The Six Five Pod | EP 292: Capital Flood, AI Disruption, and the Real Risks Ahead

The handpicked topics for this week are:

1. $650B+ AI CapEx Surge and the Bubble Debate: Hyperscalers are dramatically increasing infrastructure investment, raising questions about near-term returns versus long-term survival. Is this a bubble, or is it the cost of staying relevant in the AI era?
2. Anthropic’s Valuation, AI Acceleration, and Regulatory Pressure: Private market enthusiasm continues as Anthropic’s valuation climbs and CEO Dario Amodei sounds alarms about AI’s pace and societal readiness. At the same time, regulatory funding and lobbying activity intensify.
3. AI and Job Displacement: The “All At Once” Flip Debate: A viral essay argues AI could upend white-collar work within years. The hosts debate whether integration, friction and regulatory industries will slow disruption, or whether society is underestimating the speed of change.
4. The SaaS Rerating and Software Model Reset: With AI agents building workflows and coding environments evolving rapidly, markets are reassessing SaaS multiples. Is this a temporary repricing or a permanent structural shift?
5. AI Infrastructure Financing: Alphabet’s 100-Year Bond: The company’s $31B debt raise, including a century bond tranche, signals long-term conviction in its infrastructure build-out. Demand has reportedly exceeded supply by multiples.
6. Semiconductor Expansion and the Trillion-Dollar Forecast: Applied Materials highlights accelerating semi-cap demand, reinforcing projections that semiconductor spending could approach $1 trillion faster than previously expected.
7. Energy and Data Center Constraints: As compute scales, power becomes the gating factor. Energy-linked data center plays are emerging as strategic infrastructure enablers.

For a deeper dive into each topic, please click on the links above. Be sure to subscribe to The Six Five Pod so you never miss an episode.

‍

Listen to the audio here:

Disclaimer: The Six Five Pod is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors and we ask that you do not treat us as such.

‍

Patrick Moorhead: 

Six Five Pod is back and we are jacked. It's great to be back. We had an earlier Monday podcast. Hopefully you all enjoyed that. Uh, we had to do that because we went from CES to Davos to, gosh, the UAE. But, uh, we're back on schedule. And I can't believe a week ago I was in an airplane, a 14 hour ride from the UAE.

‍

Daniel, how are you doing? My buddy, my buddy, my friend.

‍

Daniel Newman:

We’re back and we’re jacked. And someone here is older than the last time I am.

‍

Patrick Moorhead:

Maybe on the calendar, but, uh, physiologically, we will see. We will see. No. It's great. Took about a week to get recovered. And yes, I have celebrated a birthday. Uh, but I feel like I am younger, a lot younger than I was five years ago.

‍

Daniel Newman: 

Well, you look. You look better, I don't know. I, I had, uh, it was that Monday. I did my my my most recent in-body. I think I sent it to you. I don't think you said anything back to me, but it was pretty epic.

‍

Patrick Moorhead:

It was, uh, a little jealous. You know, Daniel, I'm getting a little fat, and I'm trying to figure out why the fat's going up. My strength is going up big time, like all vectors. And I don't know if it's related. All my electronics tell me that, uh, I'm burning more calories than I'm ingesting, so I need to figure this one out.

‍

Daniel Newman: 

Buddy, it might be time to pivot from one, uh, functional medicine to to optimize health and performance. That might be where you go to get …You know.

‍

Patrick Moorhead:

Can I get a, uh can I get, uh, can I get a Rita prescription?

‍

Daniel Newman:

Can’t say that online? I think we'll get the entire show will get banned. I don't think you're allowed to even say the word and out loud.

‍

Patrick Moorhead:

I didn't, I didn't, I just said, Rita.

‍

Daniel Newman:

Rita, we'll call it re three A or something. I think that's what they do on Instagram. Yeah. There are definitely the old, uh, peptide arooney's. Uh, Huberman had a big one this week. Said it's going to it's going to break every.

‍

Patrick Moorhead:

Trillion dollar business. Right?

‍

Daniel Newman:

Trillion dollar business. Right. You want to be buying Eli Lilly right now.  You want to be buying Eli Lilly?

‍

Patrick Moorhead

Um, I want I want mine made in a bathtub somewhere in China and sent over and …

‍

Daniel Newman:

And China. Do you say that? China.

‍

Patrick Moorhead:

Yeah. China. China, China. 

‍

Daniel Newman:

It's good to be back. It's good Friday. I. I did get a finally got a good sleep. You know, I had like, two nights in a row where my sleep score was under 60. Um, don't know why. Just couldn't get, you know, just couldn't get into any sort of good restorative sleep. And I did great on the two days.

‍

And then actually, like, I just crashed. Like last night, I just, I think I fell asleep at 8:30 on my couch, um, made it into bed and I didn't get out of bed until seven.

‍

I mean, for me, I skipped my workout. I mean, that's that to to know, like how? Like crazy that is. Because normally I'm up at 4:45 and I'm in the gym by 5:30. I skipped my workout to sleep. I did get nine hours of sleep and I got four hours of restorative last night, so I should feel great. But I actually might have gotten too much because I'm actually having, like, this.

‍

Weird. Like, I'm in a in a I'm almost in a fog, so you're gonna have to bring me out of it.

‍

Patrick Moorhead:

I will, we will. We got a big week here. I mean, a lot of stuff going on, and I think the way I would summarize it is capital control and commercialization of AI. Money is flooding into infrastructure labs and shaping regulation. Agents are moving from assist to operate. And it's interesting on the market, investors are questioning the belief in AI, but they're questioning the payback period.

‍

We're going to hit that in a lot of segments. Hey, some stuff to look out for. Daniel and I are going to be in MWC in Barcelona March 2nd on the March 5th. We're going to be looking for all the AI network narratives. Of course, you've probably seen the alignment between Nokia and Nvidia. We're probably going to see a lot of device AI, device positioning if they can get enough volume with the memory memory pullbacks.

‍

So anyways, let's jump in to the Decode Segment.

‍

All right. One of the biggest questions out there $650 billion in CapEx Daniel I think that's 2 to 3 times. Um, last year what the heck is going on? It's just it's just just another proof point of a bubble.

