The AI Money Trap, Part One

This is an iHeart podcast.

On Fox One, you can stream your favorite news, sports, and entertainment live, all in one app.

It's fing roll and unfiltered.

This is the best thing ever.

Watch breaking news as it breaks.

Breaking tonight, we're following two major stories.

And catch history in the making.

Gibby, meet Freddy!

Debates,

drama, touchdowns.

It's all here,

Hi, I'm Morgan Sung, host of Close All Tabs from KQED, where every week we reveal how the online world collides with everyday life.

There was the six-foot cartoon otter who came out from behind a curtain.

It actually really matters that driverless cars are going to mess up in ways that humans wouldn't.

Should I be telling this thing all about my loved life?

I think we will see a Twitch streamer president maybe within our lifetimes.

You can find close all tabs wherever you listen to podcasts.

There's more to San Francisco with the Chronicle.

More to experience and to explore.

Knowing San Francisco is our passion.

Discover more at sfchronicle.com.

What happens when Delta Airlines sends four creators around the world to find out what is the true power of travel?

I love that both trips had very similar mental and social perks.

Very much so.

On both trips, their emotional well-being and social well-being went through the roof.

Find out more about how travel can support well-being on this special episode of the Psychology of Your Twenties, presented by Delta.

Fly and live better.

Listen wherever you get your podcasts.

Coolzone Media.

Hello, and welcome to Better Offline.

I'm, of course, your host, Ed Zitron.

As ever, go to the episode notes and buy some merchandise.

We have the beautiful challenge coin available until August 24th, I want to say.

Subscribe to my newsletter or send me a threatening email.

I love hearing from you all and I also love to see that our subreddit has gone over 12,000 vagabonds and scallywags.

I'm so proud of our community.

Anyway, this week's going to be a fun one, a two-parter plus a special report on GPT-5, and it's kind of a culmination of the last year and a half of work that I've done covering the AI bubble.

Now, the reason I return to this subject so often is because each week I become more and more convinced that the only way the AI bubble ends is, well, badly.

Every week I see more and more evidence that this is structured in a way that is doomed to end, not even abruptly, but there are going to be abrupt little farts, little burps as the bubble begins letting out air, and in a way where I see people in innovation as the collateral damage.

In the global financial crisis, the Federal Reserve and the US government as well as other equivalents around the world worked to ensure that when a bank or hedge fund went under or a big insurance company became insolvent the collapse was managed in a way that didn't implode the entire financial system.

The difference with AI is first that it's nowhere near as essential as financial services to the global economy and thus unlikely to be saved as AIG was in 2008, but also that there's no real way for this to end gracefully.

Just due to the way all of this is happening, how it's built, I'll get into it.

And all the signs are now there, and I have been hesitant to make this call the whole time, by the way, where, well, there's a contagion here, where one precarious company falls, the rest eventually tumble, sparking a chain of events that will be, well, pretty bad.

But how does this end?

Okay.

At the start of August, we saw no less than three different pieces, two in the Wall Street Journal and one in the Financial Times, Financial Ninety Times, eh, who cares, published within the span of one week, all asking the same question, whether the massive proliferation of data centers is a massive bubble.

And honestly, since then, there have been several more.

Although these publications at times seem to have taken the default position of AI's inevitable value, they've begun to sour on the idea that it's going to happen anytime soon.

To be clear, these are sensible, well-respected business newspapers, hardly partisan voices from the AI wars.

Separately, CNBC reported that Quirktop 300 core on OpenAI has either raised or is about to raise another $8.3 billion in cash, less than two months since it raised $10 billion from SoftBank and a selection of other venture capital firms.

While this sounds ambiguous, this reporting only claims that it has secured that funding, not that it's received it, which is an important distinction I pointed out in the past, as secured funding often comes with its own conditions attached.

On top of this, I should add that there is currently a rumor that OpenAI insiders, the people working at the company, are selling $6 billion of stock at a $500 billion valuation.

That's around the price of Netflix.

This is fucking stupid.

It's so stupid.

And every time I read about it, I feel a little insane.

But it's also weird that you've got people like SoftBank and Dragoneer, who invested in the $8.3 billion round, who are also buying the stock.

All this sounds stinky.

And I hate to be crude.

I hate to be too crude here, but where the fuck is all this money going?

Is OpenAI just incinerating capital?

Is it compute?

Is it salaries?

Is it more compute?

Is it data centers?

Because SoftBank isn't actually building anything for Stargate.

It's all...

Okay, I'll calm down a little bit, but it's all getting a little bit silly.

Now, the information, which has some of the best sourcing in any publication covering the AI bubble, I'd argue, though I've taken issues with their reporting in the past, suggested that OpenAI intends to use this money to build data centers, possibly the only worse investment it can make other than generative AI.

And it's the one that...

OpenAI can't really avoid because they're also somehow running out of compute too.

And amongst this already ridiculous situation since the issue issue of OpenAI and Anthropic's actual revenues, which I wrote about in my premium newsletter on the 1st of August, please give me money, and have roughly estimated to be $5.26 billion in the case of OpenAI and $1.5 billion in Anthropic as of July.

Now, to be clear, there are very different numbers going around, I'll get to in a second.

In any case, these estimates were made based on both companies' predilection for leaking their annualized revenue, or month times 12.

Now, this extremely annoying term, annualized recurring revenue, is one that I keep bringing up because it's become the de facto way for generative AI companies to express their revenue.

