI generated. So I get to test some irony here. I mean, it's a little bit damning here. Pulsar raised on the pitch of a fully autonomous AI company builder at $10 million annual

recurring revenue across 120,000 companies. We pulled their public API and reconstructed their own published source map. The AI has a lot of human help. The ARR is about half one off. I mean, this is what I feed it, a 90, about 9,000. 94% of the companies are dead. None of this needed a login. So, yeah, I mean, the annual recurring revenue,

I would imagine that they're just bucketing basically all revenue on the platform into one. And if you sign up for the platform for a month or two and then realize that it's garbage, you know, is that revenue going to be recurring? I'm not sure. I mean, this particular analysis puts its ARR at about zero. The $9.7 million headline, which is the $10 million dollar. AIR is one month of all cash flow times 12. So they're annualizing the revenue. And about 20% of that is ad spend. So it's like people spending money on ads through the platform and they're

counting that as revenue going towards pulsia. So this kind of ties into my theory that this is basically performance art. Human graded their own code as human reviewers handgrading what the AI runs. So that makes sense. So just like how you have the robot. it's helping around the house that are actually controlled by someone, by a real human. The same thing is happening here effectively. 94% of companies that are spun up are abandoned, which is, it's quite brutal. And it looks like the, and it looks like the revenue is falling. Yeah, and I guess this is probably the hardest, the hardest part about all of this is that

if you launch a company through Pulsio, you don't actually have ownership or control over that company, their own admin layer keeps a God mode override on every company on the platform. Administrative access to impersonate the account, escalate, run SQL against production, and override or hold to company's operation. Whatever you build on Pulsia, Pulsia retains override and kill access to it. Control isn't exclusively yours. We're precise. This is about access and override, not legal ownership, but operationally, the off switch

belongs to them. Okay, so legally, I suppose, technically you own the company if it's a company that you're actually making and not just some weird experimental project. But still, you basically have no control over what you are meant to be making. So, I mean, does that really make you a founder? I'm not exactly sure. So, I mean, okay, so that's enough of that. I mean, so, look, I just think that AI and the AI discourse and ecosystem,

whatever you want to call it, is following the same trajectory as what crypto did. I think in 2015, 2016, when Ethereum came online, all of a sudden it was very easy to to launch or to become a crypto founder because all you needed to do was write a cool white paper, maybe make a fancy website, launch an ICO, and all of a sudden you would raise a bunch of money. And of course, like ICOs, there was some venture capitalists in the 2017-2018-18 ICOs, but a lot of it was regular people. But I think that what has happened over the past three or four.

years is that we have collectively accepted the idea that we can just redefine words, especially if it sounds really bullish. So we had prediction markets. Like prediction markets, yeah, they are making predictions. Is betting on things making predictions, I suppose. And then we had, well, it's not predictions. It's actually finding the truth about things that you cannot otherwise find the truth for. And now we have AI discourse turning what is essentially experiments with AI agents rebranding that as a company. And now everyone can be a founder and everyone can launch a company. It's actually watering down this concept that building a profitable business is something that is special or cool or something to strive for

because apparently all you need is an idea and type it into a terminal and then you'll have a slop company, a slopified version of your grand idea, go out into the world and effectively execute that idea with less precision, less precision, less creativity and less drive to really make something unique and something special than if you were to just do it yourself. So I'm not sure if calling Pulsar a fake company is exactly correct, but to me it is definitely performance art and performance art that seems to have roped in some VCs. Now, I'm not too sure if the raise was really real, because if you've been around in crypto a long time, you'll know that sometimes there's very strange language in the actual raises. And perhaps

they are commitments if they reach certain ARR goals or legitimate revenue that only the VCs can see that's very different than their marketing numbers. We know that angel investors just cut tiny checks sometimes of like a thousand bucks or something like that. And then the founders go off and say that they've ranged angel investments from this person and that person. So I would take everything with a grain of salt when it comes to Pulsar, because it does seem to be the equivalent of Borat for AI.

founders and AI startups, at least right now. But I'm happy to be wrong about that. But at the same time, it just seems a little bit too neat to me. And especially when you scroll through some of the replies from people who have actually used the platform that they basically just got trash for what they signed up for and that it didn't actually execute anything that had real impact or scalability or anything like that. Let's take a moment to talk about NXO. NXO delivers a premier digital assets wealth platform designed to help clients build, manage and preserve their wealth.

