Tech Earnings Quadkill, Red Button vs Blue Button | Diet TBPN

0:01John Coogan:What a day in financial markets and technology markets. There is a ton of news today. Ger Gavin summed it up well. A huge day tomorrow. He's posting this yesterday. 10 a.m. Canada interest rate decision. 2 p.m. This is Eastern Time USA interest rate decision. That is in. Fed held rates constant. So sort of a nothing burger, I guess. not the best news. I think some people were hoping for a cut, but everyone sort of expected this. It met expectations, but the news is that just five minutes ago, the Federal Reserve has kept interest rates on hold at a range of 3.5 to 3.75 percent at Wednesday's meeting.

0:45John Coogan:And Jerome Powell is giving a speech in just 30 minutes at 2.30 Eastern. Then at 4 p.m., Google earnings, Amazon earnings, meta earnings and Microsoft earnings. It is a massive debt. And no big deal. No big deal. They only represent just under 20 % of the total market cap of the S &P 500. Let's go. All reporting within. But people are optimistic. Semi-analysis put out a note this morning, according to YC, YieldChad, expecting hyperscaler CapEx to be revised upwards and beat street expectations as hyperscaler cloud revenue is accelerating. And they are seeing positive ROI on cloud investments. I can sort of go through my earnings preview.

1:26John Coogan:It's a tech earnings quad kill today. And the big question is just how is the AI build out going? Obviously, the CapEx numbers were huge. We saw the first time ever we've seen a$200 billion number from Amazon last quarter. But everyone's up in the, is there even a word for triple? It's like, what, 11 digits or something like that? 100 billion. A lot of digits. Everyone's spending at least 100 billion these days. at least if you're in the Mag-7. But financial performance has actually been strong even in sort of legacy areas such as search, Google Search is growing, e-commerce sales, Amazon core business is growing, enterprise software seats, Microsoft 365.

2:09John Coogan:We're gonna have more news on all of that. Even though all of those businesses, they're working, they're chugging along. The big question is around durable revenue tied to AI infrastructure because you have this matching problem. You spend a bunch of money, there's depreciation. When do you actually get the cash flow back? How durable is this revenue? What are the moats around this revenue? What do growth rates and margins look like in this new era of something that looks maybe a little bit more like a railroad business, an oil business, as opposed to something like you build a website, people just show up, and it's 80 % margin, which was the dream of the previous software era.

2:45John Coogan:We're going into an entirely different era, but it is very exciting, and we're getting a lot of data today. So everyone has seen cash flow from these hyperscalers to NVIDIA, to power companies, to data center builders. But everyone's wondering, what's the exact conversion cycle to higher revenues and higher profits? What's the pathway there? Because you don't want to just be drawing down on cash endlessly. Eventually, you stop making money entirely. So let's start with Google. Google has the most fully integrated AI stack, arguably. They have consumer distribution. They got model training with DeepMind.

3:20John Coogan:They have custom chips with the TPU and a bunch of product services where they can stuff AI features. Google Workspace. Google Workspace. Underrated. Search, YouTube, Android, cloud. They can deploy solutions all over the place. And so the flywheel should be spinning very, very quickly. The key question that investors are asking is, does AI change the unit economics of search too quickly? Does AI do LLMs, do AI search overviews, Gemini generally? Are they able to monetize those results fast enough to offset any potential declines in search ad revenue? And so people will be looking for how is search monetizing?

4:02John Coogan:How are the new AI disruptors monetizing? But there are tons of places to pick up growth, even if growth does slow down in the core search business. tons of opportunity in cloud. But the question again is AI overviews in Gemini, are they expanding search usage? Are they increasing ad ROI? Or are they compressing the model, the financial model? Microsoft is also coming off a strong quarter. Honestly, everyone's coming off strong quarters. Everyone's doing very well. Revenue is up 17 % at Microsoft with cloud, which includes Azure, M365, some LinkedIn stuff. It's a big bucket. Microsoft cloud is growing at 26%.

4:41John Coogan:But if you dive in and you double click on Azure, Azure is growing at 39%. And of course, Azure is a bigger lever on CapEx, which is run rating around 150 billion. Not bad. The biggest number in the Microsoft earnings is RPO, remaining performance obligations. Last quarter, it was listed at 625 billion, up 110%. That was the eye-popping number of the last earnings. About 45 % of that is coming from OpenAI, but they have lots of other partners that have signed on for really long compute contracts. And that is starting to show up in the financials saying, hey, we're spending all this capex, but we have this RPO and we have these deals signed where companies are, it's not just us that we're forecasting some really high growth here.