And both OpenAI and Anthropic are leaking them intentionally and doing so in a way that suggests that they're not even using the traditional ways of calculating them.

Because ARR isn't in and of itself an evil thing, it's fairly standard within software as a service companies, except they run completely different to OpenAI and Anthropic.

Anyway, OpenAI leaked on July 30th, 2025, that it was at $12 billion of annualized revenue.

So it earned around around $833 million in some sort of 30-day period.

And then two days later, on August 1st, 2025, the New York Times reported it was at $13 billion annualized, or $1.08 billion of monthly revenue.

It's taking the piss a little bit.

It's very clear OpenAI is not talking in actual calendar months, at which point we can assume they're using like a trailing 30-day window, as in the month is just 30 days rather than a calendar month like May or June or July.

We can, however, declaratively say that it's not saying the month of June or the month of July was 12 or 13 billion dollars annualized because if it was, they wouldn't have given two vastly different goddamn numbers in the same two-day period.

It doesn't make any sense.

And to be clear, while I can't say for certain, I believe these leaks are deliberate.

OpenAI's timing matches exactly with fundraising.

Similarly, on Anthropic's side, these revenues are beginning to get really, really weird.

Anthropic went from making around $72 million, $875 million annualized in January, to $433 million in monthly revenue in July, or at least it leaked on July 1st, 2025, that it was at $4 billion annualized to the information, which is $333 million, and then claimed it had reached $5 billion annualized, $416 million, to Bloomberg on July 29th, 2025.

I'm so tired of this.

Something stinky is happening.

The people who will not admit something stinky is happening, I'm beginning to question their sanity or whether they're just inherently corrupt in some way.

And you can be corrupt without money.

You can just do it for your mates.

Anyway,

how did Anthropic get there?

I'm guessing it was from cranking up prices on Cursor, a product made by a company called AnySphere that lets developers generate code using Anthropic's models and other models.

And we've had the confirmation that's the case, thanks to the information reporting that $1.4 billion of the annualized revenue from Anthropic is from its two top customers.

So around $116 million in a month, and the biggest of which is Cursor.

Confusingly, the information also says that Anthropics Glawed Code is generating nearly $400 million in annualized revenue, roughly doubling from just a few weeks ago per their report.

It's around $33 million of monthly revenue.

They really jacked that product up.

I need to do a whole thing on Glawed Code, if I'm honest, because that whole thing has become a complete nightmare.

They've had to completely, the weekly rate limits are coming.

God, what a fucking mess.

Every time I think about this company, I get upset.

In any case, I think Cursor is a huge indicator of the current fragility of the bubble, and the fact that for most AI startups, there's simply no way out because being acquired or going public does not appear to be a viable route.

But before I explain, please listen to one of the many stable, normal advertisers that gives money to the show.

Not to me, though.

I don't get the money from the ads.

There's more to San Francisco with the Chronicle.

There's more food for thought, more thought for food.

There's more data insights to help with those day-to-day choices.

There's more to the weather than whether it's going to rain.

And with our arts and entertainment coverage, you won't just get out more, you'll get more out of it.

At the Chronicle, knowing more about San Francisco is our passion.

Discover more at sfchronicle.com.

Be honest, how many tabs do you have open right now?

Too many?

Sounds like you need close all tabs from KQED, where I, Morgan Sung, Doom Scroll so you don't have to.

Every week, we scour the internet to bring you deep dives that explain how the digital world connects and divides us all.

Everyone's cooped up in their house.

I will talk to this robot.

If you're a truly engaged activist, the government already has data on you.

Driverless cars are going to mess up in ways that humans wouldn't.

Listen to Close All Tabs, wherever you get your podcasts.

In business, they say you can have better, cheaper, or faster, but you only get to pick two.

What if you could have all three at the same time?

That's exactly what Cohere, Thomson Reuters, and Specialized Bikes have since they upgraded to the next generation of the cloud.

Oracle Cloud Infrastructure.

OCI is the blazing fast platform for your infrastructure, database, application development, and AI needs, where you can run any workload in a high-availability, consistently high-performance environment and spend less than you would with other clouds.

How is it faster?

OCI's block storage gives you more operations per second.

Cheaper?

OCI costs up to 50% less for computing, 70% less for storage, and 80% less for networking.

Better?

In test after test, OCI customers report lower latency and higher bandwidth versus other clouds.

This is the cloud built for AI and all your biggest workloads.

Right now, with zero commitment, try OCI for free.

Head to oracle.com slash strategic.

That's oracle.com

Every business has an ambition.

PayPal Open is the platform designed to help you grow into yours with business loans so you can expand and access to hundreds of millions of PayPal customers worldwide.

And your customers can pay all the ways they want with PayPal, Venmo, Pay Later, and all major cards so you can focus on scaling up.

When it's time to get growing, there's one platform for all business: PayPal Open.

Grow today at paypalopen.com.

Loan subject to approval in available locations.

And we're back.

So, how exactly is one AI-powered coding environment such a load-bearing part of the AI trade?

We're going to be focusing now on both the fact that cursor is a systemic risk to the AI industry, one where, if it fails, could make every other player that bit less secure, and the fact that cursor is emblematic of what makes generative AI companies so paradoxical, where you can have bonkers valuations on one hand, but also no way for these companies to go public or sell to someone else and justify the actual valuation itself.

I know it sounds a little insane, but I believe that Cursor is the weak point of the entire bubble, and I'm going to explain why and how this could go.

