The Hater's Guide To The AI Bubble, Pt. 1

Speaker 1 This is an iHeart podcast.

Speaker 2 Be honest, how many tabs do you have open right now?

Speaker 3 Too many?

Speaker 2 Sounds like you need Close All Tabs from KQED, where I, Morgan Sung, Doom Scroll so you don't have to.

Speaker 1 Every week, we scour the internet to bring you deep dives that explain how the digital world connects and divides us all.

Speaker 5 Everyone's cooped up in their house. I will talk to this robot.

Speaker 1 If you're a truly engaged activist, the government already has data on you.

Speaker 9 Driverless cars are going to mess up in ways that humans wouldn't.

Speaker 6 Listen to Close All Tods, wherever you get your podcasts.

Speaker 10 There's more to San Francisco with the Chronicle. More to experience and to explore.

Speaker 10 Knowing San Francisco is our passion.

Speaker 10 Discover more at sfchronicle.com.

Speaker 11 Every business has an ambition.

Speaker 3 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.

Speaker 17 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.

Speaker 19 When it's time to get growing, there's one platform for all business, PayPal Open.

Speaker 17 Grow today at PayPalOpen.com.

Speaker 20 Loans subject to approval in available locations.

Speaker 21 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?

Speaker 21 That's exactly what Cohere, Thomson Reuters, and Specialized Bikes have since they upgraded to the next generation of the cloud. Oracle Cloud Infrastructure.

Speaker 21 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.

Speaker 21 How is it faster? OCI's block storage gives you more operations per second.

Speaker 8 Cheaper?

Speaker 21 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.

Speaker 21 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/slash strategic.

Speaker 17 Call Zone Media.

Speaker 22 hello and welcome to Better Offline. I'm your host, Ed Zitron.

Speaker 22 Check out the episode notes. Got wonderful merchandise and completely separate to the podcast.

Speaker 22 Got a wonderful newsletter, where's your red dot at with the premium section that I would love you to subscribe to? But I also got some good news.

Speaker 22 You've got another three-part episode, the second of the year.

Speaker 22 And this week, we're going to be talking about how the cracks in the generative AI market are becoming harder to ignore and how recent events are making a collapse seem all the more inevitable.

Speaker 22 What that means for the wider economy, really the markets, and you.

Speaker 22 And I want to make that case because, as a journalist, I believe I have the duty to give you the information you need to make sense of a world increasingly feeling incomprehensible.

Speaker 22 And of course, detached from reality, and where everything is consequential to everyone, where it's impossible for ordinary people to shroud themselves from the consequences of decisions made by the executive and shareholder class.

Speaker 22 Good journalism is making sure that history is actively captured and appropriately described and assessed, and it's accurate to describe things as they currently are as alarming.

Speaker 22 And boy howdy, am I alarmed. Now, alarm is not a state of weakness or belligerence of myopia.

Speaker 22 My concern does not dull my vision, even though it's a convenient to frame me as somehow alarmist, like I have some hidden agenda or bias toward doom.

Speaker 22 I profoundly dislike the financial waste, the environmental destruction, and, fundamentally, I dislike the attempt to gaslight people into swearing fealty to a sickly and frail pseudo industry where everybody but NVIDIA and consultancies lose money.

Speaker 22 And I also dislike the fact that I and others like me are held to a remarkably different standard to those that paint themselves as optimists, which typically means people that agree with what the market wishes were true.

Speaker 22 Critics are continually badgered, prodded, poked, mocked, and jeered at for not automatically aligning with the idea that generative AI will be this massive industry, constantly having to prove themselves as if somehow there's something malevolent or craven about criticism.

Speaker 22 The critics do this for clicks or to be contrarian.

Speaker 22 I don't do anything for clicks or downloads or prestige. I don't have any stocks or short positions.
My agenda is simple. I like talking about this crap and it comes to me naturally.

Speaker 22 I have a podcast and it is on some level my job to try and understand what the tech industry is doing day to day.

Speaker 22 And I get it.

Speaker 22 I get it's easy to try and dismiss what I say as going against the grain because AI is big and AI is something we should all be impressed by and the AI is the thing that's going to start everything and all this fucking money is tied up in it.

Speaker 22 But look, this isn't a fad for me. This isn't something I'm doing because I feel like it or because I'm jumping to the next trend.
No, I've been railing against bullshit bubbles since 2021.

Speaker 22 The anti-remote work push and the people behind it, the clubhouse and audio social networks bubble, the NFT bubble, the made-up quitting panic, and even I even, and this one I got no credit for.

Speaker 22 I called that something was up with FTX several months before it imploded. Did I do much more than that? No, I found one thing.
Nevertheless, this isn't contrarianism.

Speaker 22 Not at all. It's the kind of scepticism of power and capital that's necessary to meet these moments.

Speaker 22 And it's if it's necessary to dismiss my work because it makes you feel icky inside, get a therapist or see a priest.

Speaker 22 Nevertheless, I'm alarmed, and while I have said some of these things separately, based on recent developments, I think it's necessary to say why.

Speaker 22 In short, I believe the AI bubble is deeply unstable, built built on vibes and blind faith. And when I say the AI bubble, I mean the entirety of the AI trade.
And it's alarmingly simple, too.

Speaker 22 He says before doing three episodes on it. But this isn't going to be some Sacher and whiny or simply Warrison podcast.