‍

Daniel Newman:

I think it's the. I think it's the opposite, but I think it is a drain near term. On operating cash flows, it does create some risk. You've got companies that have huge profitable businesses that are raising debt, 100 year debt coming from from Google. Um, you know, you kind of saw Oracle as maybe a bit of the first shoe.

‍

And Oracle, you know, announced all of its fundraising. And of course people are worried about cash flow. So this is one of those kind of infle ction moments that, you know, creates some panic within investors. But it's also one where you have to kind of ask, are the people running these companies smart and do they know what they're doing?

‍

Um, because, yeah, I mean, if you're looking at short term cash flow, uh, spending this much money is going to impact these businesses abilities to show returns. They're not going to buy back as much stock. They're not going to be able to potentially give the immediate returns to shareholders. But if you believe in the transition and transformation revolution of AI, You have to kind of consider the fact that these companies are investing in, uh, preventing their own extinction.

‍

They all understand that if they do not make these bets, they do not invest heavily enough that they will not have the flexibility to, uh, operate through this, uh, through the unknown, through the, the, uh, mysterious and fast paced transformation of a pet. Here's what I look at. You asked me about. Is it a bubble?

‍

I actually think what we've heard the last two weeks of AI is eating. Every business, in every industry is the ultimate proof point that AI is not a bubble. And this spending is the requirement to basically participate in the game. These companies are the most valuable companies in the world. They have to spend.

‍

Nvidia is going to be the biggest beneficiary of this spend. I think something around 50% of this goes straight to Nvidia. But, um, I just I don't think they have an option. I don't think they have the alternative to not do it. And my biggest proof point so far has been the one company that kind of tried to. Well, there's two companies that I think have made under spending mistakes.

‍

One is Apple as a whole, and there's been a lot of people that have put a risk around Apple, even though they have the deepest moat in their kind of user base community, which is what they'll have to monetize. And then Amazon, which decided to go light on CapEx early in the cycle. Yeah.

‍

Patrick Moorhead:

And they paid the biggest price.

‍

Daniel Newman:

And now they're having to spend more later in the cycle than everyone else to play catch up. But people want it to be a bubble, but it's not a bubble.

‍

Patrick Moorhead:

Yeah. It's interesting. Um, I do think the spend is warranted, but I do understand, as you call the bubble bearers, because here's the fact, Daniel, we have not seen anything near to the downstream impact yet. Uh, we have anecdotes of, of anthropic, uh, Creating cloud core work with no coding. By the way, my my coder son says that that there's a lot of caveats on that big time.

‍

You got to have the ability to integrate that into everything else that you've built. And we really haven't seen a, you know, we would see a big efficiency improvement inside of enterprises. We would see a lot more than these random layoffs, um, out there. So I do understand it, but I do I am a believer in, uh, in the investment.

‍

Um, the interesting part about this is that the giant sucking sound of, of, uh, chips and infrastructure and memory is pulling away from everything else. Phone makers aren't going to get what they need. PC makers, they're last in line. Right? Windows PCs is going to be the last in line of of, uh, for this.

‍

The good news is the memory guys are taking a mature look at this and not starving a certain businesses, realizing that they don't want to kill an entire category of of of products.

‍

Daniel Newman

Oh, by the way, did you see yesterday? I think the AMA CEO said they think semi spend could hit a trillion this year.

‍

Patrick Moorhead:

Yeah. It's incredible. Don't you think power is going to be the next, uh, thing to to go here. Isn't going to be the next big discussion.

‍

Daniel Newman: 

I think everybody kind of knows that, especially here in the US, maybe not in China, which is interesting because I remember I said something about it possibly hitting a trillion, and I had I got some real hate mail about that. Um, but I think part of this is just ASP, right. When you have all the you talked about this shifting from consumer PCs devices to much higher price and higher margins like HBM, the the price tags are going up substantially.

‍

Um, it's just really interesting. But I mean, clearly all these smart CEOs and their boards all think it's the right thing to do is to spend.

‍

Patrick Moorehead

Yeah.

‍

Daniel Newman:

I mean.

‍

Patrick Moorhead:

You have to you have to spend. There is there is no option. Hey, in the context of spending a bunch of money in the saga of the AI company of the week, that's the darling, right? We went, you know, for two years that, um, um, OpenAI was going to run away with everything. You had Gemini come in. They got, you know, bloodied and bruised with their early models, and Gemini three comes out and and then everybody whipsawed over the other side of the room, and then, uh, Claude, um, uh, basically opus comes out clawed co work and stuff like that.

‍

And then, you know, we're whipsawed over the anthropic and Blackstone, uh, boosted stake to $1 billion. To put it about, implies a $350 billion valuation there, and at least according to Reuters. Anthropic is at a $14 billion run rate. Cloud code alone is over 2.5 billion run rates, with subs quadrupling, um, every every year.

‍

Um, and, you know, it's interesting what makes this special. Like when Blackstone starts investing, um, it's not, uh, I think it changes the game to, to kind of a financing strategy. Right. That's when Blackstone, uh, gets involved in these types of, uh, of of deal. And I'm very interested on anthropic to see where they go in the enterprise.

‍

You know, Microsoft looked like the absolute clear runaway, uh, winner. And they still are the preferred, um, agent in uh, inside of of of enterprises. And, uh, anthropic is a, I think a number of 4 or 5 player at this point, but I'm wondering if they can move. Uh, move forward. Even Microsoft is even introduced a anthropic in into their, um, application as a service development tools.

‍

Daniel Newman:

Yeah. I mean, look, the in the dev environments, the like Microsoft or Bedrock and you know, there's always going to be some model, uh, agnosticism that's going to exist because I think different models do do different things. Well, anthropic has done a great job. I mean, this this run up and by the way, being able to raise this much money at this valuation and to your point, uh, diverse kind of set of companies getting in.

‍

But, I mean, this just goes to show. There's probably one of the biggest disconnects right now between like public and private markets, where like the private markets are almost insatiably willing to invest into these startups. It's super high multiples, and public people are devaluing AI right now in some strange way.