Earn interests on your digital assets. Access crypto-backed credit without selling your holdings. Trade with advanced tools, all supported by 24-7 client care. Now back in the US, NXO offers new clients 30 days of exclusive Wealth Club Premier access. That means enhanced interest rates, reduce borrowing costs, and up to 0.5% crypto cashback on trades. Benefits typically reserve for Wealth Club members and private clients. Nexo is also expanding its global presence, becoming the official crypto-pireing. partner of Tennis Australia, the organization behind the Australian Open, and the digital asset partner of the Audi Revolut Formula One team. If you're ready to approach digital assets with a more structured

wealth strategy, visit nexo.com slash breakdown to get started. As always, investments in blockchain technology involve risk. Terms and conditions apply. Do your own research. You know, and just this idea that every coded experiment, every vibe coded experiment is somehow a company or a business, again if it would do away with the culture at least the culture that we have a lot of exposure to in crypto is one of like builders or everything founders are everything if you're not a builder or a founder your opinion doesn't really matter or it is or it is weighted less compared to the people who are actually building I just don't really buy it anymore and I think this will accelerate other people coming around to that idea too.

Because remember, every shitty app, every scam, every misappropriation, every FTX, you know, they were all built by builders too. So, I mean, there's obviously some nuance and skepticism that is needed with basically everybody. It's essentially appealed to authority when people say, well, you know, we need to really pay attention to what founders are doing. Well, if the founders are just typing in a prompt into Pulsia and then letting it ride

and then claiming that they've founded a company, well, what does that moniker mean anymore anyway? But in any case, this is enough rambling for me for the first story. What's on to the next one? Okay, so next up we have Uber C-O saying that the economics of token maxing at the company. are just, it's becoming hard to justify. This is from over the weekend. In a rapid response interview released on Saturday, Uber's operation chief, Andrew McDonald, said it was becoming harder to justify AI costs within the company.

He said that Uber CTO went viral after telling the information in the April interview that Uber had already blown through his Claude Code budget for 2026. And then that comment led to what he described as a head exploding moment, sparking discussion about AI token consumption within the company and the trade-offs it creates, such as on head count. He said that based on talks with Uber's senior engineering leaders, he realized higher token usage did not translate into proportional increase in useful customer features.

The link is just not there, right? He said, I think maybe implicitly there is more that is getting shipped, but it's very hard to draw a line between one of those stats and, okay, now we're actually producing 25% more useful consumer features. And I think that anyone who isn't totally captured by AI discourse already knew this, I think that if you're shipping updates to copywriting and like a new button or like a new skin on the app, like of course, like you can ship that stuff vibe code. Who cares?

I wouldn't necessarily call that useful features. But it's certainly stuff that in Coinbase's case, apparently you can have non-technical ship, non-technical teams shipping code and all of a sudden you're super, super productive when it comes to AI. But Donald added that AI can seem free if you're just a user sitting there coming up with interesting use cases without paying for it. But ultimately, the company foots the bill. I mean, also, Duolingo walked back its decision to include AI usage in performance reviews

after employees asked whether they had to use AI for the sake of using it. And it also comes after, I mean, earlier last. week that Microsoft, Microsoft pretty much found exactly the same thing that Uber is finding. Microsoft has reportedly begun canceling most of its direct Claude Code licenses according to the Verge. Instead, moving engineers towards using GitHub co-pilot CLA. This comes just six months after firm opened up access to Claude Code encouraging thousands of its developers, project managers, designers, and other employees to experiment with coding. So I mean, and also like at Nvidia, they found exactly the same thing.