Speaker 22 I think at this point, it's become a little ridiculous to not see we're in a bubble. We are in a goddamn bubble, by the way.
It's so obvious we're in a bubble. It's been so obvious we're in a bubble.

Speaker 22 It's been obvious for months, if not years, a bubble that seems so strong, but it's actually very weak, with a central point of failure. I may not be a contrarian, contrarian, but I am a hater.

Speaker 22 I hate the waste, the loss, the destruction, the theft, the damage to our planet, and the sheer excitement that some executives and yes, some writers have that workers may be replaced by AI.

Speaker 22 And the bold-faced fucking lie that it's actually happening, and what generative AI is doing is somehow proof that it will.

Speaker 22 And so I present to you The Hater's Guide to the AI Bubble, a comprehensive rundown of the arguments I have against the current AI boom's existence.

Speaker 22 Send this podcast to your friends, your loved ones, or I don't know, blare it in their ears like you're torturing them.

Speaker 22 But no, this isn't going to be a traditional guide, but something you can listen to and say, oh, that's why the AI bubble is so bad.

Speaker 22 And at this point, I know I'm tired of being gaslit by guys in Gingham shirts who desperately want to curry favor with other guys in Gingham shirts, but who also have PhDs.

Speaker 22 I'm tired of hearing people talk about how we're in the era of agents that don't fucking work and will never fucking work.

Speaker 22 I'm tired of hearing about powerful AI that's actually crap, and I'm tired of being told the future is here while having the world's least useful, most expensive cloud software shoved down my throat and up my asshole.

Speaker 22 Look, the generative AI boom is a mirage. It hasn't got the revenue or the returns or the product efficacy for it to matter.
Everything you're seeing is ridiculous and wasteful.

Speaker 22 And when it August tits up, I want you to remember that I said this. I tried to say something.
I've been trying to say something for a while. But let's start with something real obvious.

Speaker 22 Let's start by talking about the so-called Magnificent Sevens weak point. And it's not the one you'd think, because it's NVIDIA.

Speaker 22 As I write the script for this podcast, NVIDIA is sitting around $170 a share, a dramatic reversal of faith after the pummeling it took from the DeepSeek situation in January, which sent it tumbling to a brief late April trip below $100 before things turned around.

Speaker 22 The Mag 7 stocks, NVIDIA, Microsoft, Alphabet, which is Google, Apple, Meta, Tesla, and Amazon, make up around 35% of the value of the US stock market.

Speaker 22 And of that, Nvidia's market value takes up about 19% of the Magnificent 7. They're about like 8% to 9% of the entire US stock market.

Speaker 21 It's not brilliant.

Speaker 22 This dominance is also why ordinary people ought to be deeply concerned about the AI bubble.

Speaker 22 The Magnificent 7 is almost certainly a big part of their retirement plans, even if they're not directly invested.

Speaker 22 Back in May, the wonderful Laura Bratton from Yahoo Finance reported that Microsoft, Amazon, Meta, Alphabet, and Tesla alone make up 42.4% of NVIDIA's revenue.

Speaker 22 The breakdown doesn't make things better. Meta spends 25% and Microsoft an alarming 47% of their capital expenditures on NVIDIA chips.

Speaker 22 And as Bratton notes, Microsoft also spends money renting servers from Core Weave, which analyst Gil Luria of D.A.

Speaker 22 Davidson estimates accounted for $8 billion, more than 6% of NVIDIA's revenue in 2024.

Speaker 22 Luria also estimates that neo-cloud companies like Coreweave and Crusoe that exist only to provide AI compute services account for as much as 10% of NVIDIA's revenue, or at least did so in 2024.

Speaker 22 NVIDIA's climbing stock value comes from one thing, its continued revenue growth.

Speaker 22 In the past four quarters, NVIDIA has seen over year a growth of 101 94 78 and 69 and in the last quarter a little statistic was carefully brushed under the rug nvidia missed though narrowly on data center revenue and data center revenue is by the way where the gpus go and all the associated hardware and so and kind of server architecture switches and the like and yeah this is exactly what it sounds like gpus that are used in servers rather than gaming consoles and pcs i get a lot of emails saying oh will it be easier for me to buy consumer graphics cards?

Speaker 22 I don't fucking know, mate. I'm just here to talk about enterprise bullshit.
Okay, not really enterprise, but enterprise scale GPUs. We're getting off track.
Analysts estimated

Speaker 22 it would make $39.4 billion from the data center category. And NVIDIA only, only, I know, pathetic amount, brought in $39.1 billion.

Speaker 22 And again, this could be attributed to their problems in China, especially as the H20 ban, they were banned from selling a specific chip in China, has only just been lifted.

Speaker 22 In any case, this was amiss. And I'm not sure why no one wanted to talk about it.

Speaker 22 But there's another problem. There's so many little problems here.
Like, NVIDIA's quarter over quarter growth has also become aggressively normal. It went from 69% to 59% to 12%

Speaker 22 to 12% again,

Speaker 22 quarter over quarter, which isn't bad. It's pretty great, in fact.
But when 88% of your revenue is based on one particular line in your earnings, it's a pretty big concern, at least for me.

Speaker 22 Look, I'm no stock analyst. Do not take stock advice from me.
I don't know about stocks. I don't know what will go up and down, but I'll tell you,

Speaker 22 I'll tell you something's not right here. So, but I'm going to keep this simple.

Speaker 22 NVIDIA relies on not only selling lots of GPUs each quarter, but it must always sell more of them the following quarter.