‍

Um, and obviously there's the difference of near and long term. But everybody, I think, kind of knows that anthropic gets 380. And then they do a public at 600, and they're kind of looking at just the step functions to exit liquidity, whereas public markets, um, you know, are more or more, uh, they throw tantrums and they're more judgmental on the daily.

‍

But, uh, I mean, Anthropic is crushing it. They really are. Their models are really good. I mean, you know, you do a lot of this, and I know you've said how you kind of play with it and you play with MCP and agents, and then you play with just sort of directly prompting things to like create stuff for you. And I mean, when I'm, when I, when I want a good outcome.

‍

I put opus 4.6 is incredible. I mean, it just it just does an incredibly good job. I mean, what they've done shaking up these different industries and to your point, like everybody wants to oversimplify like, yes, you're not just going to turn Claude coworker on and instantaneously replace copilot across your entire enterprise.

‍

It's not how it works. There's going to be a ton of work to do. Um, for a small business, though. Like like ours, like smaller businesses. Like, you can do this. Like you absolutely can do this. It's in big, large companies on large enterprises. And that's why I think all the kind of software selling and down pressure is not nearly as significant as people want to make it out to be, is because big enterprises can't move like that.

‍

They can't apply the tools that quickly. Um, so Anthropic deserves props, though, man. They've been super disruptive.

‍

Patrick Moorhead: 

They do. I mean, I downloaded it, I downloaded coworker for the first time yesterday and and was playing around with it. And what I found with opus 4.6 is the complex operations. It does really well, doesn't do research as well as, um, ChatGPT in my opinion. Uh, but we will we will see. Hey, let's move to the next topic here.

‍

More anthropic. Right. Uh, kind of related to the flip topic that we're going to bring up here, but, uh, anthropic just, uh, put in $20 million, uh, to back us candidates to support the regulation of of of of AI. And, you know, this, uh, strikes me at the outset as a bunch of doomerism, um, and I initially had.

‍

Yeah, but anyways, this is your topic. I mean, what's what's going on here? Like, why why is anthropic doing this?

‍

Daniel Newman: 

Well, I think Dario, the CEO anthropic has been one of the most vocal whistleblowers about how fast AI is proliferating. I think there is a sort of.

‍

There's the arms race with like the US and China and the desire to sort of win on all AI fronts. And then there's the fact, and I think you and I have had some, some you know, we don't talk about our private conversations, but like when you finally come to the recognition that AI isn't a bubble necessarily and that AI has utility, and then you start to see all the things that it can do and when it's applied, how effective it can be and how many.

‍

Let's just say for now, just task based roles it could potentially replace. And then you start to see as it becomes more iterative and collaborative, how much knowledge work it can do. Um, because once you by the way, most people use AI wrong. Sidenote they kind of one shot it, you know, thinking like I ask a question, get an answer and didn't like it.

‍

Oh, AI doesn't work. I mean, you you know, when people are having those conversations telling you like AI hallucinates and doesn't work. Do you ever ask them like one of the most recent time they used it was and like how they use it because I think I think there's a big gap. But anyways, the point is, is when you learn how to use this stuff and you apply it and you see what people are doing, um, you start to go, what happens to the economy?

‍

What happens to jobs like. And I know these the, the, I think the script that gets sent out to all technology CEOs is to say every technology industrial revolution has come with more jobs, not less, right. That is kind of the the script that all of the CEOs get. And I think historically speaking, it's been accurate.

‍

But the difference is, is I've never seen tech diffusion happen in such a short period of time, meaning that the diffusion of this, this, this AI revolution, we're seeing change that used to take ten years or 20 years during an industrial revolution, or is taking place in two, one every week. You're seeing breakthroughs.

‍

You're seeing technologies being delivered that can disrupt data services industries. They disrupt legal services industries, disrupting software services industry, like where you're literally looking at stuff and going, oh my gosh. And, um, it's it's, you know, I mean, you're seeing robots cleaning toilets now and and like, yeah, maybe that's not a scale thing at yet at some point, but where, where is work.

‍

So my point is like I think there has to be more of that conversation. Like I'm not exactly sure what benefit we get from handing a lot of regulation to the states. Um, I know why we do that sometimes. I don't know, in this case, if that works well for the sake of our kind of US-China diffusion race. But I do know for sure that we, we don't have an answer for if Dario is right, an AI is as good as he says it is.

‍

And by the way, I didn't even mention coding. If it's as good as he says it is, you know, and it can do, uh, 50, 60% of knowledge workers jobs in the next 12, 18 months.

‍

Then what? How are we going to run the. How are we going to run the economy?

‍

Patrick Moorhead:

Yeah. None of this makes a lot of sense to me. I mean, if if Dario is so concerned, then he should just slow down. He should slow roll. Uh, his his AI and limit it to do gooders and and stuff like that. I mean, it's just it's just weird to me. It kind of reminds me of just, I don't know, some woke, uh, agenda, um, and out of the other side of your mouth, I will say that, you know, if if this goes as quickly as it could, uh, society is not prepared for, uh, what's going to happen with the sharp, uh, job loss like we've never seen in the history of the world? Um, and security issues.

‍

Daniel Newman:

That small price for him to pay Pat to just to be able to say I told you so, like 20 million to be able to say I told you so.

‍

Patrick Moorhead:

Yeah. I'm part of the problem. I'm part of the solution. Right? And maybe it keeps I don't know if you saw, uh, one of his chief safety people quit and then, you know, decided he was going to write poems. Did you?

‍

Daniel Newman:

I saw there was, like, a lot of that this week, though. I think you and I get the same speed, but, like, all these people leaving, I basically decided, you know, that it's there's a there was like a very big algorithmic push on humanism this week was like a, you know, I had to stay off line. It's depressing.

‍

Patrick Moorhead: 

Listen, I that was the other day on my birthday that that I tried to, I realized, is how addicted I am to the phone and just how terrible it is. And I need a wake up call. And there's whole like, looking down versus looking up. Uh, I really am trying to look, uh, more up. And that doesn't mean looking at a PC screen or a TV.

‍

Daniel Newman: 

I mean, exist in the physical world.