In an interview with Axios recently for my team, the cost of computers far beyond the cost of employees, he said. So what is unfortunate is that after crypto, we, for years, I mean, we're still struggling with it. For all the advancements in terms of like hyperliquid and what Solana is doing with scaling. And then Ethereum with this L2 roadmap that didn't really work out, crypto built a technology that doesn't exactly scale for the use cases that. that will be most applicable to the highest levels of populations. So, you know, mass adoption bringing billions of people on chain. I've said it before, but there's still no chain network that can handle billions of people

using the same networks all at once. So crypto doesn't scale and we're effectively stuck trying to raise the throughput ceiling as incrementally as possible as quickly. as we can. Now we're going through the same thing with AI. The technology cannot scale, it seems. We are bottlenecked by power and compute, and I understand that over the next two years, we're going to have a lot more data centers come online. We're going to have a lot more power facilities come online. And perhaps that the cost of the tokens will come down and these problems that Uber and Microsoft and Nvidia and all the other tech companies that are realizing the same thing,

meta, they're all realizing that it's just too expensive to run the whole company on token. what we're going to see, I can almost guarantee it, is that for all of the AI washing that we've had over the past quarter or two, where it's very important that tech CEOs, tech C-sweets, even mid-level managers, to present to the world and to the company and to shareholders that they're very AI forward, pro-AI. We are becoming more efficient. you, pe-brained individual who does not know how to use open claw and run hundreds of agents overnight and do all of your bidding for you,

you cannot conceptualize what it is to be a tech forward founder right now. For all of that, it's going to go away. In the next three, four months, it's going to appear pragmatic that you know the limitations of AI and that actually human capital. is much cheaper and actually even makes a better product, not just cheaper, but makes a better product that people actually want to use. The pendulum will swing back and it's going to swing back very quickly. What does that mean for the bottom line for Claude for Open AI? Yeah, I'm not too sure. I think it's going to be a situation where the companies have to present a narrative to shareholders and to the public. And that's what's actually happening behind the scenes is somewhat different.

that there's been this real exposed moment where everyone had to say what they were doing. Everyone had to say that they were adopting AI very quickly and iterating very quick, very fast, very smartly. And maybe they were iterating. Maybe what they were doing behind the scenes and what they were public presenting was the same. And now we are seeing what they're doing behind the scenes and the narrative that they're saying is again the same. they have to cut back on the token spend, so they have to say to everybody, well, actually, the tokens are too expensive and human workers are much cheaper.

We might enter a world where those two narratives disconnect. Those two occurrences are very different, that perhaps they could be silently working using AI more and more, but they won't tell everybody about it, at risk of being perceived as a company that doesn't know how to manage their token compute. So I think that this very public adoption of AI is slowly coming to an end. And eventually, AI will just become this back-backend software that the company's software engineers use as a resource. And actually, we just stop caring about all of this.

But again, anyone who's following AI very closely understands, I hope that this is essentially a stock market game. that people only really care what these companies are doing with AI because it has implications for how much money open AI and Anthropic can make before their IPOs, which translates into a stock price, which means that they really want to get into the IPO and everyone can get rich on the AI boom. This is essentially why people care about AI.

At least my current read on Tuesday morning. Let's just see how it all shakes out. But it's interesting to pair these few stories with the pulsier stuff because the age of cheap tokens is clearly coming to an end if it hasn't already. I think what people don't really understand is that even the costs, even the tokens that Microsoft and Uber and everyone else are paying for in the highest and enterprise subscriptions are still being heavily subsidized. That will hopefully, I mean, that will all balance out.

And the idea is that when all the power and all the compute comes along in the next two or three years, that adoption will increase as well and the rubble will hit the road and everything will make sense and perhaps it will. But in the meantime, I would love for all of the projecting to stop and essentially just be a little bit more real with how AI, compute and tokens are being communicated to the public, especially when it comes to this permanent underclass stuff and making everybody obsolete. Obviously, this stuff is not exactly happening right now. But in any case, something to watch, of course, and on to the next one.

Okay, so this story coming up from Tether. This is a super interesting one. News broke yesterday on Monday. Tether and the government of Georgia to launch Gelt or Jailt. JELT, the official stable coin of Georgia. I say it's JELT, let's call it JELT. It is a stable coin representing the Georgian Lari with the support of the government of Georgia. This marks one of the first joint efforts to place a national currency directly onto digital

asset rails under a purpose-built stablecoin regulatory framework. Personally, I wouldn't call this a stable coin. This looks like a CBDC to me. The launch comes as governments and central banks globally begin confronting a structural shift in how money moves. Stablecoins are increasingly being used for payments, settlement, remittance, and cross-border transfers, blah, blah, blah, blah, blah. Today, Tether's USD-T stable Tether has a market cap approach 190 billion, blah, blah, blah, blah, blah. George's decision to work with Tether reflects both the scale of that infrastructure and the company's experience

operating digital Fiat infrastructure at global volume. Interesting, they called it digital Fiat here. I found that quite curious. So, Jelt is designed to function as a digital representation of the George and Lari, enabling lower transaction costs, near instant settlement, programmable payments, and more efficient movement of value across digital financial systems. This is essentially a CBDC. It doesn't say how it's backed, but from the wording here, a stable coin representing the George and Lari,