Speaker 22 More than 42% of NVIDIA's revenue comes from Microsoft, Amazon, Meta, Alphabet, and Tesla continuing to buy more GPUs.

Speaker 22 Remember, it's not about buying the same amount, number must go up nvidia's continued value and continued growth is heavily reliant on hyperscaler purchases and continued interest in generative ai but really just the buying part the gpus are what matter and the u.s stock market's continued health relies on some level of five or six companies and it's unclear how many gpus apple buys spending billions of dollars on gpus from nvidia and more every quarter in fact i found an an analysis from portfolio manager Danke Wang from January 2025 that found that the Magnificent 7 stocks accounted for 47.87% of the Russell 1000 index's returns in 2024.

Speaker 22 And that's an index fund of the thousand highest-ranked stocks on the FTSE Russell Index, which in simpler terms means 35% of the US stock market is held up by five or six companies buying GPUs.

Speaker 22 If Nvidia's growth story stumbles, it will reverberate through the rest of the Mag7 too, making them rely on their own AI trade stories. And would you know it? Wouldn't you flipping know it?

Speaker 22 When you look at the stories, there is no AI trade because generative AI is not making anybody any goddamn money.

Speaker 22 But I have to make money, which is why you need to listen to the following advertisement. It's the only one of these bloody things you're going to get.
Here's an ad.

Speaker 10 There's more to San Francisco with the Chronicle. More to experience and to explore.

Speaker 10 Knowing San Francisco is our passion.

Speaker 10 Discover more at sfchronicle.com.

Speaker 2 Be honest, how many tabs do you have open right now?

Speaker 3 Too many?

Speaker 2 Sounds like you need close all tabs from KQED, where I, Morgan Sung, doom scroll so you don't have to.

Speaker 1 Every week, we scour the internet to bring you deep dives that explain how the digital world connects and divides us all.

Speaker 5 Everyone's cooped up in their house. I will talk to this robot.

Speaker 1 If you're a truly engaged activist, the government already has data on you.

Speaker 9 Driverless cars are going to mess up in ways that humans wouldn't.

Speaker 6 Listen to Close All Tabs, wherever you get your podcasts.

Speaker 24 So, I've shopped with Quince before they were an advertiser and after they became one.

Speaker 22 And then again, before I had to record this ad, I really like them.

Speaker 24 My green overshirt in particular looks great.

Speaker 25 I use it like a jacket.

Speaker 24 It's breathable and comfortable and hangs on my body nicely. I get a lot of compliments.

Speaker 24 And I liked it so much, I got it in all the different colours, along with one of their corduroy ones, which I think I pull off. And really, that's the only person that matters.

Speaker 24 I also really love their linen shirts, too. They're comfortable, they're breathable, and they look nice.
Get a lot of compliments there, too.

Speaker 22 I have a few of them.

Speaker 24 Love their rust-coloured ones as well. And in general, I really like Quince.

Speaker 22 The shirts fit nicely, and the rest of their clothes do too.

Speaker 24 They ship quickly, they look good, they're high quality, and they partner directly with ethical factories and skip the middleman.

Speaker 22 So you get top-tier fabrics and craftsmanship at half the price of similar brands.

Speaker 24 And I'm probably going to buy more from them very, very soon. Keep it classic and cool this fall.

Speaker 24 With long-lasting staples from Quince, go to quince.com slash better for free shipping on your order and 365-day returns.

Speaker 22 That's q-u-in-ce-e.com slash better.

Speaker 24 Free shipping and 365-day returns. Quince.com slash better.

Speaker 11 Every business has an ambition.

Speaker 3 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.

Speaker 17 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.

Speaker 19 When it's time to get growing, there's one platform for all business: PayPal Open.

Speaker 17 Grow today at paypalopen.com.

Speaker 20 Loan subject to approval in available locations.

Speaker 22 And we're back.

Speaker 22 And I am so tired of people telling me that companies are making tons of money on AI. They are not.
Anyone saying this to you is lying or ignorant or both.

Speaker 22 The Magnificent 7 spent an insane $560 billion between 2024 and 2025 on CapEx, with the overwhelming majority going towards generative AI and their effort for this wonderful effort.

Speaker 22 For more than half a trillion dollars, these companies have made about $35 billion in revenue and no profit. And I must say, and this is a technical term, this is egregiously fucking stupid.

Speaker 22 But let's break it down, starting with starting out in Redmond with our friends at Microsoft. And they plan to spend $80 billion on CapEx in 2025.

Speaker 22 Now, as of January 2025, Microsoft's annualized revenue, meaning best month times 12, from artificial intelligence was $13 billion, a number that it's chosen not to update since, likely because said number is either flat or not growing.

Speaker 22 Though it could in its upcoming, I think at the end of this week or next one, it's got earnings coming up. Maybe there'll be good news.

Speaker 22 Yet the problem with this revenue is that $10 billion of that revenue, according to the information, comes from OpenAI spend on Microsoft's Azure Cloud.

Speaker 22 And Microsoft offers preferential pricing, I'm quoting the information here, at a heavily discounted rental rate that essentially only covers Microsoft's costs for operating the servers.

Speaker 22 That's not good, right?

Speaker 22 Like, it's not good.

Speaker 22 It's not good that 76.9% of Microsoft's AI revenue comes from OpenAI, and that revenue is made at cost or just above it, which makes Microsoft's real AI revenue about $3 billion, or about 3.75% of this year's capital expenditures, or 16.25% if you count OpenAI's revenue, which costs Microsoft likely more money than it earns.