5:26John Coogan:The entire industry is forecasting high growth. And so we have done deals to justify the capex that we're spending right now, even with depreciation, which seems like a less of an issue than people thought it was since H100 still seemed to be monetizing just fine, but we can go into that. So Microsoft has the cleanest read on enterprise AI monetization, which I think is something people have been really looking for, looking for numbers. They know that subscription, LLM chat apps monetize at a decent rate, that the margins at a lot of these companies are okay, the token, the APIs are working. But what does it actually mean to deploy AI into the American economy, into everyday businesses, into large-scale businesses.

6:04John Coogan:All of those businesses are on Microsoft overwhelmingly. And so there are a ton of data points that can help you understand how AI is flowing through the global economy, but also the American economy. So what are we talking about specifically? Azure growth, how much cloud hosting is going on, gross margins for Microsoft Cloud, that's very important. Are you seeing cloud compression? Because if you go to your Microsoft Cloud provider and you say, hey, I'm gonna, this Teams thing, I'm paying a lot of money for it. I actually vibe coded something. You got to give me a discount. That would show up in cloud margins.

6:35John Coogan:Will it show up? I don't think it will, but we'll see. I think it's going to be fine. Copilot adoption and ARPU. How much are they actually rolling these out? Are they actually getting incremental spend? Are companies willing to send more of their hard-earned dollars to Microsoft for better AI services, better features, copilots? Then M365 seat growth. That's a really important one. Are people adding more seats? Are they hiring? Like we've seen some layoffs in big tech, but how is the overall economy doing? How many more seats are being rolled out? Again, with the question of like, does your AI agent need a seat and so you have more seats?

7:12John Coogan:Or does your AI agent replace 20 seats and then you only have one seat and you have 20 agents. They're all logging into the same M365 seat. These are like more long-term questions, but we're getting an early read today. GitHub co-pilot momentum. This will all paint a picture of what is happening with AI adoption in enterprises broadly. Interestingly, Copilots at Microsoft, it hasn't been the most hyped product. GitHub Copilot, early, very, very early to the party. Huge run rate. Rocketed up to 500 million ARR very, very quickly. But the horse race has always been, you know, windsurf cognition and cursor and codex and cloud code.

7:51John Coogan:And like the battle has been the startups for the most part. Gemini's been in there. GitHub Copilot has felt like it hasn't been dominating the narrative, but we're going to find out, is it still growing? Because the market is really, really big. It's possible that everything's growing, even if there's, you know, a horse race back and forth between the leading labs. There will be a bunch of other questions answered around, you know, how nuanced is the diffusion adoption question? So Microsoft has an incredible go-to-market team, an incredible go-to-market motion, enterprise sales motion. And so is there an advantage that they have that they can press GitHub Copilot, even if it's not the sexiest product next to whatever's hot this week?

8:33John Coogan:Can they get that into people's teams? Same thing. Yeah. Yeah. Slack was definitely like the hot one, the hyped one, and teams wound up doing very well. And so we'll get a stronger read on what's going on there. And then, of course, that M365 seat growth will tell us a lot about what does the future look like for seat-based enterprise software, the SaaS model broadly. Because if all of a sudden that's falling off a cliff, well, it probably doesn't look good for other seat-based SaaS companies. For Amazon, everyone wants to see strong AWS acceleration to justify the MAG7 topping CapEx numbers. They got it at$200 billion in CapEx in 2026.

9:13John Coogan:So expect a lot of focus on AWS revenue growth and margins. Q4 was healthy, though. Net sales up 14 % for Amazon. AWS growing at 24%. The sneakily huge ads business over at Amazon grew 23 % to$21.3 billion. They're almost making$100 billion a year just on ads. That's remarkable. And that's a lot of cash flow to fund CapEx and other AI initiatives. And so operating income overall was$25 billion for last quarter. And they generated free cash flow of$11.2 billion. So still huge cash flows. They're obviously drawing down on those. And that$11.2 billion number was down because of increased AI spending.