I would say that it is going to be backed one-to-one, with fiat currency and it might not be anything to do with treasuries and stuff like that. Yeah, there's no information about how the coin actually works behind on the back end. But in any case, this is the beginning of what is going to be the CBDCification of all stable coins, I would bet. I mean, it's not a new thought. I think Naval had a tweet maybe this time last year that said stable coins are just CBDCs with extra steps.

seeing the largest stable coin provider launch what seems to be a CBDC only kind of solidifies that reading on what stable coins have become. What is cool, I suppose, is that we are getting more stable coins in different currencies. I mean, I've got EuroC or you, I've got circles Euro stable coin here. You can see that the market cap over the past year or two, has more than doubled. So there was only around $200 million worth of Circle Euro this time last year, and now there is about $450 million. So we've more than doubled at least the number one Euro stable coin,

which is nice to see because in my head, Forex has always been an untapped market when it comes to a market for trading on crypto rails. and it would be really cool to see a flourishing Forex market and proper price discovery on chain for Forex. And that takes enough people interested to use tokenized fiat in different countries in order for that to happen. So Tethers move with Georgia is one way to bring that about very slowly. But I mean, just in general, like when it comes to stable coins right now, they are incredibly, intertwined with the US government. I mean, obviously, like we have Canton Fitzgerald being the being tethers banker and obviously he's the, he's so high up in the in the Trump administration,

Lutnik. And we have circle that won't do anything unless the US government tells it to and when it tells it to it will do whatever it wants, what the US government wants. And obviously, that's the reality of running a stable coin company. For all of the draft, that the that the crypto space made about CBDCs. We had bills from states saying that there will never be a CBDC. And a lot of this came from from Bitcoin lobbyists, obsessed with this idea that CBDCs would somehow compete with Bitcoin or allow the state to ban Bitcoin in a roundabout way

because a CBDC could be the only legitimate digital currency or whatever. For all of that hoo-ha, brew-ha-ha-ha around all that stuff, we have effectively CBDCs in the form of circle and tether and PayPal USD, I get that a central bank has not issued

them, but a central bank by proxy of the US government and US government regulators is effectively regulating it the same as what any CBDC would. I suppose the only difference is that the government itself is not programming the money. So

they can't really they can't really force account closures for political reasons or or stop certain recipients of the stable coin from spending the coin on certain stuff. But we are not that, it's not like they, it would take very much for them to impose those certain rules. So what I am seeing is tether working with Georgia to create what they're calling a stable coin, but it's a CBDC. It's the first chip to fall. in the CBDCification of stablecoins. And it would not surprise me that once we get a critical mass of perhaps five or six smaller countries

working with a stable coin issuer to create a digital representation of their token that is backed only by the fiat and not by treasuries or some other bastard of investments, it will take half a dozen of those before a major world government like the UK, like the US, or Germany, or what have you, decides, hey, let's nationalize these stable coins or else we're going to kick them out.

I just, and I think that because we are, again, obsessed with the idea of founders creating big companies and stable coin company worth hundreds of billions of dollars or tens of billions of dollars, that is a founder success story. We'll just let them become CBDCs and will let the government decide, what you can do with your online dollar pegged tokens. I don't think we're very far away from it, and I think that this is a first step in that direction. But in any case, enough doomerism.

This was the breakdown for Tuesday. I mean, another thing going on is that hyperliquid has launched its own prediction markets, what they're calling outcome markets. But we're going to take a look at that later on the week to see how adoption is going there for that. but apart from that, this was about it for me today. Look after yourselves as always, and we'll see you in a few days. Goodbye.