‍

Patrick Moorhead:

Yeah. Like, how long can I go without looking at my phone or looking at my watch? Uh, it's hard, and, uh, I got sucked in, man. It's like hook, line and sinker. It's. It's pretty awful. Hey, let's move to the next, uh, topic here, Amazon. Uh, rumors of a content, uh, marketplace to train AI data. I mean, listen, what distinguishes a lot of different outcomes?

‍

All outcomes is personalization and content. And there has been a, a, I would say not a robbing, uh, but a pillaging of content out there, uh, that people typically get paid for or at least get some, uh, advertising on. And the rumor is that, uh, Amazon is going to put together, uh, kind of a content platform for content creators to make money.

‍

Um, and, you know, you've already seen deals with OpenAI. Uh, I don't know what Anthropic has done. Uh, you you've seen a lot of people, you know, Sue. perplexity as an example as as well. And this could be an intermediary for everybody else. If you're not one of the two top level labs out there. I do think this is important.

‍

And this kind of gets into this long game of, um, you know, it's funny, we like the results that we get out of, um, uh, out of out of these agents, but a lot, you know, 99% of the content and thinking is coming to them. I know there's a thought that wants to get to AGI and its thinking on its own that will that will, you know, lessen the requirement to to go there.

‍

But I can't even imagine a world where all we have is AI slop, um, uh, out there. So yeah, I we'll see where this goes.

‍

Daniel Newman:

Yeah, I think, I think we're gonna end up with mostly AI slop.

‍

Patrick Moorhead:

Doesn’t AI  just get dumb at that point. Or.

‍

Daniel Newman: 

I mean, that's that's the question. Like for.

‍

Patrick Moorhead: 

Like principle.

‍

Daniel Newman:  

It's like the recursive approach to make it continue to get better. When you no longer are feeding it kind of high value. But I guess the point is, is how collaborative and iterative is able to be on the existing high value to synthetically create more high value.

‍

Um, you know, again, I think a lot of people get stuck in this kind of mindset of what we see today as the best that will ever be, and they really have a hard time envisioning what is possible. I think this is why I always say, like, you know, optimists just outperform pessimists in every part of life all the time.

‍

So I really smart people, I think sometimes they get stuck in their own heads and then dummies like me figure it out because I'm just more optimistic. Um, but yeah, I, you know, I think on that topic, you kind of said it all so we can keep moving.

‍

Patrick Moorhead:

Yeah, that sounds good. Let's move to the next one. Uh, SaaS apocalypse here. Uh, actually, no, let's let's hit Spotify first. Um, Spotify says they haven't, uh. What's that?

‍

Daniel Newman:

Yeah. No, this is a this is a quick one, man.

I mean, this is just proof of everything we've been talking about on the show so far. This Spotify.

‍

Patrick Moorhead:

I ask you this, though, what is writing a single line of code? Does that mean that dude, I, I have been doing a lot with generative AI and and vibe coding and my son obviously does. And you know, you can't just go in there like you're the data feeds don't even work. You got to go in and modify those. Does that mean you're not modifying code?

‍

Daniel Newman:

Yeah. You know the weren't that prescriptive about it.

‍

What I can tell you from talking to the developers on our platform is that the tooling is really where all the work is taking place right now. So the code itself is almost. Most real good developers will tell you there's very little code being written, but the tooling that connects all the everything together and makes everything work.

‍

There's still real work to be done there. And so I'm I'm guessing that what they're saying is there is probably more of the former than the latter, meaning they are using Codex or using cloud code to do the generation of, of code. But I would bet that their full stack engineering team is still spending a lot of time to make sure everything works.

‍

I just because, you know, if it was this good, Pat, I can tell you this is you know, I've spent a lot of time and money building a platform and a tool, and I've seen a lot of progress, but it doesn't move that fast. It's like you can show me something in a, in a, in a dev environment, in a staging environment very quickly.

‍

Like this is how it could work if there was no database and there was no outside products, you know, outside tools to connect to If you just want to see a visual of something, but like when you actually need it to connect to six, seven, eight different concurrent systems and make sure everything works. And by the way, it looks elegant as it works.

‍

There's still real work to be done. So. But I do think the point that kind of because remember a year ago the debate was that AI wouldn't build a code. It wasn't going to be good enough to ever replace, you know, your top developers. I do think most developers now have come to the conclusion that that's not the case.

‍

Patrick Moorhead:

Hey, I never thought that would be the case, particularly because, um, heck, coding is more structured than language, right? In a sense, it should be easier than doing an LLM, and I'm sure I'll be challenged on that. And I would love that. But, um, each each language, um, has structure, uh, has different specific words.

‍

It, it it it needs, it needs to use so.

‍

Daniel Newman: 

Well it sounds like though, it's like it's just there won't really be machine code, right? I mean, that's what I'm hearing. And reading a lot is like the translation layer is kind of going away.

‍

Patrick Moorhead:

Yeah. It's kind of like going from machine language to an IDE or the interim Pascal and stuff like that. Um, yeah. I think the question is kind of what what are our programmers, uh, become? Hey, related to this is this SaaS apocalypse, right? We saw the market. This is in a bulls and bear segment. But for context, um, I think basically SaaS is down, what, 25%?

‍

And a lot of questions are coming in. Uh, and, you know, these SaaS companies are are changing, right? Even Salesforce has gone from, you know, they've got three different ways that you can uh, by their by their agents. Right? You can do it by outcome. You can buy them in bulk. And the third way is kind of a blend, uh, between this.

‍

So, um, you know, the big question is, um, you know, what exactly is going to happen to these, these business models? And I am a believer that I don't think SaaS is dead, but the people aren't using SaaS. Uh, hardcore are going to be dead. And then if you're an enterprise, you have to have to understand how they operate.

‍

And we've seen a lot of companies demonstrate this already where, um, uh, if your core competency is manufacturing, you're probably going to be putting a lot of your own oomph into your ERP, um, and your SCM, and you might, uh, kind of outsource your, your sales and marketing agents to, to Salesforce. And, you know, if you don't do a lot of manufacturing or distribution And you're mostly, you know, your value add comes, you know, at the front office, you know, you're you're probably going to put most of your in-house AI effort, um, uh, on sales and marketing as opposed to and you might, you know, pay sap for, uh, for Jewel.