9:59John Coogan:So if AWS accelerates, all the CapEx looks like buying scarce capacity ahead of demand. Like they will be GPU rich or just compute rich generally at a time when it's good to be GPU rich and compute rich and they will look like geniuses. So people are hoping for strong AWS acceleration. Consensus for AWS revenue is around$36.7 billion with growth in the mid-20 % range. But the market's really hoping that it starts with a three. Everyone's hoping for 30%, something like that, at least this year at some point. You know, re-acceleration would be a treat for the market. Lastly, Meta. Meta is an interesting spot.

10:41John Coogan:Super clean Q4. Nothing really to prove today. Despite all the thud, all the fear, uncertainty, and doubt around the talent wars, new team members, the new lab, the sometimes clunky model releases. They didn't get behemoth out. You know, different adoption of Meta vibes, the video app. Like, there's been all these questions about the AI strategy. All of that needs to be put to the side in the face of just AI is already working for Meta and their ads platform. Let's get an update on Reels. Yeah, they're absolutely crushing it. So 3.58 billion DAUs, daily active users, and they grew the ad business a ton.

11:21John Coogan:And so the revenue overall grew 24 % to almost$60 billion last quarter. Ad impressions rose 18%. Average price per ad rose 6%. family of apps operating income was$30.8 billion, which makes losing$6 billion at reality labs quaint. It's just like, who cares? It's totally worth taking a side bet on the future of devices, and maybe you get a platform out of that. It makes a ton of sense. And so CapEx was$72.2 billion last year. The guide this year is somewhere between$115 to$135. I think we might see that tighten up today because it's a little bit wider than some of the other hyperscalers that are targeting, but clearly a near doubling, almost doubling of CapEx.

12:05John Coogan:What are you laughing about? There's one possibility where he goes, you know, he blows it out. 250. We can't, we can't, we can't count that out. We can't count that out. I mean, he's got mirrors in space. They came out with a solid model. He's putting mirrors in space. He wants to be a player. Yeah, yeah. Yeah. I mean, a lot of focus is on the new meta models, but all of that, just financially, at least in the quarterly earnings, like it will just take a backseat to what's going on with the AI in the ad placement, ad monetization funnel, because that's where meta makes so much money. And so the question is, how much more juice will AI bring to the ad business?

12:48John Coogan:Expectations imply revenue growth of around 31%. And the question is always the same. Is there enough incremental growth to justify the capex? An AI advancement delivers better ad performance basically immediately. And this is the good news for meta. So improvements in AI move the needle at meta incredibly quickly. Like a new model can come out and there can be some breakthrough in like thinking, reasoning, coding agents. And if you're in the enterprise software world, it can take time to do a deal, roll it out to different developers, different organizations, change management, understand the guardrails, the security.

13:25John Coogan:How does this actually flow through an enterprise to drive incremental demand? Not so at Meta. If there's a breakthrough and they have a model that is placing ads more effectively, they can quickly A-B test that, run that across the entire family of apps and you don't even notice it as a consumer. You're just like, oh, like I was shopping for a car and I saw an ad for a car. I was shopping for a shirt and I saw an ad for the perfect shirt and the ads just got a little bit better and just a little bit better means like another $20 billion in profit. So it can be, so they just don't have a diffusion question.

14:01John Coogan:They don't, they don't have a diffusion problem. There's no, oh, we have this amazing model. If only we could get people that open Instagram to use it. Like there's nothing there because they use it when they scroll and they see an ad. And so that's an incredibly fortunate place to be in. So every company is trying to answer the same question. Can you turn AI CapEx into proprietary distribution, higher customer retention, measurable revenue growth before the depreciation catches up? It's the Super Bowl for big tech. So get your popcorn ready. Everyone's excited. There are AI worries, AI jitters in the Wall Street Journal.

14:38John Coogan:AI worries have returned to Wall Street, now come earnings. And so everyone is going back and forth. On the open AI news, shares of Oracle, CoreWeave, and SoftBank slumped. But CoreWeave is up majorly today, 8.4%. They dropped at least 4 % after hours yesterday, but then they're back up. That revived investors' worries that technology giants' massive investments in artificial intelligence won't produce the blockbuster profits many expect. The report was particularly jarring coming ahead of key earnings from major players, Alphabet, Amazon, Microsoft, Meta, all scheduled to report Wednesday with Apple following on Thursday.