Speaker 22 The information also reports that Microsoft made $4.7 billion in AI revenue in 2024, of which OpenAI accounted for $2 billion, meaning that for the $135.7 billion that Microsoft has spent in two years in AI infrastructure, it's made $17.7 billion, of which OpenAI was $12.7 billion.

Speaker 22 It's kind of crap, isn't it? It's not very good at all. And things do not improve when we get to Amazon.

Speaker 22 An analyst estimates that Amazon, which plans to spend $105 billion in capital expenditures this year, will make $5 billion in AI in 2025, rising, and I quote, as much as 80%, suggesting that Amazon might have made a measly $2.77 billion in 2024 on AI in a year when it spent $83 billion in capital expenditures.

Speaker 22 And last year, Amazon CEO Andy Jassy said that, and I quote, AI represents for sure the biggest opportunity since cloud and probably the biggest technology shift and opportunity in business since the internet i personally think he is full of shit and it's a similar story over with google which plans to spend 75 billion dollars in capex 2025

Speaker 22 bank of america analyst justin post estimated a few weeks ago that google's ai revenue would be in the region of 7.7 billion dollars though his math if i'm honest is a little generous because it includes subscribers to packages that include a lot of non-ai stuff too Google's one subscription includes increased cloud storage across Google Drive, Gmail, and Google Photos, and added a $20 a month premium plan in February 2024 that included access to Google's various AI models.

Speaker 22 Google has claimed that the premium AI tier accounts for millions of the 150 million subscribers to Google One, though how many millions is impossible to estimate. That won't stop me trying, though.

Speaker 22 Assuming the $3.1 billion in 2025 revenue would work out to $258 million a month, that would mean there were 12.9 million Google One subscribers also paying for the premium AI tier.

Speaker 22 This isn't out of the realm of possibility. After all, OpenAI has like 15.5 million paying subscribers, but Post is making a kind of a generous assumption here.

Speaker 22 Nevertheless, we'll accept the numbers as they are because they fucking stink.

Speaker 22 Google's $1.1 billion in workspace revenue came from a forced price hike on those who use Google's services to run their businesses, my ass included, meaning that it's not likely a number that they can significantly increase in the future because it was just raising the rent on everyone.

Speaker 22 And that's $7.7 billion of revenue, not profit on $75 billion of capital expenditures. Very nasty.
But let's move on to one of my faves, Meta, which plans to spend $72 billion in 2025.

Speaker 22 Someone's going to get mad at me for saying this, but I believe that Meta is simply burning cash on generative AI.

Speaker 22 There is no product that Meta sells that monetizes large language models, that I can tell at least, but every Meta product now has them kind of shoved in there.

Speaker 22 Your Instagram DMs oinking at you to generate artwork based on your conversation.

Speaker 22 Nevertheless, they do make some money allegedly, and we do have some sort of knowledge of what Meta is saying they make due to a copyright infringement case, Cadre versus Meta.

Speaker 22 Unsealed judgment briefs revealed in April that Meta is claiming that Gen AI-driven revenue will be more than $2 billion in this year, with estimates as high as $3 billion.

Speaker 22 The same document also claims that Meta expects to make $460 billion to $1.4 trillion in total revenue through 2035. And this is from AI, by the way.

Speaker 22 And this is the kind of thing that should have you like wrenched out of the, your key card should stop working when the words leave your mouth.

Speaker 22 Because Meta makes 99% of its revenue from advertising and the unsealed documents state that it generates from its Llama models and will continue earning revenue from each iteration and share a percentage of the revenue it generates from users of the Llama models hosted by those companies with the companies in question redacted.

Speaker 22 Mr. Max Zeff of TechRunch adds that Meta lists host partners like Amazon Web Services, Nvidia, Databricks, Grok, Dell, Microsoft Azure, Google Cloud, and Snowflake.

Speaker 22 So it's possible that Meta makes money licensing to those companies. Sadly, the exhibits further discussing these numbers are filed on the seal.
And also, their large language model is open source.

Speaker 22 What service is Meta providing? Are these companies so goddamn lazy that they need Meta to come in and set up the fruit?

Speaker 22 Jesus Christ. Jesus Christ.
When I read these numbers, I just, when I read about these people, they drive me a little insane.

Speaker 22 Either way, we are now at $332 billion of capital expenditures in 2025 for $28.7 billion of revenue, of which $10 billion of it is OpenAI's at cost or just above cost revenue. Not great.

Speaker 22 Then there's Tesla, which doesn't appear to make money from generative AI and plans to spend $11 billion on CapEx in 2025.

Speaker 22 Despite its media prominence in the Magnificent 7 at least, Tesla is one of the least exposed companies of the Mag 7 to the AI trade, as Elon Musk has turned it into a meme stock company where what they do doesn't really matter.

Speaker 22 That doesn't mean, of course, that Musk isn't touching AI.

Speaker 22 The XAI, the company that develops racist large language model Grok and owns what remains of Twitter, apparently burns a billion dollars a month.

Speaker 22 And the information reports that it makes a whopping $100 million of annualized revenue, so about $8.33 million a month.

Speaker 22 Now, there's a shareholder vote for Tesla to potentially invest in XAI, which will probably happen, allowing Musk to continue to pull leverage from his Tesla stock until the company's decaying sales and brand eventually swallow him whole.

Speaker 22 But we're not talking about Elon Musk today.