‍

Daniel Newman: 

Yeah. I mean, this is one of those new where nuance matters. It started with Satya, and his whole enterprise software is just Crud databases with logic on top. He kind of sparked that well before the kind of sell off happened. But, you know, then you started to see what you could build with Claude. You saw, you know, a CNBC anchor build, a Monday.com replica, uh, you know, in her home office. 

‍

Patrick Moorhead: 

Didn't work. 

‍

Daniel Newman:

What's that?

‍

Patrick Moorehead:

Didn't work, by the way.

‍

Daniel Newman: 

No, it doesn't fully work. But the point is, is just being able to kind of see what can be done in minutes, and then you start to say, okay, I think a couple of the big problems I see with the apocalypse are one is we are grouping all things to be the same when they're not. We're sort of we're sort of not appropriately demarcating between companies that have really big moats and deep, uh, uh, critical data, rules, governance, rails, compliance that sits in their software versus like features that became public companies that probably like a SaaS company that did nothing but rev ops, or a SaaS company that does nothing but a project workflow, like just like one little thing. And it became like a big business.

‍

And there's a difference with an SAP or an oracle that has every bit of a company's proprietary enterprise resource data, managing their supply chains. It's managing their transactions. It's managing every like customer interaction for years and years. And by the way, doing it all in a way that's governed across hundreds of countries, um, compliant across different data, regulations and rules updated every minute to the minute, every day to make sure that this stuff is, you know, is able to to meet the requirements of business.

‍

And then, by the way, they're taking on the responsibility of the engineering that needs to go on. So a business can just keep running their business. That's a meaningful amount of work that I don't think most people can vibe code. Um, but there are like I said, that is like there are companies that are less differentiated, they have less of a moat there.

‍

And I do think there will be certain tools. So what I think is going to happen is one, you're going to get consolidation. I think the big platforms will consolidate some of those kind of features into their bigger platforms to give more diversification of what they're capable of, and have it sit on top of more critical and proprietary data.

‍

And then two, I think there'll be more collaboration. I think that these these companies are going to work with AI labs to make basically ad because because this is the one underwriting that I think most of the SaaS industry just never got right. And that's making customers love using your software. I don't think most people love using their enterprise software.

‍

Patrick Moorhead:

I think pissed off, they're pissed off to you that they got overcharged and they got locked in and prices went up 20% without adding adding any new features.

‍

Daniel Newman:

It became like an insurance industry.

‍

Patrick Moorhead:

Yeah man, I hear it. My CIO roundtable is like every single time. And I think their aspirations to literally get rid of all of SaaS. Uh, is is bigger than the, the, the reality, but just the vigor, uh, is there as well. And the whole C-suite and the board are, uh, wanted out as well.

‍

Daniel Newman: 

Yeah. So, I mean, what people want is, is the ChatGPT experience. They want to have an intimate relationship with their AI. I know, sounds weird. I mean, it's why you can't get off your device, but they wanted that. The ability to interact. They want to ask you questions. You know, I want to know, like, hey, where are the.

‍

Where's the best place to spend my time to help grow my business? Yes. The way we talk, you know, You know Benioff when he kind of talked about having AI in the boardroom like that actually was very salient. Like that was prescient of him to say yeah, the problem was is it has to actually be good. It has to be good.

‍

And so, you know, where we're getting to now is if the semantic layer has access to the right data, and then there's the right amount of iteration on top of the AI. These things can be really, really great partners. And so I think, um, I think we're directionally, uh, I think it's probably a little oversold, but I think the rerating in that industry is real, and I think it's permanent.

‍

You lost your background. You're still there. I think we’re Having technical difficulties.

‍

Patrick Moorhead:

Stimulated argument. Okay. Actually, let me go. Okay. All right, Daniel, let's get into the flip here. Do our simulated argument that I fortunately win every single week. Let's dive in.

‍

Oh that's right. Hang on a second. Sorry.

‍

Daniel Newman:

You gotta start the topic.

‍

Patrick Moorhead:

That's right. Well, today we're going to react to that essay from Matt Schumer, I think got 75 million views on X. That was just talking about the rapid change that AI is going to bring. Is going to be bigger than Covid 19 and fundamentally up on up in most jobs. Wherever you have a display in front of you within the next few years. So let's flip the coin. Who's going to agree with that and who's going to disagree with that?

‍

Let's go.

‍

All right Daniel, why don't you argue that pretty much every job is going to be upended in a few years. And a few typically means three. So jump in. Good luck.

‍

Daniel Newman:

Well thank you. First of all, we need to get the fat faced beard off my coin, make it gorgeous and jacked. Just like his answer is going to be. Um, here's the thing is, people and I said this, even on the show today is most people tend to struggle to see the future or even to recognize that change can continue to happen faster.

‍

We get very stuck in our mind where we are today, but fortunately with AI, the speed of disruption is very front and center. You know, whether it was three weeks ago, we got coworker two weeks ago we got clogged. Bot bot um, you know, we got Codex 5.3. You know, we talked on the show about companies that have engineers that no longer write code.

‍

You've got, you know, alarm bells being sounded by CEOs of companies that benefit from this proliferation and growth, that are basically almost saying halt, slow down, put the brakes on. And this is because the advancements in these technologies is very real, and it's happening just so quickly. And so you start to look at, you know, whether it's been the disruption of software, which we talked about today, whether it was the decision that, you know, lawyers would no longer be needed.

‍

We've got theories that, uh, that certain models are already diagnosing better than doctors. We have different ones saying that, you know, key

‍

AI tools are enabling AI to outperform the best traders in the stock market. Like every single job that you think of, you know, in the beginning it was, oh, it'll just be the monotonous work. It'll be the routine, the repetitive, the stuff that RPA and automation did, you know. But now you're looking at and saying, you know, do I need to have a big marketing team, or can my entire comms and PR team be built by a bunch of agents and bots that that create press releases and social content and write marketing copy?

‍

What about a website that I need generated? Oh, we don't need any, you know. And this, like I said, is just the early days. So, you know, Covid in my opinion, was transformative. And I think the analogy was less about COVID's transformation of work, and it was more about the fact that people couldn't see what was coming until it hit.