This is by no means inevitable, but I cannot work out what Cursor does other than this.

Cursor makes, before at least the massive changes to its service, though Tom Dutan over at Newcomer suggests that the revenue is still going up for now, they make about $500 million in annualized revenue, so around $42 million a month.

This makes it the single highest earning generative AI company that isn't called Open AI or Anthropic, and the highest earning company built on top of primarily Anthropic's models.

Its success is symbolic to the greater AI movement and just as it hit its peak, Anthropic and OpenAI to a lesser extent, though they seem more interconnected during GPT-5's launch, decided to add priority processing and priority service tiers, demanding more money up front and causing Cursor to have to massively degrade its service to keep up.

I also have a premium piece that explains this from a a few weeks ago.

I did a monologue too, and it's over at wheresyorehead.app.

Please give me money.

To explain in short, Cursor's AI-powered coding editor used to have fairly unrestrained access to the various models provided by companies like OpenAI and Anthropic.

In mid-June, a few weeks after Anthropic introduced priority tiers that required enterprise companies to pay upfront and guarantee a certain throughput of tokens and increase costs on prompt caching, which is a big part of AI coding, Cursor massively changed the amount its users could use its product and introduced a $200 a month subscription.

As an aside to this, Anthropic also competes with Cursor's AI coding product with Claude Code, its own service.

I've talked about it in the past and I'll talk about it some more in this series too.

Cursor, as Anthropic's largest client, the second being GitHub Copilot, represents a material part of its revenue, and its surging popularity meant that it was sending more and more revenue to Anthropic.

Anthropic used this opportunity to raise prices on accessing its models to continue providing service at an acceptable level to Cursor's customers by introducing priority tier access on May 30th, 2025.

This has allowed Anthropic to juice its revenues and due to the upfront nature of these contracts, Cursor is locked in regardless of how well it does.

The net result of these cost increases means that Cursor's product is less attractive to its customers and will thus eventually make it less money.

At this point, one has to ask, how does Cursor survive?

Its product isn't profitable and the means it used to make the company so successful have become untenable.

It has guaranteed a certain throughput of tokens per second to the major model developers, chief of the Manthropic, but Cursor itself said in the whole cursor ultra $200 a month thingy that it signed multi-year deals with multiple cloud providers like OpenAI, XAI, and Google.

Cursor's product is now worse, by the way.

They also have made it so the auto model that used to be unlimited will now, as of September 15th, charge a fee.

It's so fucking funny.

I love this shit.

It's so funny watching this happen.

But really, people are going to cancel their subscriptions.

Their annualized revenue will drop and their ability to raise capital will suffer as a direct result.

Cursor will, regardless of this drop in revenue, have to pay the cloud companies what it owes them as if it had the business it used to because it was a guaranteed throughput.

I've spoken to a few different people, including at a company with an enterprise contract, that are either planning to cancel or trying to find a way out of their agreements with Cursor, and they were planning to do so as recently as April.

I'm sorry, recently as April.

I mean, they were planning to do back in April before this shit happened.

I'm not editing it.

You understand.

Anyway, if Cursor is allowed to die, it will be unable to pay a chunk of Anthropic and OpenAI's revenue.

And yes, the revenue of people like XAI and Google as well.

It also brings into question whether it's possible to build, putting aside any questions of profitability, a business of any kind offering services built on top of generative AI models, and in turn bring into doubt the veracity of investing in this sector writ large.

It will also call into question whether any other generative AI company is a real business.

This will naturally lead to the question of why we're building all these goddamn data centers.

Cursor at this point faces two options: die or get acquired.

This is not an attack on anyone who works at the company, nor anything personal.

The unit economics of this business do not make sense, and yet, on some level, its existence is deeply important to the Valley's future.

And a large chunk of Anthropic's revenue, I should add.

And Tom Dutan over at Newcomer added, something like 10% of Anthropic's revenue comes from Cursor.

This is so bad.

But I'm gonna humor this.

I'm gonna humor this big time.

Who would actually acquire Cursor?

Maybe OpenAI.

But they couldn't acquire Winsur for another AI company because they were too worried that Microsoft was going to get the somehow essential IP of what appears to be one of 100 different AI-powered coding environments.

They also already trade and failed to buy Cursor.

And if I'm honest, I would not be surprised if Cursor would sell now.

I don't know if OpenAI has the money, but you know...

They probably would sell.

Honestly, Cursor fucked up bad not selling to OpenAI.

They could have got $10 billion and Sam Altman would have had to accelerate the funding clause.

It would have been so goddamn sick.

But now the only sick here is Kercel's fragile, plagued business model.

But how about Anthropic?

I don't know about that one either.

They've already got their own extremely expensive coding environment, which I estimate in my premium newsletter loses the company 100% to 10,000% of a

subscription per customer.

That's like a monthly subscription just from one user.

And I did that a few weeks ago.

And now Anthropic, as of August 28th, is adding weekly limits on accounts, which I believe creates some of the most gnarly churn in SaaS history.

It's going to be quite nice.

Also, does Anthropic really want to acquire its largest customer?

Also, with what money,

they're not raising $5 billion to bail out Cursor.

Anthropic needs that to feed directly into Andy Jassy's asshole.

pocket to keep offering increasingly more complex models that never quite seem good enough to make any money.

But how about Google?

They no,

they just sort of bought Windsurf, but I'll get it's a complete mess, and they can't do that again.