Speaker 22 We are not.

Speaker 22 We have to talk about Apple now. And honestly, they're the least interesting part of this story.

Speaker 22 Their capital expenditures in 2025 are expected to also be around $11 billion and they arguably have the weirdest

Speaker 22 AI story in the Magnificent 7.

Speaker 22 Apple intelligence radicalized millions of people against AI, mostly because it fucking sucks.

Speaker 22 Apple clearly got into AI reluctantly and now faces stories about how they feel left behind in the AI race, which mostly means that Apple aggressively introduced people to the actual features of generative AI by force.

Speaker 22 And it turns out that people don't really want to summarize documents or write emails or make custom emoji. And anyone who thinks they would is a fucking alien.

Speaker 22 In any case, Apple hasn't bet the farm on AI inso much as it hasn't spent $200 billion in infrastructure for a product with a limited market that only loses money.

Speaker 22 And again, if you want to give me some money, I'm going to put an ad break here. So after this, whatever comes next, buy it.
Or don't if you don't want to, but really you should.

Speaker 22 If I'm speaking, if you hear my voice in the ad,

Speaker 22 then you should buy it.

Speaker 26 Unless you don't want to.

Speaker 1 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.

Speaker 27 There was the six-foot cartoon otter who came out from behind a curtain.

Speaker 8 It actually really matters that driverless cars are going to mess up in ways that humans wouldn't.

Speaker 23 Should I be telling this thing all about my loved life?

Speaker 7 I think we will see a Twitch stream or a president maybe within our lifetimes.

Speaker 1 You can find Close All tabs wherever you listen to podcasts.

Speaker 11 Every business has an ambition.

Speaker 3 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.

Speaker 16 And your customers can pay all the ways they want with PayPal, Venmo, Pay Later, and all major cards.

Speaker 18 So you can focus on scaling up.

Speaker 19 When it's time to get growing, there's one platform for all business, PayPal Open.

Speaker 17 Grow today at PayPalOpen.com.

Speaker 20 Loan subject to approval in available locations.

Speaker 25 As a founder, you're moving fast toward product market fit, your next round, or your first big enterprise deal.

Speaker 25 But with AI accelerating how quickly startups build and ship, security expectations are higher earlier earlier than ever.

Speaker 25 Getting security and compliance right can unlock growth or stall it if you wait too long.

Speaker 25 With deep integrations and automated workflows built for fast-moving teams, Vanta gets you audit-ready fast and keeps you secure with continuous monitoring as your models, infra, and customers evolve.

Speaker 25 Fast-growing startups like Langchain, Writer, and Cursor trust Advanta to build a scalable foundation from the start.

Speaker 25 Go to Vanta.com slash iHeart to save $1,000 today through the Vanta for Startups program and join over 10,000 ambitious companies already scaling with Vanta.

Speaker 25 That's vanta.com slash iHeart to save 1,000 for a limited time.

Speaker 28 Hey, it's Ryan Reynolds here for Mint Mobile. Now, I was looking for fun ways to tell you that Mint's offer of unlimited premium wireless for $15 a month is back.

Speaker 8 So I thought it would be fun if we made $15 bills. But it turns out that's very illegal.

Speaker 26 So there goes my big idea for the commercial.

Speaker 28 Give it a try at mintmobile.com slash switch.

Speaker 30 Up from payment of $45 $45 for a three-month plan equivalent to $15 per month required. New customer offer for first three months only.
Speed slow after 35 gigabytes if networks busy.

Speaker 31 Taxes and fees extra.

Speaker 30 See mintmobile.com.

Speaker 22 And we're back.

Speaker 22 Now I'm going to use a new term I came up with that's really, really bad, but the fragile five, I call them Amazon, Google, Microsoft, Metra, and Tesla, the ones investing all the money in the GPUs, are holding up the US stock market by funding NVIDIA's future growth story.

Speaker 22 And this is really the first big takeaway I want you to take from this three-parter.

Speaker 22 To be clear, I'm not saying that any of the Mag 7 are going to die, just that five companies spend on NVIDIA GPUs largely dictate how stable the US stock market will be.

Speaker 22 If any of these companies, but especially NVIDIA sneeze, your 401k and your kids' college fund will probably catch a cold.

Speaker 22 I realize this sounds a little simplistic, but by my calculations, NVIDIA's value underpins about 8% of the value of the US stock market.

Speaker 22 At the time of writing, it accounts for roughly 7.5% of the SP 500, an index of the 500 largest US publicly traded companies.

Speaker 22 A disturbing 88%, as I mentioned, of NVIDIA's revenue comes from enterprise-scale GPUs primarily used for generative AI, of which five companies spend makes up over 42% of its revenue.

Speaker 22 In the event that any one of these companies makes significant changes to their investments in NVIDIA chips, it will likely have a direct and meaningful negative impact on the wider economy and markets.

Speaker 22 NVIDIA's earnings are effectively the US stock market's confidence and everything rides on five companies. And if we're honest here, really four companies as Tesla is 0.9%

Speaker 22 of the investment in GPUs of those five companies buying GPUs for generative AI or to train generative AI models worse still these services while losing these companies massive amounts of money don't really produce much revenue meaning that the AI trade is not driven by any real meaningful revenue growth but Ed Ed Ed

Speaker 22 they said they said points of growth silence quiet

Speaker 22 nothing more out of you any of these companies talking about growth from AI or the jobs that AI will replace or how AI has changed their organization are hand-waving to avoid telling you how much money these services are actually making them.