‍

And then it hit incredibly hard. And I think that's kind of what's happening here is, you know, remember, almost 3 or 3 months or less ago, the only thing that that the media would talk about is AI is a bubble. They're overspending. There's no use. There's no utility. There was that there was that example that went around of what happened to the AI over three years, and it was Will Smith eating spaghetti, and it showed Will Smith and the progression of Will Smith eating spaghetti in 23 and 24 and 25, and now in 26.

‍

And it went from a absolutely bizarre looking video that would be created with AI to this almost perfect cinematic experience that could be created with nothing more than language. So you look at movie producers, directors and content creators all at risk of jobs right now. So what I think is people, there's a phrase that says it's slow at first and all at once.

‍

Well, right now we had the slow at first, which actually wasn't all that slow. And then the all at once moment seems to be coming very quickly, and I think many are sounding the alarm, as was the person in this essay. Um, Right now, we have no plan in society to deal with the fact, when technology does enable so many of these jobs to be displaced.

‍

And I think what ends up happening is it happens very quickly. The the job creation starts to fall at a much more precipitous pace. We're going to get caught off guard in government and leadership. We will not have the ability to figure out how to, uh, you know, provide income and stability to all the people that used to depend on these jobs that AI can now replace.

‍

Um, and so while we will get a level of productivity boom, it will be a boom that will not require nearly as many humans to execute it. And then once that happens, um, we're going to be completely caught off guard. I think the essay was on the right track. I think we're underestimating the speed of disruption, and I think we are underprepared for it.

‍

And I think this is a really important topic that we need to start covering more regularly, because when we talk about the productivity, we talk about the efficiency, we talk about the profitability of AI. We need to talk about what is the risks and who, Who and what will be left behind. If this works as well as it appears that it will.

‍

Patrick Moorhead:

Yeah. So it's going to be an easy one. I'm just going to do some quick quotes here. It says most, if not all white collar tasks to be automated within by AI within 12 to 18 months, and that AI can impact 50% of entry level white collar jobs and 1 to 5 years with unemployment potentially jumping materially. Uh, what this thesis doesn't fully incorporate, which I think you and I talked about on the show before, is just how hard it is to integrate this into a current business that has has anything I mean, even if, you know, you have uh, net suite is your is your back end and Google Google is your, your front end.

‍

The ability to do first of all, who is going to go in there and and do these things? Um, you have uh, current IT. Who's doing their their jobs, keeping the lights on. Who is going to actually go in? Who are the incremental resources to go in and do that? And I know the boomers are going to say, you know, have have mult book.

‍

Sorry, have Claude Bot go in there. But that's just not the reality. And then when I look at all of the regulated industries and in health care, in finance, in government and transportation and, and utilities, I do not see that fast rolling at all. And I think there's a big difference between capabilities and deployment.

‍

I think the capabilities if if you're starting a new business, are are absolutely, absolutely there. And I think that's a start ups that that completely start with AI will be like the companies that started with the web like Salesforce, right where they came in. And they just kind of. They took all the stuff we were doing on paper and spreadsheets and, and then they came in and did that.

‍

We there was no concept of an ERP that that was used across multiple companies until SAP SA, SAP came in and then e-commerce hit. And um, um, so those new startups have the ability, but I don't think that impacts the regulated industries because the regulated industries haven't really seen a complete turnover.

‍

The only exception could be trading. Used to have to go to a stockbroker to put a trade in. And then and then E-Trade hit. So, uh, yeah, I mean, I think that's the that's the confusion here of the deployment versus the, the capabilities, uh, out here. And I also don't want to, um, I also don't want to dismiss, um, the intuition that humans have that, uh, AI doesn't have because it hasn't sat in that conference room for 35 years to actually get a, um, get a true assessment of it. And there's not a there's no metrics per se, uh, on trust that, that you have with somebody. And until machines are just buying machines and it's a, uh, a right brain, uh, type of activity.

‍

Um, I just don't see this hitting, uh, as, as quickly as it is. I mean, let's say you do believe in this. If you do believe. Uh, in in the essay, then fundamentally, uh, you believe. Yeah. Let's let's let's peel, peel the onion, uh, back here and, you know, it's so funny, like, you put out a what you think is a banger, a banger post, and it ends up like nobody reads it.

‍

Um, I thought what I, I put out what I thought was a, uh, a pretty good assessment that if you believed in this, here's what would happen. Right. Uh, good thing is. Right. Diseases are cured. We've never seen. I saw that over in the UAE big time, whether it's genotypes or phenotypes. But, you know, there will be massive job loss, right?

‍

And what that means is massive consumer spending cuts outside of housing, let's say transportation and food. And then any business outside of those that serve, uh, housing, transportation and food. Uh, they have to do a bunch of layoffs. So consumer advertising, electronics, home goods, entertainment, travel.

‍

People are going to delay medical and dental appointments. Uh, you're going to see more layoffs, and there's going to be a credit crash due to all the defaults, uh, on the mortgages, credit cards and loans. Uh, and that's interest rates will skyrocket. And that means that spending will decrease even more, and then businesses won't have the money to spend an AI as they conserve cash and lay off more people.

‍

Uh, and then AI companies are then, uh, impacted because the end user, uh, spent from even businesses plummet. And then you get to that, who's going to pay for for everything? Um, stock market crashes like we've never seen before. We've got civil unrest, uh, preppers and complete calamity and disasters.

‍

So if you believe in Schumer's essay, then you believe that essentially we're going to have civil unrest and a complete catastrophic decline in this stock market. Have a nice day, everybody.

‍

Daniel Newman: 

Are you prepping?

‍

Patrick Moorhead:

I am, I am I am officially becoming a prepper. Food. Guns, water. I mean.

‍

Daniel Newman:

 So. What does that mean?

‍

Patrick Moorhead: 

Well, there's hedging, you know, as a as a father and a husband. Right. I feel a duty to protect the family. And, you know, if, if something might happen, I mean, why do we think all these rich people have all of these hideouts in New Zealand with security guards and nuke proof, you know, what do they know that what do they know that we don't?

