저자: DavidCarbutt · 외국인
요약 보기
AI 버블 논쟁. 기업들의 AI 지출은 현금흐름 기반이라 1990년대 닷컴과 다르다고 주장. Salesforce·ServiceNow 등 SaaS는 AI 래퍼로 플랫폼(팔란티어)에 점유율을 잃고 있으며 시장이 이를 냉정하게 벌주고 있다고 평가. 클라우드 인프라 지출 급증, 엔터프라이즈 AI 도입 확대(의료·금융) 근거로 제시. 팔란티어를 플랫폼 승자로 긍정 평가. 영상 내 자산별 명시적 가격 방향 예측은 거의 없음.
다듬은 글
Arch Invests, Kathy Wood, and Brett Winton just had one of their most important conversations I've seen in a while. They made some pretty big predictions and in my opinion, this is a mustwatch. So, some skeptics would say, "All right, let's set aside the pure play, right? You've made the case for anthropic.
You've made the case for Open AI. You've made the case, right, for some of the well-known pure plays. Does it speak to a bubble when a CEO can simply go onto their earnings report and say, "We're putting AI into everything we do. I expect my stock price to go up. You know, it's very interesting uh to watch these CEOs uh doing exactly that.
And we're seeing it in the SAS space, especially Salesforce.com, Service Now. Salesforce.com was supposed to be one of the biggest beneficiaries. I remember when ChatG all their commercials still reference it on TV, Einstein, and now they have everything AI. Uh and yet their growth, topline growth has not accelerated.
it's sort of in that you know low double digit range and has not accelerated the market and the market's punishing those stocks. Same with Service Now even though its growth rate has been very strong much stronger than uh than Salesforce.com and we had done the research over the last few years showing that over the last five years uh the software as a service part of the tech stack or the applications part of the tech stack has lost share to the platform as a service part of the tech stack like Palunteer and it seems like that is still happening.
Uh and in fact what Palanteer is doing brilliantly is really just laying its software on top of legacy software in any enterprise saying don't worry you don't have to rip and replace and what they're doing over time is usurping the role that all of those SAS players were playing and you know as they build on top of the pal Palunteer platform as Palanteer builds uh you know the enterprise software on top of its platform.
So uh I think the market is quite discerning and and I I would say the other in terms of the hype cycle uh compared to the tech and telecom bubble. I mean face I mean meta platforms Amazon uh Alphabet uh Microsoft uh they are cash fortresses and so they have basically said we are upping our capital spending.
Y and they've got a shareholder base that is looking at where they're putting uh where the companies are putting their cash and you know there's a little bit of a turnover in shareholder base saying wait a minute I I I thought I was getting this kind of free cash flow and now you're telling me I'm not.
uh but nonetheless those stocks have done pretty well and their growth rates for the most part have accelerated uh and and they are being rewarded not all equally uh but they are being rewarded so uh one thing we didn't have I remember Amazon in the late 90s Jeff Bezos would come out and say we are going to lose more money and you should be really happy about it because that's the opportunity we see and the market bought it for a while and then in 2000 it didn't.
So, uh, that went way too far. Right now, if we're looking at the kind of cycle we've been in in the last five years, I'm going to say, uh, investors have been demanding cash cash flow and now they're quite discerning about, okay, you're you're spending on AI, but we're not seeing it in the revenue growth line.
This is a problem. and others are seeing it and they're being rewarded for it. So there's I think we're quite rational out there. I think also people under or don't remember or didn't live through how crazy the late 90s were. Like in in in 2000 you had more than an IPO per trading day. That would quadruple quadruple on the first day or quintuple.
I was there. I was there. So So you it's clear. I mean, I do think that at some point we will get there. At some point there will be froth. Uh, and it will be and the activity I I think the people misremember how frantic the activity level can get. And like Kathy alluded to earlier, right now people are so hungry for this to be a bubble because they feel uncomfortable.
You know, in fact, the big incumbent software providers would prefer it to be a bubble because they're not very well positioned to actually take on this capability. If you're trying to make sense of AI, I made a course that gives you a complete breakdown of the entire industry from frontier models to chips, data centers, and everything in between.
I built the course I wish I had when I started researching AI, so you can skip the confusion and get straight to the clarity. We'll be adding new modules over time as the AI world evolves, and the price will be increasing by two to 3x on Sunday. If you want the best deal, and lifetime access to all updates, now's the time to jump in.