15:18John Coogan:The ice is thin, the leash is very tight, said Dan Morgan, portfolio analyst at Synovus Trust. Any evidence that would come out that would add doubt about OpenAI, Anthropic, or any of these companies is obviously going to create a sell-off. Morgan said he hadn't adjusted his positions on Tuesday and didn't think investor concern was broad-based because companies, including IBM, Texas Instruments, and Intel had reported strong earnings in recent days. Instead, Tuesday's losses centered on the companies with biggest stakes in OpenAI's business. OpenAI had forged close ties with these companies to secure funding and gain access to computing resources, which are essential for training AI systems and providing answers to queries and executing user requests.

16:03John Coogan:And so Oracle and CoreWeave, we talked about, fell. OpenAI has defended its financial footing and said its leaders are aligned on securing computing resources. The business is firing on all cylinders and the mood internally is incredibly positive, the company said in a response to the journal's article on Tuesday. Colin Tedard says, my children might not eat tonight if this doesn't go well because he is excited for all four of these earnings on Robinhood. and we have Vlad from Robin Hood coming on the show later today. Kramer says Intel is such a horse. I have no bear case. I have no bear case.

16:41John Coogan:And Intel is up another 10 % today? Yeah, up 10 % today. Wow. 40 % on the week so far. 40 % on the week. Is that good? Wow. Based on everything you've seen so far in your six-month career, Tyler, is that good? That seems incredible. Looks pretty good to me so far. Yeah, I mean, there is a, there's like the idea of a CPU shortage generally is just being digested by the market right now. I think it took, I mean, I remember Nat Friedman and Daniel Gross going on Stratechery and talking about after ChatGPT saying like, yeah, NVIDIA seems like it could be important. Like quadrupled in value. and this tends to happen as people digest the new reality of demand around certain pieces of the AI supply chain.

17:33And we have an interactive game set up?

17:36John Coogan:We do. So there has been a viral, I don't know, what is it, a brain teaser? It's like a poll. A poll, but it's a thought experiment. It's a thought experiment. It's not purely a poll. It's a brain teaser. It's a thought experiment. and did this has this existed before did this start with it did this didn't start with mr beast right no so it started with tim urban yeah tim urban i i don't think tim urban like created this either but i think his post was the one that first went like really big okay i don't think this is new i feel like i've heard about this before at some point yeah but tim urban was the first one to make it like go viral this yeah cycle so he got 24 million views on april 24th and then mr beast came from behind four days later with 10 million views on a very similar question.

18:29John Coogan:Let's start with the original Tim Urban question, and we will ask everyone in the chat to participate and vote for what you would pick in this thought experiment. So here is the question posed by Tim Urban of Wait But Why. Everyone in the world has to take a private vote by pressing a red button or a blue button. If more than 50 % press the blue button, everyone survives. If more than 50 % of people press the blue button, everyone survives. If less than 50 % of people press the blue button, only the people who press the red button survived. Which button would you press? What are you, what are you pressing so so the sort of like normie answer is you immediately just go blue because it feels like a lot of people would go blue but like the safe answer is you just if you if you're hyper rational you'd be like obviously you hit red because if everyone's just super rational everyone hits red and then everyone's all good you don't hit over 50 right yes yeah because like no matter what so if i press red no matter what anyone else picks like i survive right Yes.

19:40John Coogan:But there is a risk. But but like, you know, if you want everyone to survive, then you need 100 percent of people to pick red. Right. Yes. Yeah. But OK. So OK. So originally I was blue. Yeah. I think I was like complete normie. Right. And then I went to Jordy's take. Yes. I was like, wait, no, that doesn't make sense. Like, even though I mean, we'll see what happened in the poll. Right. But, you know, I was like, OK, if I'm thinking about it super rationally, I should just pick red. Right. Then I save myself 100 percent of the time. Yeah. But then like if there's some people that pick blue, then like they're gonna die for sure.

20:11Like I'm basically voting against them.

20:13John Coogan:Yeah, you're sort of, you're complicit in their demise. If you want everyone to live and you're going for red, you need 100 % of people. But if you're going for blue and you want everyone to live, you only need 51 % of people, right? So you need to convince way less people. Quickly, let's tell the chat how to vote. Yeah, so one is for red. One is for red. I forgot the term been here. And two is for blue. Two is for blue. So type a single character, just a one, if you want to vote for red and just a two if you want to vote for blue. Oh, do we have this up? Ooh, we are tilted towards red. This is crazy.