‍

Daniel Newman: 

Or like you said, they're hedging.

‍

Patrick Moorhead:

Yes.

‍

Daniel Newman: 

I mean, I don't know how long you want to hide in a bunker. You know, the event that the the the public goes cold on you because you're an AI survivor? Um, but I don't know, I, I don't think it's either. I think it's somewhere in the middle. But I do think that we are well under underprepared for a potential tidal wave in terms of knowledge work.

‍

Patrick Moorhead: 

Yeah, I think in the long term, uh, we'll, we'll we'll probably probably figure this out. But in the short term, it's going to be spiky. We've been able to digest all the change because it came slow agriculture to industrial electrification, uh, computerization, um, web and all that stuff. All right, baby, let's jump.

‍

We are in the homestretch here. Let's dive into bulls and bears. Let's talk about the market.

‍

All right, everybody knows bonds. Bonds are fun. Uh, they operate the opposite of stocks, typically in a normal market. And typically they're 30 years. They're 30 years, they're seven years. There's five years. But now we have Alphabet's century bond $31 billion global debt raise. What is going on?

‍

Why do we need a 100 year tranche?

‍

Daniel Newman: 

I guess why not? If you can offer a reasonable rate to an investor on a long term bond, it's going to underwrite a company that people, I think believe will be around for a long time. It's a it's a, you know, it's a low risk,

‍

uh, yield. And it's, you know, attached to something that people know needs to be invested in. I don't know, I think it's kind of a smart play. It's a way to raise money and way to raise it relatively cheaply. Um, but I don't know, Pat. I mean, I think right now new vehicles are coming out, um, and they're going to find ways to raise capital and they're going to do as much as they can to try to, you know, not harm their balance sheets too much and set themselves up for, you know, the long term to make sure they can get the cash they need to continue the build out.

‍

So I wouldn't be surprised to see more products like this hit the market from other other big seven names.

‍

Patrick Moorhead: 

I mean, when you do a 100 year float, which by the way, the demand was ten x the offer. Okay. That gives you the idea of the demand for that. You have to bet that company is going to be around in 100 years. Or am I just taking this way too literally? Daniel. 

‍

Daniel Newman: 

I think that's that. There is a underlying assumption, but I think most people won't be alive for 100 years. So I think there's, you know, sort of a, sort of a, a two sided bet there. I don't know that everybody's sitting there thinking of it because I'm sure they'll trade the bonds. Right. It's like just because they buy them doesn't mean they cannot exit them. But the product has the has a lifespan of that long. And yeah, I mean you would have the the bond would have to underwrite the asset for at least the period of which the bond exists.

‍

Yeah.

‍

Patrick Moorhead: 

Yeah. It's crazy. All right, baby, let's, um.

‍

Daniel Newman: 

You right. You investing?

‍

Patrick Moorhead: 

Uh, well, I mean, I'm going to be alive and tell them a hundred, so, you know, I wouldn't get that note back. My grandkids would get that note back, probably. So. No, thanks. All right, let's move on to, uh, meat and potatoes here. And that is earnings. Uh, Cisco came out essentially had a, uh, had a double beat and, uh, stock went down, I don't know, 12% initially, I think 10% the day after, really based on, uh, margin pressure.

‍

I mean, it wasn't a huge amount of margin pressure, but it was enough to to get people, um, riled up that I think really kind of overshadowed the good things they were doing. I mean, a 21% increase in, in networking, uh, Primarily AI driven switch demand and $2.1 billion of orders from hyperscalers, which again, I think about a year ago, that was a surprise to everybody and anybody that they participated in the hyper in the hyper scalar market.

‍

It's pretty well known that they also participate in what I call the the neo clouds. I mean, when I was over in the UAE, there was a lot of discussion about what Cisco was was doing there. And the fact is that if you're a neo cloud, you need help. But there are a lot of chips that that Cisco delivers into the, um, into the hyperscalers, uh, plus some, some hardcore, uh, switching in a particularly AI AI fabric in there as it's competing with Broadcom and and Nvidia.

‍

Daniel Newman: 

Yeah. I mean, look, this is one where I think a lot of things are falling. It's hard to say. Was it specific to a Cisco? I mean, it looks like they guided, uh, right to the midpoint on their, their EPs, and they guided above for revenue. Do people want to see more growth from AI? I mean, I don't know. I guess, you know, I think this is somewhat just a this is somewhat of a technical, uh, response because a lot of companies did well beat and still got sold off. This is a bit more of the market conditions than anything Cisco is doing. Cisco's continue to improve. Uh, it made a big pivot over the last year.

‍

It's playing in the right spaces. It's got the AI narrative is definitely coming up. The security demand that AI creates is a tailwind for the company. Um, the networking need for more bandwidth silicon, it's developing. So it's in all the right spaces. Um, I think the gyrations or the gyrations, as I like to always say, markets will market, but you know, you got to keep watching the business.

‍

And I think it's been trading up and around an all time high. So it's also not a name that's, you know, beaten down by any means. But when you're when you start to get towards those all time highs, the pressures to keep growing and outperforming become continuously higher. So good on I'm not nothing there, nothing there that worries me.

‍

Patrick Moorhead:

Yeah I think it's it's it's everything you said. Plus the percentage of revenue that goes to the hyperscalers. If 50% of their revenue came from hyperscalers, they'd be screaming, right. Just like just like Broadcom. Just like, um, um, uh, you know all the silicon photonics companies. So hey let's jump into uh let's jump in in Applied Materials they had a triple beat.

‍

Daniel Newman:

What is that yawn?

‍

Patrick Moorhead:

Uh, let's go into a triple beat Dan what this is.

‍

Daniel Newman:

Yeah. I mean. It was a triple beat. It wasn't a massive beat. But this is all about the future. I mean, this is a they basically just came out and said, like, you know, the demand is and we got this tip off from TSMC, but then the demand for semi cap is just going to explode. I mentioned earlier that trillion dollar number.

‍

I mean literally this is the first kind of CEO that's come out and said the tipping point has arrived. We remember we were thinking a trillion might take till 29 or 30.