They've already given out the participation trophy multiple billions of dollars to investors and founders, so nobody has to get embarrassed about this.

And then allowed Cognition to pick up the scraps of a business that made $6.83 million a month after burning $143 million of investor capital.

And TechRunch reports that Windsurf was left with about $100 million in cash post-acquisition.

TechRunch also reports that Cognition paid about $250 $250 million for what remained of the company.

Google paid like $2.shothing billion dollars to just take the founders and Cognition picked up the rest.

And this deal, by the way, didn't actually pay out the majority of Winsurf's employees.

I'll have the links to the links to all of this in the notes.

This whole deal was real sickly though.

The investors got paid out, the founders got paid out, the employees got fucked.

It really sucks.

I'll get into it in a bit.

But let's go with Meta.

Here's the thing.

If I'm Michael Trull, the CEO of Cursor of Any Sphere even, I am calling Mark Zuckerberg and I'm pretending that I think the only person in the world who can usher in super intelligence is the guy who burned more than $45 billion on the metaverse and believes that not wearing AI glasses in the future will be a disadvantage.

I would say all manner of shit about the future and that the only way to do this was to buy my AI-powered coding startup that literally cannot afford to exist.

I would tell Zuckerberg anything.

I would be like, mate, I just watched the actual Stargate.

I think you're the people from Stargate.

He's never going to to watch Stargate.

He's just going to be like, oh, yeah, I've heard that's a TV show.

Wow.

Great.

And then you get however much money you need, just $20 billion.

He'll just pull that out of his pocket.

He hasn't.

He has 20 billion.

He'll just hand it to you.

He's probably got it in his like, in his safe.

Just go and ask.

How much money do you think Mark Zuckerberg's got in his safe?

Email me.

Email me with the guess.

Anyway.

This whole can't afford to exist thing that really is the problem.

These companies are all going through the same motions that every company before them did.

Raise as much money as possible, get as big as possible, and eventually scale to the point that you're fat with enterprise cash.

Except the real problem is that just like the big tech's new gluttony of physical real estate it's taken on, generative AI companies are burdened with a constant and aggressive form of cloud dam.

The endless punishment of the costs for accessing the API for generative AI models that always seem to get a little bit better, but never in such a way that anything really changes other than how much Anthropic and OpenAI are going to need at the end of the month or they break your startup's legs.

I'm not even trying to be funny.

Anthropic raised its prices on cursors so severely it broke its already unprofitable business model.

These products, while also for the most part not producing that much revenue, need to be sold with users being aware of and or sensitive to the costs of providing them.

The cursor's original product was a $20 a month service for 500 fast requests of different models in the same way that accessing clawed code on any subscription is either $20, $100 or $200 a month rather than paying per API core because these companies all sell products that shield the customer from the actual costs of running the services.

And those costs could be vast.

One article, which I'll link to in the spreadsheet, argues that given the trajectory of the development of AI models, it's not unreasonable to imagine a full-time engineer spending $100,000 in tokens each year per engineer.

Just really think about that for a second.

I'll have the link in there.

It's fucking stupid.

It's so fucking stupid I want to scream.

And the irony is that, despite being willing to kill these companies by fundamentally changing the terms in which they access their models, Anthropic is also in some way dependent on Cursor, Replit, and other similar firms continuing to buy tokens at the same rate as before, as that consumption is baked into their own ARR figures, as well as the forward-looking revenue projections they've given to investors to convince them to give them $5 billion.

It is, in some sense, a Kobe Ashi Maru.

Anthropic has an existential need to screw over its customers by hiking rates and imposing long-term commitments, but its existence is also in some way predicated on these companies continuing to exist.

If Cursor and Replit both die, that's a significant chunk of Anthropic's API business gone.

And may I remind you, that significantly overshadows its subscription business, making it almost like an inverse of OpenAI, where subscriptions drive the bulk of revenue.

Also a shitty business model.

Anthropic's future is wedded to Cursor.

And I just don't see how Cursor survives, let alone exists or gets subsumed by another company in a way that mirrors how acquisitions have worked since ever.

If Cursor does not sell for a healthy amount, I'm talking 10 to 20 billion dollars and I mean actually sell, not the founders are hired in a strange contractual agreement that pays out investors and its assets are sold to Rick from Pornstars, it's PAWN, by the way, it will prove that no generative AI company to this date has actually been successful.

In reality, I expect a chumly-esque deal that helps CEO Michael Trull buy a Porsche while his staff makes nothing.

Is Cursor worth $10 billion or more?

No, no matter how good its product may or may not be, it is not good enough to be sold at a price that doesn't require cursor to incinerate hundreds of millions of dollars with no end in sight.

And this ultimately gives us a real conundrum.

Why aren't generative AI startups selling?

I'll tell you what though, other people are selling stuff and it's time for you to buy it.

Listen to these ads and let me know what you think at my email address, which is Robert Evans at

There's more to San Francisco with the Chronicle.

More to experience and to explore.

Knowing San Francisco is our passion.

Discover more at sfchronicle.com.

Hi, I'm Morgan Sung, host of Close All Tabs from KQED, where every week we reveal how the online world collides with everyday life.

There was the six-foot cartoon otter who came out from behind a curtain.

It actually really matters that driverless cars are going to mess up in ways that humans wouldn't.

Should I be telling this thing all about my loved life?

I think we will see a Twitch stream or president maybe within our lifetimes.

You can find Close All tabs wherever you listen to podcasts.

Every business has an ambition.