Speaker 22 If they were making good money and experiencing real growth as a result of AI, they wouldn't shut the fuck up about it.

Speaker 22 They'd be in your ear and up your ass hooting and hollering about how much cash they were rolling in.

Speaker 22 And they're not, because they're not rolling in cash and are, in fact, blowing nearly $100 billion each to build massive, power-hungry, costly data centers for no real reason.

Speaker 22 Don't watch the mouth, watch the hands. These companies are going to say they're seeing growth from AI, but unless they actually show you the growth and enumerate it, they are kind of lying.

Speaker 22 They're lying in the way that you're allowed to. But hey, hey, Amazon Web Services took years to become profitable.
People said Amazon would fail.

Speaker 22 So this is one of the most annoying and consistent responses to my work.

Speaker 22 And it's when people say that either Amazon or Amazon Web Services ran at a loss and that Amazon Web Services, which pretty much was the invention of modern mass market cloud compute infrastructure for running stuff on the cloud, lost money and then didn't.

Speaker 22 Here's the thing. This statement is one of the things that people say because it sounds rational.
Amazon did lose money and Amazon web services was expensive. That's right, right? It's obvious, right?

Speaker 22 The thing is, I've never really had anyone explain this point to me. So I finally sat down.

Speaker 22 I'm going to deal with this criticism because every fucking person who mentions it thinks they just pulled Excalibur from the stone and can now decapitate me.

Speaker 22 They claim that because people in the past doubted Amazon because, or in addition to the burn rate of the AWS systems as the company built out its infrastructure, that I too am wrong because the analysts were wrong about that.

Speaker 22 This isn't Camelot, you're a Rube, you are not King Arthur. And now I will address both the argument itself and the they part of it too.

Speaker 22 Because if the argument is that the people who got Amazon web services wrong should not be trusted, then we should no longer trust them.

Speaker 22 The people who actively propagandize something wrong, we shouldn't trust them, right?

Speaker 22 Right?

Speaker 22 Well, you'll never guess who's now saying AI is good. Oh, I'm I'm going to get there.
Don't you flippin worry. But if I'm honest, I'm not sure where this argument came from.

Speaker 22 Because there is, to my knowledge, no story about Amazon web services where somebody suggests its burn rate would kill Amazon. But I'm a curious little critter.
So let's start with an obvious one.

Speaker 22 The obvious point. I want to give a shout out to Harry McCracken of Fast Company for bringing this one up to me.
In May 31st, 1999, there was a piece that everybody is thinking of called Amazon.bomb.

Speaker 22 And the writer, Jacqueline Doherty, was mocked soundly for being wrong about Amazon, which has now become quite profitable.

Speaker 22 The article, along with the other sources that form the basis of this episode, are going to be linked in the spreadsheet. And as a surprise, I'll actually update you.

Speaker 22 I also want to be clear that Amazon Web Services did not launch until 2006. And Amazon itself would become reliably profitable in 2003.

Speaker 22 Technically, Amazon had opened up Amazon.com's web services for developers to incorporate Amazon content into their applications in 2002.

Speaker 22 But what we consider Amazon Web Services today, cloud storage and compute, launched in 2006. But okay, fancy pans, what did she actually say? We quote Doherty.

Speaker 22 Unfortunately for Bezos, Amazon is now entering a stage in which investors will be less willing to rely on its charisma and more demanding advances to tough questions like: when will this company actually turn a profit?

Speaker 22 And how will Amazon triumph over a slew of new competitors who have deep pockets and new technologies?

Speaker 22 We tried to ask Bezos, but he declined to make himself or any other executives of the company available. He could ignore Barons, but he can't ignore the questions.
Bang a line, by the way.

Speaker 22 Amazon last year posted a loss of $125 million, which is about $242.6 million in today's money, on revenues of $610 million, so about $1.183 billion in today's money.

Speaker 22 And then this year's first quarter, referring, of course, to 1999, as the company posted a loss of $61.7 million, which is $119.75 million in today's money, on revenues of $293.6 million, $569.82 million in today's money.

Speaker 22 I realize that was a real motherfucker of a quote, but it's necessary.

Speaker 22 Her argument for the most part is that Amazon was burning cash and had a ton of competition from other people doing similar things and that analysts backed her up and they really did by the way.

Speaker 22 Again, I quote, the first mover does not always win. The importance of being first is a mantra in the internet world, but it's wrong.

Speaker 22 The ones that are the most efficient will be successful, says one retail analyst. In retailing, anyone can build a great-looking store.

Speaker 22 The hard part is building a great-looking store that makes money, which is a good point.

Speaker 22 Fair arguments for the time, though perhaps a little narrow-minded.

Speaker 22 The assumption wasn't what Amazon was building, and we, by the way, are referring to Amazon.com, the store, was a bad idea, but that Amazon wouldn't be the ones to build it.

Speaker 22 And again, we quote, once Walmart decides to go after Amazon, there's no contest, declares Kart Barnard, president of Bernard's retail trend report.

Speaker 22 Walmart has the resources that Amazon can't even dream about. Which is true at the time, but in simpler terms, Amazon's business model was not in question.
People were buying shit online.

Speaker 22 In fact, this was just before the dot-com bubble burst when people had insane optimism about the future of the web. Yet the comparison stops there.
People obviously like buying shit online.