20:49John Coogan:This is not what happened with Tim Urban's and with Mr. Beast's because Tim Urban, 58 % said blue, blue won, no one died. And with Mr. Beast's, 55.7 % went blue. No one died. It's not looking good for the blue team here. 33 % of the chat is going to get cooked right now. It's possible a nation state is voting. Yes, that would be very, very high risk. Mike Solana says, for what it's worth, and this is embarrassing, but I'm going to admit it. My instinct was blue, and I pressed blue. Then I thought about it for a moment, and the answer was clearly red. I can't make a rational case for blue, but understand where blues are coming from.

21:37John Coogan:Parentheses, they are wrong. So your final answer, you flipped. I think I'm still going blue. You're still going blue. Because I think you have to assume that there's going to be some variance. Some people will just pick the wrong one. Yes. The correct one is red. If everyone is 100 % logical, everyone should be picking red. Yes. Yes. If 100 % of people are rational, they will all pick red and no one will die. Correct. But clearly some people are not going to do that. If 90 % of people are doing it, if 90 % of people are rational, then everyone should just be picking blue because you want everyone to live.

22:11John Coogan:Yes. Yes. And then the people that make the mistake, they go for the red, they are saved anyway. But there are some folks who are sharing different illustrations. Cremieux has one here. He says, if this was how the buttons looked, what portion of humanity would press blue. And he has two images pulled up. One is the red button, nothing happens. And the other one is the ultimate death gamble. Press blue. You enter the ultimate death gamble. You and everyone who presses this button dies unless more than 50 % of people choose to press this button. And so that makes it pretty clear. I think this is the Mike Solano position as well.

22:55John Coogan:You got to press red in this case because nothing happens. But that's not entirely true. And so people have sort of flipped this around because pre-rat here has the opposite, which is instead of the ultimate death gamble, it's the ultimate murder gamble. So blue, if everyone presses blue, nothing happens. Red is the ultimate murder gamble because you initiate the ultimate murder gamble, everyone who pressed blue will die unless 50 % of people choose red. And so it goes back and forth and you can see both of those. Do you want the blood on your hands from playing the ultimate murder gamble? Or do you want the risk of the ultimate death gamble?

23:34John Coogan:I think it's underrated to just be a little thrill seeking and want to press blue because you're bored and you want... No, you can tell blue is the correct answer just because of the massive polls where over 50 % of people picked blue. That's true. That's true. Pick blue because you don't want to kill yeah you know once you're at mr b scale this is like a real world simulation it is it is well we're still we're still red uh cook uh simone say it says red button pushers will leave their wives and children to fend for themselves in disastrous situations because their kids can't run fast enough mike solana says blue button pushers are literally leaving their wives and children behind to fend for themselves.

24:15John Coogan:They aren't guaranteed to push red. Yes, people are going back and forth. Red cells won't believe this, but some people make a distinction between nothing and mass genocide. So they see pushing the red button as effectively committing mass genocide because you are potentially condemning the blue button pushers to death in this scenario. Self-Maxer shares a version of the trolley problem where the red team is not on the rails, but the blue team is pushing the lever to save themselves. Is this an accurate description? Is this a fair contextualization around this? It's obviously charitable to the red team, which, of course, the TVPN audience seems to be represented by, standing next to the trolley lines.

25:06John Coogan:lines and not standing on the rails. But you would be there, Tyler. You would be a blue button pusher pushing on the trolley. It seems a little risky in this scenario. It does. When you have the option just to stand there and watch the trolley go by, it's all about frame of mind here. No, but you can tell that from the polls, from the big polls, you can just tell that blue is the correct option. Potentially. Do you know that beforehand? No, you don't. But those people who voted in it did not know beforehand yeah and so if if 10 million or how many hundreds of thousand people voted yeah i think you can just assume that that's the question is like if this actually was a real vote yes would people spend a little bit more time thinking about the problem and would that influence their yes yes yes answer it is it is a very social media poll people are scrolling they see mr beast ask a question they're just like curious there's no real stakes yeah and so they just answer quickly.

26:02But the blue is very much a, it's a faith in humanity restored moment,