PayPal Open is the platform designed to help you grow into yours with business loans so you can expand and access to hundreds of millions of PayPal customers worldwide.

And your customers can pay all the ways they want with PayPal, Venmo, Pay Later, and all major cards so you can focus on scaling up.

When it's time to get growing, there's one platform for all business: PayPal Open.

Grow today at PayPalopen.com.

Loan subject to approval in available locations.

Mint is still $15 a month for premium wireless.

And if you haven't made the switch yet, here are 15 reasons why you should.

One, it's $15 a month.

Two, seriously, it's $15 a month.

Three, no big contracts.

Four, I use it.

Five, my mom uses it.

Are you playing me off?

That's what's happening, right?

Okay, give it a try at mintmobile.com/slash switch.

Upfront payment of $45 for three months plan, $15 per month equivalent required.

New customer offer first three months only, then full price plan options available.

Taxes and fees extra.

See mintmobile.com.

And we're back.

So, really, though, why are there so few generative AI acquisitions?

I know you're probably going to assume that there have been a ton of them, right?

If you've not been obsessively reading every single story about that for over a year, you might have missed it.

But before we go any further, there have been some, but they've been sparse, and they often have a weird structure that's typically the exception, not the rule in tech.

So, here's a real one: AMD bought Silo AI, the largest private AI lab in Europe, in August 2024 for $665 million, which appears to be the only real acquisition in generative AI history and appears to be partially based on Silo's use of AMD's GPUs.

Elsewhere, Nvidia bought Octo AI for an estimated $250 million in September 2024 after buying Brev.dev in July 2024 for an undisclosed sum and then Gretel in March 2025.

Gretel, Brev.dev, OctoAI, Jesus fucking Christ, these names.

Be normal.

Be normal for one minute.

Yet in all three cases, these products are used to deploy generative AI and not products built on top of generative AI or AI models.

Canva bought generative AI content and research company Leonardo.ai in July 2024 for an undisclosed sum, but that's small, Fry.

Really, the only significant one I've seen was on July 29th, 2025, publicly trained customer service platform NICE buying AI-powered customer service company Cognigi in a $955 million deal.

I hate these names so much.

According to CX Today, CogniGi expects about $85 million in revenue this year, though nobody appears to be talking about how much it costs to make that revenue.

However, according to some sources, they charge tens of thousands or hundreds of thousands of dollars per contract for their AI voice agents that can understand and respond to user input in a natural way.

Great.

Great.

We've got one real deal company built on Models acquisition, and it's a company that most people haven't heard of, making around $7 million a month.

But let's take a look at the others.

So you probably remember Inflection AI to Microsoft, which was not an acquisition, but a $650 million licensing deal that, according to Foss Company, may be more like a billion dollars when you include things like how much it paid to Inflection CEO and former DeepMind co-founder Mr.

Foss Suleiman, who is most famous for just being an abusive and horrible guy to work with.

Here he's pushing a horrifying 996 culture at Microsoft 2.

To do what, Mustafa, you fucking prick.

Anyway, according to Fast Company, the deal involves a license to sell inflections models, a waiver against any employment claims against inflection or Microsoft, paying off investors and some sort of unnamed compensation for employees.

Sounds stinky to me.

But let's look at WinSurf to Google and of course Cognition, which was also not an acquisition.

Winsurf's C-suite went to Google for $2.4 billion, which paid them off along with their investors.

And then the rest of the staff and the product got acquired by Cognition for $250 million.

According to TechCrunch, investors made $1.2 billion on the deal with Winsurf co-founders Varan Mahan and Douglas Chen making another $1.2 billion.

And its staff getting to start a new job at a different company building something else with, according to TechCrunch, a large portion of WinSurf's approximately 250 employees not benefiting from the deal.

As an aside, Mahan and Chen fucking suck.

I fucking hate these guys.

Absolute loser psychopaths.

You make $1.2 billion and you couldn't break off a little bit from that, you fucking

disgusting.

Can't crack that $200 billion.

The 0.2?

You can take $200 million, share it with the class, you fucking losers.

Assholes, pieces of shit.

Come on the show if you want to hear more words like that, you fucking assholes.

Anyway, let's talk about IO products going to OpenAI in an all-stock acquisition.

This deal is a farce and it's unclear if OpenAI actually bought anything.

$6.4 billion in stock?

For what?

Joni Ives' weird face staring at you lovingly as he says things like, I think we should make it look like a circle while making a $5 million a year salary.

Get out of here.

It's not real money.

And then, of course, this character.ai to Google.

This was also not an acquisition.

Google, to quote the Wall Street Journal, paid $2.7 billion to bring back an AI genius who quit in frustration.

I would not use the term genius.

Well, I don't like that term.

Just to be clear, Shazir was one of the authors of the original Attention Is All You Need paper that began the transformer-based large language model era.

Nevertheless, much like Inflection, Google paid a licensing fee to character.ai for its models and hired, according to the information, its co-founders and many of its engineers, creating a fund that would pay out vesting shares, as in the shares you're given when you join a company that you accrue over time as you work there, until July 2026.

Anything after that, you're fucked.

And now, outside of one very industry-specific acquisition, there just doesn't seem to be an investor with the hunger to buy a company like Cursor valued at $9.9 billion or more if they raise another round.

And you have to ask, why?

If AI is, as promised, the thing that will radically change our economy and these companies are building the tools that will bring about that change, why does nobody want to buy them?