Speaker 22 It was the business models of many of these web pioneers that sucked. Looking at you, web van.
But we're going to talk about Amazon Web Services.

Speaker 22 And the less technical of you, I want to explain something. AWS is a really important company.
I'll kind of get into those details.

Speaker 22 But people like to argue about it and say, well, it lost a bunch of money. So, you know, that means the generative AI should lose a bunch of money too.
And that's how it works.

Speaker 22 I'm going to substantively and repeatedly explain why that is so goddamn stupid. I'm sick of the argument.
I'm sick of it.

Speaker 22 Breathe, Edward. Breathe.
They can't get you behind the microphone.

Speaker 22 Okay.

Speaker 22 Amazon Web Services was an outgrowth of Amazon's own infrastructure, which had to expand rapidly to deal with the influx of web traffic from Amazon.com, which had become one of the world's most popular websites and was becoming increasingly more complex as it sold things other than books to multiple international locations as well.

Speaker 22 Other companies had created their own infrastructure, but if a smaller company wanted to scale, they basically needed to build their own thing.

Speaker 22 It was a massive barrier between companies and building web services. And it's actually kind of cool what Amazon did.
I hate to look rosy eye behind rose-colored lenses.

Speaker 22 I don't know the phrase at Jeff Bezos, but I don't know.

Speaker 22 Remember, this was early 2000s before Facebook, Twitter, and a lot of modern internet we know that runs on services like Amazon Web Services or Microsoft Azure or Google Cloud.

Speaker 22 They basically invented the modern concept of cloud compute. But we're here to talk about Amazon web services being dangerous for Amazon and people hating on it, the thing that allegedly happened.

Speaker 22 Right?

Speaker 22 I do hope all the people that said this to me didn't just make it up. Oh my god, they did.

Speaker 22 A November 2006 story from Bloomberg talked about Jeff Bezos' risky bet to run your business with the technology behind his website, saying that Wall Street wanted him to mine the store.

Speaker 22 Bezos referred to as a one-time internet poster boy that became a post.com pinata. Fuck, they were so good, but where is this piss and vinegar, by the way? This is fun.

Speaker 22 Nevertheless, this article, which again is linked in the spreadsheet for the episode notes, has what I think my haters crave.

Speaker 22 And I quote: But if techies are wowed by Bezos' grand plan, it's not likely to win many converts on Wall Street. To many observers, it conjures up the ghost of Amazon past.

Speaker 22 During the dot-com boom, Bezos spent hundreds of millions of dollars to build distribution centers and computer systems in the promise that they would eventually pay off with outsized returns.

Speaker 22 That helped set the stage for the world's biggest web retail operation with expected sales of $10.5 billion this year.

Speaker 22 All that has investors restless and many analysts throwing up their hands, wondering if Bezos is merely flailing around for an alternative to his retail operation.

Speaker 22 11 of 27 analysts who follow the company of Underport perform more sale ratings on the stock. A stunning vote of no confidence.

Speaker 22 That number of sale recommendations is matched among large companies only by Quest Communications International Inc., according to investment consultant Starmine Corp.

Speaker 22 It's more than even the eight sale options on the struggling Ford Motor Company. Pretty bad, right? Pretty bad.
My goose is cooked. All those analysts seem pretty mad, except it's not.

Speaker 22 My goose is raw. Yours, however, has been in the oven for over a year.
As one analyst, Scott W.

Speaker 22 DeWitt, noted at the time, the direct costs of providing Amazon web services at first were minuscule, because much of the startup infrastructure already existed.

Speaker 22 It was surplus capacity Amazon already owned, software Amazon already used, and Amazon was in it for the long haul.

Speaker 22 It knew that this would take some time before it became a profitable business unit, as the company was basically scaling up the infrastructure of the internet.

Speaker 22 And by the way, let's just go back to that quote here. The quote says, the costs were minuscule.
The costs weren't the problem. Hey, wait a second.
That's a name. Scott W.
DeWitt. I can look him up.

Speaker 22 I wonder what he's up to right now. Oh, oh, looks like he's working at Wedbush as its managing director of equity research and has said that AI companies would enter a new stage in early 25.

Speaker 22 He said, oh, oh god, just listen to this.

Speaker 22 The second stage is the application phase of the cycle, which should benefit software companies as well as the cloud providers.

Speaker 22 And then phase three of this will ultimately be the consumer-facing companies figuring out how to use the technology in ways that can actually drive increased interactions with customers.

Speaker 22 The analyst says the market will enter phase two in 2025 with software companies and cloud provider stocks expected to see gains.

Speaker 22 He adds that cybersecurity companies could also benefit as the technology evolves.

Speaker 22 I know I meant to be more mature, but DeVit also calls out Palantir, Snowflake, and Salesforce as those who would gain. In none of these cases am I able to see any actual revenue from AI.

Speaker 22 And Salesforce themselves said, according to the information, that they'd see no revenue growth from AI in 2025.

Speaker 22 Palantir also has, as discovered by the Autonomy Institute's recent study, recently added the following to its public disclosures.

Speaker 22 There are significant risks involved in deploying AI, and there can be no assurance that using AI in our platforms and products will enhance or be beneficial to our business, including our profitability.

Speaker 22 What I'm trying to say here is that analysts can be wrong and they can be wrong at scale. There is no analyst consensus that agrees with me.

Speaker 22 In fact, most analysts appear to be bullish on AI despite the significantly worse costs and total lack of growth. But, Ed, Ed, Amazon web services cost money, Ed.
Now you should meet your end.