The market is actually being pretty smart about this. Companies that just slap an AI label on their products without showing real revenue are getting crushed. Salesforce appears to be the perfect example of the AI wrapper problem. The AI tools are basically wrappers around third-party foundational models from anthropic and open AI.
So why is this happening? Let me explain. The traditional SAS playbook stopped working the moment AI became commoditized. Think about it. If you're just wrapping third party models, you have zero differentiation. Your competitor can access the exact same foundational model you're using and the data backs this up.
Salesforce need to prove their agentic and data cloud products produce measurable changes in win rates churns and margins. Many customers and partners have struggled to identify real gains despite rebrands and naming shifts. Even Beni off their CEO and founder acknowledged the division between rapid consumer AI chatbot adoption and slower enterprise adoption.
Service Now is winning because they're doing something completely different. Their AIcentric orchestration strategy is translating into significant and accelerating market success. They're positioning AI agents as first class workers with agentic workforce management orchestrating human and AI collaboration.
Service Now is becoming essential enterprise infrastructure by embedding decisionmaking software deep into operations. And then you have Palanteer taking a completely different approach. They're building an AI powered decision layer with software that captures context, adapts workflows, and that learns institutional behavior.
Companies use Palanteer to integrate, govern, and activate information at the enterprise scale. Palanteer overlays on top of legacy systems. You don't have to rip and replace anything. They gradually become the new control layer, while traditional SAS applications become more commoditized backends. So, here's the logic.
Applications serve specific functions. Platforms orchestrate decisions across the entire enterprise. In a world where foundational models are commoditized, the platforms that orchestrate these models create defensible value. The application that just wraps them doesn't. So Kathy is absolutely right that the market is being rational.
And here's where both Kathy and Brett nailed it. Brett is correct that we're not hitting late 1990s frenzy levels. And here's what's different from the late 1990s. Companies today are cash generating machines building the infrastructure that's going to power the next decade. So here's why Cathy's cash fortress argument is actually brilliant.
In the late 1990s, companies were building infrastructure on borrowed money with no cash flow to sustain it. When the market turned, they went bankrupt because they couldn't service their debt. Today is completely different. These companies are generating hundreds of billions of dollars in cash flow.
They're not taking on unsustainable debt. They're investing profits into infrastructure that they own outright and the late comparison actually proves Cathy's point. The fiber optic cables laid back then are still being used today. The data centers that survived became incredibly valuable. The infrastructure was real.
The problem was the companies building it went bankrupt before they could monetize it. But when you have Google's cash flow, Meta's cash flow, Microsoft's cloud revenue, you can afford to build through the entire adoption cycle. You're not forced to cut spending when sentiment shifts. You don't need to monetize it immediately.
You can afford to be 3 years early. You can afford to be 5 years early and you still survive. Spending on cloud infrastructure reached massive levels in the most recent quarter, increasing dramatically compared to a year ago. That's actually workloads being deployed. That's actual enterprises moving AI from pilots to production.
And the deployment patterns show this is real. The overwhelming majority of enterprises adopt hybrid architectures to balance cost efficiency and performance requirements. Most organizations needed to change or modernize their IT infrastructure to support generative AI applications. This is companies rebuilding their technology stacks because AI actually works.
AI saves workers meaningful time every single day. And professionals predict AI could save dramatically more time within a few years. When you multiply that across millions of knowledge workers, you're talking about massive productivity value. Healthcare is leading generative AI adoption with massive spending.
Financial services show the highest ROI. Cyber security initiatives are exceeding expectations with a significant portion delivering ROI above expectations. Mission critical applications where enterprises will pay premium prices. The top cloud providers will account for half of global data center capex within a few years.
Accelerated servers may account for more than half of total server revenue soon. That's not over capacity. That's market consolidation around winners who can actually deliver results. The argument Cathy's making is when you build infrastructure at scale and you have the balance sheet to hold it, you don't just have one revenue model.
You have multiple paths to monetization. The market is rewarding execution exactly the way Kathy described it. Companies delivering measurable ROI are getting funded. Companies showing productivity gains are scaling and companies with the infrastructures to support AI workloads are winning. ?
Comments 0