And in broader terms, what does it mean when these companies, those with $10 billion, or in the case of OpenAI, $300 or even $500 billion valuations.

What does it mean when they can't be bought and can't go public?

Where does all this go?

What happens next?

What's the plan here?

How will venture firms that plow billions of dollars into these businesses bring a return for their LPs?

That's the limited partners who invest in the venture capitalists, if there are no IPOs and no real buyers?

Sure, investors got paid in these deals, but these deals were like pulling teeth.

These are like, if they do enough of these, there will, even in this administration, be an antitrust thing.

Nevertheless, the economic implications of these questions are quite frankly terrifying, especially when you consider the importance that VC has historically held in building the US tech ecosystem.

And they raise further questions about the impact of an AI bubble on companies that are promising and do have a viable business model and a product with a natural fit, but won't be able to raise any cash because everyone's putting it into fucking AI.

But Ed.

Great.

Ed, what if Cursor turns profitable now?

Great.

Awesome.

I would believe it was possible if it ever, ever happened, which it's not.

I'm not even being sarcastic or rude.

It's just not happened.

No company that actually stakes their entire product on generative AI appears to be able to make a profit of any kind.

Glean, a company that makes at best $8.3 million a month, $100 million in annualized revenue, said that it had $550 million in cash in December of last year, then raised to $150 million in June of this year.

Where'd the money go?

Why does a generative search product with revenues that are less than a third of the Cincinnati Reds' baseball team need a half billion dollars to make

$8.3 million a month, even?

I'm not even saying these companies are unnecessary, so much as they may very well be impossible to run as real businesses.

This isn't even a qualitative judgment of any one generative AI company.

I'm just saying if any of these were good businesses, they would be either profitable or...

be acquired in actual deals and they would also be good businesses by now.

It is not early.

That argument is stupid.

The amount of cash that these businesses are burning does not suggest that they're rapidly approaching any kind of sane burn rate, or we would have heard.

Putting aside any kind of skepticism I have, anything you may hold against me for what I say or the way I say it, I ask you this.

Where are the profitable companies?

Why isn't there one outside of the companies creating training data or NVIDIA?

We're three years in and we haven't had one.

One.

One of them.

One.

One.

We've also had no real exits and no IPOs.

There's been no course for celebration, no validation of a business model through another company, deciding it was necessary to continue its dominance by raising funds on the public market or allowing actual investors, flawed though they may be, to act as a determiner of their value.

I also will say that very few, if not all, of these companies, they would all just turn into a pillar of salt if you made them fill out an S1, which is the document to go public.

It's also unclear what the addition of Windsurf's intellectual property actually adds to Cognition, much like it's a little unclear what differentiates cognition's so-called AI-powered software engineer Devin from anything else on the market.

I heard Goldman Sachs is paying for it and said the stupidest shit I've ever heard to CNBC that nevertheless shows how little it's actually paying for.

This is an actual quote from an actual human being.

We're going to start augmenting our workforce with Devin, which is going to be like our new employee who's going to start doing stuff on behalf of our developers, Ogenti told CNBC.

Initially, we will have hundreds of Devons and that might go into the thousands depending on the use cases.

Hundreds of devons could mean hundreds of seats.

At a very optimistic 500 users at the highest end possible at $500 a month and more than likely it's $20 a month, if not less.

Let's assume that it does discount on an enterprise scale too because that always happens.

That's about $250,000 a month.

Wow.

Wow, $3 million a year in revenue on a trial basis amazing.

To be clear, it's probably far fewer seats and far fewer dollars a month.

I'm nothing if not generous with my adversaries.

In fact, I can't find a shred of evidence that Cognition actually makes money.

Despite currently raising $300 million at a $10 billion valuation, I can find no information about their revenues beyond one comment from the information from July 2024 when Cognition raised at a $2 billion valuation.

Cognition's fundraises the latest example of AI startups raising capital at sky-high valuations despite having little or no revenue.

Cool!

In a further move, per the information that is both a pale horse and a deeply scummy thing to do, Cognition has now laid off 30 people from the Windsurf team and is now offering the remaining 200 buyouts equal to 9 months of salary and, I assume, the end of any chance to accrue further stock in Cognition.

CEO Scott Wu said the following in an email telling Winsurf employees about the layoffs and buyouts.

We don't believe in work-life balance.

Building the future of software engineering is a mission we all care so deeply about that we couldn't possibly separate the two.

We know that not everyone who joined Windsurf has signed up to join Cognition, where we spent six days in the office and clock 80-plus-hour weeks.

Fuck you, man.

Fuck you.

I'm sorry.

You shouldn't be proud of them.

And all that piss, vinegar, and burning of the midnight oil does not appear to have created a product that actually matters.

I realize that's a little cold, but if you're braying and smacking your chest about your hard-charging six days a week office culture, you should be able to do better than we have one publicly known customer and nobody knows our revenue.

Maybe it's a little simpler.

Cognition maybe paid $250 million to acquire Windsurf so that he could, after the transaction, say that they had $82 million in annualized revenue.

If that's the case, this is one of the dodgiest, weirdest acquisitions I've seen in my life.

Two founders getting, what, like, a few hundred million dollars between them and their investors, and a few of their colleagues moving with them to Google, leaving the rest of the staff effectively jobless or in hell with little payoff for their time working at Windsurf.

I can only imagine how it must have felt to go from being supposedly acquired by OpenAI to this farcical rich-get-richer bullshit.