Speaker 22 Nice try, Chuckles.

Speaker 22 In 2015, the year that Amazon web services became profitable, Morgan Stanley analyst Katy Huberty believed that it was running at a material loss, suggesting that the $5.5 billion of Amazon's technology and content expenses was actually AWS expenses with a negative contribution of $1.3 billion.

Speaker 22 And by the way, want to know what she's up to nowadays?

Speaker 22 I wanted to know because six months ago, she declared that 2025 would be the year of Agentic AI robust enterprise adoption and broadening AI winners.

Speaker 22 So yes, analysts really got AWS wrong, but putting that aside, there might actually be a comparison here. Amazon Web Services absolutely created a capital expenditure strain on Amazon.

Speaker 22 From Forbes' Chuck Jones, in 2014, Amazon had $4.9 billion in capital expenditures, up 42% from 2013's $3.4 billion.

Speaker 22 The company has a wide range of items that it buys to support and grow its businesses, ranging from warehouses, robots, and computer systems for its core retail business in AWS.

Speaker 22 While I don't expect Amazon to detail how much goes to AWS, I suspect it is a decent percentage, which means Amazon needs to generate appropriate returns on the capital deployed from AWS.

Speaker 22 In today's money, this means that Amazon spent $6.7 billion in capital expenditures in 2014, likely on AWS. Assuming it was this much every year, it wasn't.

Speaker 22 But I want to make an example of every person claiming that this is a gotcha.

Speaker 22 It took $67.6 billion, and that's in today's money, and about nine or ten years of pure capital expenditures, even though all that capex wasn't just AWS, to turn Amazon Web Services into a business that now makes billions of dollars a quarter in profit.

Speaker 22 And that's $15.4 billion less than Amazon's capex for 2024.

Speaker 22 And even less than the $105 billion they spent this year. It's a fucking joke.
And to be clear, the actual capital expenditure numbers are like the AWS cost in totality were likely much lower.

Speaker 22 I just want to make it clear that even when factoring in inflation, AWS was A, a bargain and B, a fraction of the cost of what Amazon has spent in 2024 or 2025.

Speaker 22 Here's a funny little thing. On March 30th, 2015, New York magazine published a piece from none other than than Mr.
Kevin Roos about the cloud compute wars in which he claimed that, and I quote,

Speaker 22 there's no reason to suspect that Amazon would ever need to raise prices on AWS or turn the fabled profits which the pundits have been speculating about for years.

Speaker 22 Less than a month later, Amazon revealed that Amazon Web Services was profitable. They don't call him the most right man in tech journalism for nothing.

Speaker 22 I think it's so funny when you go back and read all of Kevin Roos's stuff, how many just like rates he steps on and how quickly they whammy him in the face and how nobody says anything.

Speaker 22 I'm saying something. But here's the good news.
We're at the end of this first part.

Speaker 22 Next episode, we're going to continue exploring the comparison between AWS and generative AI and talk about why that comparison fundamentally doesn't work and why everything's kind of brittle.

Speaker 22 I'll catch you on the flip side. Thanks for listening.

Speaker 22 Thank you for listening to Better Offline.

Speaker 26 The editor and composer of the Better Offline theme song is Matasowski. You can check out more of his music and audio projects at matasowski.com.
M-A-T-T-O-S-O-W-S-K-I dot com.

Speaker 32 You can email me at easy at betteroffline.com or visit betteroffline.com to find more podcast links and of course my newsletter.

Speaker 32 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.

Speaker 31 Better Offline is a production of CoolZone Media.

Speaker 31 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.

Speaker 4 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.

Speaker 27 There was the six-foot cartoon otter who came out from behind a curtain.

Speaker 8 It actually really matters that driverless cars are going to mess up in ways that humans wouldn't.

Speaker 23 Should I be telling this thing all about my love life?

Speaker 7 I think we will see a Twitch stream or president maybe within our lifetimes.

Speaker 1 You can find Close All Tabs wherever you listen to podcasts.

Speaker 23 Ah, smart water. Pure, crisp taste, perfectly refreshing.

Speaker 2 Wow, that's really good water.

Speaker 23 With electrolytes for taste, it's the kind of water that says, I have my life together.

Speaker 20 I'm still pretending the laundry on the chair is part of the decor.

Speaker 23 Yet, here you are, making excellent hydration choices.

Speaker 2 I do feel more sophisticated.

Speaker 23 That's called having a taste for taste.

Speaker 20 Huh, a taste for taste.

Speaker 31 I like that. Smartwater.

Speaker 23 For those with a taste for taste, grab yours today.

Speaker 11 Every business has an ambition.

Speaker 3 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.

Speaker 13 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.

Speaker 19 When it's time to get growing, there's one platform for all business: PayPal Open.

Speaker 17 Grow today at PayPalOpen.com.

Speaker 20 Loan subject to approval in available locations.

Speaker 29 I'm Bretski, and you're in California, and I'm here to tell you about SpinQuest.com, my favorite social casino.

Speaker 29 With over a thousand slots and table games, absolutely free with the ability to win real cash prizes instantly to your bank account.

Speaker 29 There's no better time to hop on our $30 coin package for only $10 deal. Head over today.
I love you. I'll see you there.

Speaker 33 SpinQuest is a free-to-play social casino. Voidwear Prohibited.
Visit SpinQuest.com for more details.

Speaker 1 This is an iHeart podcast.