‍

Patrick Moorhead: 

Yes.

‍

Daniel Newman:

I mean, that's a massive pull forward. And there's only way this happens is an explosion in semi cap because we're going to build out foundry. We're gonna have to build out the fabs. We're going to need all the equipment. And there's a handful of beneficiaries. And that is obviously one of the largest. So the numbers are good.

‍

The guide you know we're good. The guy was was good. But this kind of uh upwards growth. And of course I think there's a rotation to hardware because no matter what happens to software, no matter what happens to these other industries, the hardware to build this out is stable. So this is the most stable part of the market.

‍

Um, and I think people are rotating that way. We've seen it with some of the memory players. I mean, these are not companies that have historically been super high growth or interesting, but people want to be in these companies because they are now becoming the predictable picks and shovels of the AI revolution.

‍

So, I mean, it's hard not to like the prospects of any company that's meaningfully supplying equipment to foundries during this transformation.

‍

Patrick Moorhead;

Yeah. I mean, as investors get a lot smarter about the the value chain, you know, I, I would posit to say that most retail investors had no idea who applied materials even was three, 3 or 4 years ago. And you know what? One of the things I really like about applied is that it's not just about, you know, doing a node shrink like ASML.

‍

It's about it's about everything else. It's about Dram, it's about HBM, it's about packaging. It's about, you know. You know, all of the cool features that Intel talks about with 18 A and like backside power. Uh, 18 A and 14 A is essentially delivered by by Applied Materials. So yeah they're definitely a a future forward uh story here.

‍

But they're delivering today. You know, they're they're essentially their inventory to booking is is like one right. They, they they're carrying. No they're carrying no inventory. Uh, they just got slapped with, uh, a fine, uh, for shipping stuff to, uh, China subsidiaries through Korea. I read I actually read the entire complaint, uh, and essentially came down to the definition of transformation, uh, which is, um, applied thought it was doing enough transformation and changing, uh, to the equipment before it shipped.

‍

Shipped it to China. But apparently it it wasn't. It was the biggest fine levied, I think, in the history of, of these types of, of fines. I don't know if you had seen that, Daniel.

‍

Daniel Newman:

I did I didn't read the the details though. So thanks for sharing that.

‍

Patrick Newman:

Yeah. Yeah. Sometimes details matter. Sometimes they just don't.

‍

Daniel Newman:

I'll ask Claude to give me the breakdown later.

‍

Patrick Moorhead:

Yeah, just just add, uh, by the way, the whole mold book thing, um, this one guy got on X and said everybody was fawning over his post of, like, a new religion, and and it was a bunch of human programmers who created that just to just to get people's reaction.

‍

Daniel Newman:

Well, I think I think what they're doing is they're prompting it, and it's just like, you know, when you do read about, like, AI's blackmailing their their humans. Yeah. When they are being told, they're like, you know, there is a certain amount of their being trained by humans, and humans are survival.

‍

We're survivors. Right. So the inputs do yield the outputs. And even if it wasn't totally pushed by humans, I mean it is driven by the prompt. Yeah, right. It's just that prompt is a spiral anyway. All right. What else we got?

‍

Patrick Moorhead: 

We got lattice. A really simple story here. Uh, just keep an eyeball on the time here. Simple story as goes AI CapEx. Uh, goes, uh, lattice. Uh, they had a beat, a meet and a beat and, uh, stock, you know, went nuts. I think it was up 9%. Um, after hours. And, you know, they're leaning into AI servers where they should and they're investing in the future, which is robotics and physical AI.

‍

It's as simple as that.

‍

Daniel Newman:

And data center, they of course have a strong data center play. And I mean, they are the collaborative chip in many systems. And you know, they finally are going to tell. The nice thing happened was just that they had they'd been at one of the semi companies that really didn't get any pole over the last 2 or 3 years.

‍

They were sitting on a lot of, you know, sitting and a lot of inventory. Uh, they'd had some margin pressures. And you actually look at the kind of revenue curve. They're definitely the sine wave is back on the on the side for the company.

‍

Patrick Moorhead:

Hey, let's do a quickie. I know you know them pretty well. Tara Wolfe. Daniel, uh. What happened?

‍

Daniel Newman:

Yeah. I mean, you know, they there were some really big ratings increases. These are what we call the AGP. Pat. This is one of them. Tara Wolfe, cipher mining. Uh, yeah. These are the companies that are basically were they were mostly crypto miners and they turned into energy capacity plays for data center companies.

‍

And they partnered, you know, with the big hyperscalers to offer additional capacity. Um, I own Wolfe, so I say that out loud. I bought it, um, and, I don't know, ten bucks. And basically the reason I did was because right now, I do think that while we talk a lot about compute, we talk about memory, we talk about all these constraints.

‍

Like, I still don't see what we're doing to fix the energy constraints. So these companies have access to power. Um, they have facilities. They're building out the data centers, and then they're basically plugging in these really, uh, long term leases with hyperscalers. Um, and, you know, they're not bringing in most cases, any of the actual data center build out the AI expertise.

‍

They're just bringing the facility power, expertise. And they've become really, uh, you know, interesting strategic partners to this build out of kind of fast capacity to go in addition to this sort of more ongoing builds that each of these big clouds, neo clouds and hyperscalers are trying to execute.

‍

Patrick Moorhead:

Yeah. The company that I visited in the UAE last week, uh, G40 two, is a big, uh, as a big investor there, and it's a like a 70 it's a 70 megawatt implementation up there in upstate New York. Interestingly enough, it has Dell Power Edge servers. Liquid cooling, uh, classic, uh, neo cloud that, uh, Dell has been successful with.

‍

So, yeah, it's it's funny the things you, you can uncover when you, uh, you know, dig under the cover a little bit.

‍

Daniel Newman:

It's all one. Big connected thing, dude.

‍

Patrick Moorhead:

Yeah, it really is. Daniel, we, uh. We made it. Um, I want to thank everybody for tuning in. If you have complaints at Daniel Newman Comm, if you want to fawn over the show at Patrick Moorhead on X, we will be there. Thanks for tuning in, everybody. And we will try to be more consistent in the future. Thanks and take care.

‍