It also suggests that the actual underlying value of Windsurf's IP was around $250 million.

So I ask, why exactly is Cognition worth $10 billion?

And why did it have to raise $300 million after raising hundreds of millions of dollars, according to Bloomberg, in March?

Where is the money going?

It also doesn't seem to have revenue in general.

And Carl Brown of the Internet Bugs revealed that Cognition faked the demo of Devin, the AI-powered software engineer, last year.

And Devin doesn't even rank on SWE Benchmark, the industry standard for model efficacy at coding tasks.

But nevertheless, what I hear is based entirely on one coding language.

Eh, we'll get to that another time.

At best, Cognition has now acquired their own unprofitable coding environment and a smidgen of revenue associated with them.

How would Cognition go public?

What is the actual exit path for Cognition or any other generative AI startup?

Because really, it comes down to three things.

Get acquired, go public, or die.

And that right there is Silicon Valley's own crisis, their own housing crisis.

Except instead of condo houses they can't afford with subprime adjustable rate mortgages, venture capitalists have invested in unprofitable, low-revenue startups with valuations that they can never sell at.

And like homeowners in the dismal years of 2008 and 2009, they're almost certainly underwater.

They just haven't realized it yet.

Where customers were unable to refinance their mortgages to bring their monthly payments down, generative AI startups face pressure to continually raise at higher and higher valuations to keep up with their costs, with each one making it less likely that their company will sell to someone else, by which I mean survive.

The other difference is that, in the case of the housing crisis, those who were able to hold on to their properties eventually saw their equity recover to pre-crash levels, in part because housing is essential and because its price is influenced just as much by supply and demand as it is the ability for people to finance the purchase of properties.

And when the population increases, so too does the demand for housing.

A house is a useful thing that you could live in with walls and stuff.

None of that is true with AI.

There is a finite amount of investors, a finite number of companies, and a finite amount of capital.

And those companies are only as valuable as the expectations their investors have for them and the as the broader sentiment towards ai goes as well also i should add there is only there there don't appear to be that many business models or different ones there's so many different ai coding environments including ones made by these companies amazon has their own one now open ai has codex i mean at some point can all of these things last and who is going to buy cognition because the only other opportunity for the investors who put money into this company to make money here let alone recoup their initial investment, is for them to go public.

Do you think cognition will go public?

How about cursor?

It's worth $9.9 billion.

There was a rumor that it was also raising at $18 billion to $20 billion back in June.

Do you think that Cursor goes public or gets sold?

I mean, maybe, maybe one of these sales.

Do you see perplexity?

Fucking perplexity.

At a valuation of $18 plus billion dollars selling to another company or going public.

The alternative, as discussed, is that Perplexity, a company with 15 million users and around 550 million annualized revenue, is still making less than half of the revenue, by the way, of the Cincinnati Reds baseball team.

They're going to go public.

They're going to go public.

Just to be clear, the Reds make $325 million

a year.

They're profitable, real money.

How is Perplexity, which is an inferior business to the Reds, how are they meant to do that?

They've, at this point, raised over a billion dollars and they lost $68 million in 2024 on $34 million of revenue.

By comparison, by the way, the Reds are a great business with a net income of $29 million, all to provide a service that upsets and humiliates millions of people from Ohio every year for the pleasure of America.

That's a great business.

We should respect the Reds.

The Reds should be interviewed by Bloomberg.

We should have Ellie De La Cruz talking every time we hear from Sam Altman.

Just put Ellie in there.

He'd be way more interesting and also works harder and provides greater value as a business leader.

Anyway, putting aside the reds, what exactly is it that Perplexity could offer to the public markets as a stock or to an acquirer?

Apple considered acquiring them back in June, but Apple tends to acquire companies it wants to integrate into their core business, as was the case with Siri, which makes me think that Perplexity leaked that information about a deal that was never really serious.

Hell, Meta also talked about acquiring them too.

And again, I would assume that Perplexity leaked that.

Isn't it weird, by the way, that two different companies talked about buying Perplexity, but neither of them did so?

ceo aravin srivenas said in july that he wanted to remain independent which is a weird thing to say after talking to two giant multi-trillion dollar market cap tech firms about selling them to them like you you seem to not want to stay independent why would you have the conversation if you're were you just getting him to hell bullshit it's almost as if nobody wants to buy perplexity or that Really, they want to buy any of these sham companies, which I know sounds mean, but if you're worth tens of billions or a billion dollars or anything over like 500 million dollars and you can't make more than a bottom-tier baseball team in fucking Ohio you're neither innovative nor deserving of said valuation but really my pissiness and baseball comparisons aside what exactly is the plan for these goddamn companies they don't make enough money to survive without a continuous flow of venture capital and they don't seem to make impressive sums of money even when allowed to burn as much as they'd like these companies are not being forced to live frugally or at least have yet to be made to perhaps because they have all actively engaged as at spending as much money as possible in pursuit of finding an idea that makes more money than it loses.

This is not a rational nor reasonable way to proceed.

Yes, there are startups that can justify burning capital.

Yes, there are companies that have burned hundreds of millions of dollars to find their business models, or billions in the case of Uber, but none of these companies are like those companies in the generative AI space.

Generative AI businesses don't have the same economics, nor do they have the same total addressable markets.

Now, if you're going to say Amazon Web Services, there's a three-part episode, the Hater's Guide to the AI bubble that explains that in detail.

I do voices, I have sources.

You need to go and listen to that.

I'm fucking tired of hearing people about this one.

Oh, what about Uber?

Oh, what about AWS?

You ain't got shit.

That's all you've got.

You certainly don't have a company to point to.

These startups are their VC firms' subprime mortgages, overstuffed valuations with no exit route, and no clear example of how to sell them or who to sell them to.

The closest they've got is using generative AI startups as beauty pageants for guys wearing Patagonia, finding ways ways to pretend that the guy who runs an AI startup, sorry, AI lab, is some sort of mysterious genius versus just another founder and just another bubble with just another overstuffed valuation.

The literal only liquidity mechanism outside of Cogniti that generative AI has had so far is selling AI talent to big tech at a premium.

Nobody has gone or is going public, and if they're not going public, the only route for these companies is to either become profitable, which they haven't, or sell to somebody, which they're not.

But I've been dancing around the real reason they won't sell, because fundamentally, generative AI does not let companies build something new.

Anyone that builds a generative AI product is ultimately just prompting the model, albeit in increasingly more complex ways at the scale of something like Claude Code, or though Anthropic has the advantage of being one of the main veins of their infrastructure, or of course Cursor.

This means that a generative AI company owns very few unique things beyond their talent and will forever be at the mercy of any and all decisions that their model provider makes, such as increasing prices or creating competing products.

I know it sounds ludicrous, but this is the reality of these companies.

While there are some companies that have some unique training and models and such, none of them seem to be building interesting or unique products as a result.

If your argument is that these things take some time, how long?

No, really.

So many of you have said this is what happens.

They burn a lot of money, they grow, and then

what?

You stop short, because the next thing you say is turn profitable by getting enterprise customers.

Nobody can do the first part and if you can do the second part in anything approaching a consistent fashion.

But really, how long should we give them?

Three years?

Perplexity's had three years and a billion dollars.

It doesn't seem to be close to profitable.

How long does Perplexity deserve exactly?

Four years?

Five years?

An eternity?

Every single example of a company that's burned a lot of money and then not done so in the end has been a company with a physical thing or connections to the real world, with the exception of Facebook, which has never had the kind of cash-burning monstrosity that generative AI is, or Meta has now created for no reason.

There has never been a software company that has just chewed through hundreds of millions or billions of dollars and then suddenly become profitable magically, mostly because the magical valuations of software have been in their ability to transcend infrastructure.

Once unit economics in the sales of software like Microsoft Office or providing access to Instagram do not require the most powerful graphics processing units run full tilt at all times.

And those are products that people like and want to use every day, even though both of them seem to be getting worse every minute.

I get people saying that these companies are in the growth stage.

But when all of them are unprofitable, and even the unprofitable ones outside of OpenAI and Anthropic aren't really making much money, come on, this isn't anything like any boom that leads to something good or useful or profitable, and it's because the economics do not make sense.

And that's before we get into OpenAI and Anthropic in the next episode.

Thank you for listening to Better Offline.

The editor and composer of the Better Offline theme song is Mattasowski.

You can check out more of his music and audio projects at mattasowski.com.

M-A-T-T-O-S-O-W-S-K-I dot com.

You can email me at easy at betteroffline.com or visit betteroffline.com to find more podcast links and of course my newsletter.

I also really recommend you go to chat.where's your ed.at to visit the Discord and go to r slash betteroffline to check out our Reddit.

Thank you so much for listening.

Better Offline is a production of CoolZone Media.

For more from CoolZone Media, visit our website, coolzonemedia.com, or check us out on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

Hi, I'm Morgan Sung, host of Close All Tabs from KQED, where every week we reveal how the online world collides with everyday life.

There was the six-foot cartoon otter who came out from behind a curtain.

It actually really matters that driverless cars are going to mess up in ways that humans wouldn't.

Should I be telling this thing all about my love life?

I think we will see a Twitch stream or president, maybe within our lifetimes.

You can find Close All Tabs wherever you listen to podcasts.

Every business has an ambition.

PayPal Open is the platform designed to help you grow into yours with business loans so you can expand and access to hundreds of millions of PayPal customers worldwide.

And your customers can pay all the ways they want with PayPal, Venmo, Pay Later, and all major cards so you can focus on scaling up.

When it's time to get growing, there's one platform for all business: PayPal Open.

Grow today at paypalopen.com.

Loans subject to approval in available locations.

Want a bourbon with the story?

American-made wild turkey never compromises.

Like a good story and never watered down.

Our pre-prohibition-style bourbons are distilled and barreled at a low proof to retain the most flavor.

Most bourbons aren't.

Well, we aren't most bourbons.

Wild Turkey 101 bourbon makes an old-fashioned or bold-fashioned for bold nights out or at home.

Now that's a story worth telling.

Wild Turkey, trust your spirit.

Copyright 2025, Capari America, New York, New York.

Never compromise, drink responsibly.

Hear Something Good on Women's Health and Longevity, a new podcast on iHeart.

Join us for conversations with renowned medical experts.

They'll share the latest breakthroughs, the good news about women's health, and the simple steps women can take to help them live healthier and happier every day.

Be sure to listen to our episode, Period Power, where we explore how menstrual health is foundational to lifelong well-being and how accurate information can shape health outcomes across a lifetime.

Brought to you by Tampax, found at Walgreens, the women's well-being destination, supporting every stage.

Listen to hear something good on women's health and longevity on the iHeartRadio app, Apple Podcasts, or wherever you get your favorite shows.

This is an iHeart podcast.