Palantir Hits $29 & Exclusive Interview With A Palantir Employee

First $29 touch since 2021

Welcome back to DailyPalantir! Palantir hit $29 for the first time since September 2021. We analyze a new research note from Dan Ives and then go into my interview with a Palantir employee. Let’s get into it!

Palantir Hits $29

Palantir was up 71% YTD and had the first $29 touch since September 2021. We closed at $28.74 yesterday after hovering over $29 during the day.

I could be wrong, but I feel the recent strength in Palantir $PLTR is institutional. Even if we head back to $27, we have seen incredible closes over key levels like $27.5 and $28 for the past 2 weeks. It just feels like Wall Street is buying Palantir before earnings, and if they can put up strong Q2 growth, then the street may be ready to jump on board and we don’t see the low 20s again just like we haven’t seen below $20 since Feb 5th (Q4 2023 Earnings).

I don’t think retail keeps bidding at these high levels — the only way that happens is when millions of NEW retail investors discover a stock and we haven’t had a major catalyst (usually tends to be a solid earnings) for them to discover it yet. I think MAYBE the big guys have discovered Palantir and they are getting in before Q2 (which would explain this new base around $27) and are willing to buy it at higher prices given they think AIP growth will showcase Palantir as one of the main AI SaaS names that has the potential to innovate and grow.

Dan Ives Note

Here’s a sentence from Dan Ive’s latest research note to his Wedbush clients that came out yesterday:

"Our myriad of field checks globally over the past month give us a high level of confidence that the AI revolution monetization has now begun to hit its next gear of growth as the baton has been handed from semis to the software phase with use cases exploding across the board," the analysts said.

So, I feel this run may be more institutional. I could be wrong, but given the strength of the price action, I do not feel retail is controlling this rally. If I am right, what Dan Ives is saying makes sense — if people like him are advising their clients to look to the software applications of AI over the hardware/semi plays, then that could explain why Palantir is getting more attention from the street. It will be up to them on Q2 earnings to justify their multiple and continue to expand, but the street may be getting ready to allocate capital near software applications of AI because they are beginning to see the monetization of AI in the enterprise.

Earnings is August 5th

I’ll be live on my channel covering earnings in August 5th. The street is expecting about $650M in revenues. If they don’t blow out expectations but can signal the monetization of AIP is beginning, this is why the street may be front running the stock before they see the growth. They know the revenue growth from monetizing AI is coming, they know Palantir is one of the key leaders in implementing AI in the enterprise, and they know software is where most of the use cases for AI are going to be relevant in.

Interview With Chad Wahlquist

Chad Wahlquist, Chief AIP Architect, came on to my show to discuss a recent tweet he put out. I’ll attach the tweet below and here is the video of this interview. Below is the written form:

Question: Okay, so Chad, you've been a friend of the channel for a long time. People know about you in this mysterious way, but you put out this tweet today that is making you less mysterious. This tweet not only is it doing well in the Twitter algorithm, but it's an incredible tweet, and it's about your kind of sales call with a company that's worth a hundred billion dollars that's thinking of Palantir. So, before I get into this tweet itself, who are you, what do you do at Palantir, and I guess in general, why do you like working at Palantir?

Answer: So, Chad W, yeah, architect is my official title, but I do whatever needs to get done. And so, like, I think that the biggest reason I came to Palantir was, you know, I was a prior customer. I brought Palantir into Tyson Foods. I led all the Tyson Foods organization there around IT, analytics, that kind of stuff, and I saw the power firsthand of the value that I could add with the tooling. And the tooling was focused on outcomes, not IT focused of "Hey, I can build a data lake." What am I going to do with that? How do I add value? And so, that was the original thesis I had. And then, when kind of the Gen AI boom happened, I was actually working on Amazon on some generative AI components, and when I saw how things were transforming and how my work was going, it became obvious to me that I needed ontology to actually really make the stuff work in a meaningful way in the enterprise space. And so, that thesis became apparent, and that's why I joined Palantir, to work on that, to see how I can bring Gen AI functionality into the enterprise.

Question: This is really exciting. I've gotten to know you, and you, I credit, as one of the reasons Arny and I actually understand ontology in depth because you sat with us for three hours and kind of forced us to understand what Palantir is actually offering here. And obviously, that ontology is a differentiating moat for Palantir that we all believe is probably one of the reasons clients get excited about them. So, in this tweet that we read about 30 minutes ago—I'm going to go through it again—you had an introductory call, basically a sales call, with a hundred billion dollar business, and their question was, "How are you different from the Snowflakes and the Databricks of the world?" So, I guess, Chad, can you answer that question for all of us that are probably wondering, why is Palantir different from some of those bigger companies? And in the context of selling Palantir to some of these clients, what are the things they're looking for to be answered to go with Palantir over a competitor?

Answer: Yeah, well, this is the reason I put this tweet out this morning. This is a theme I see—you know, it happened yesterday at a big potential customer—but this is a theme I hear across many IT practitioners. I was an IT practitioner for 20 years, you know, running a $50 billion organization's IT, and so I know the pain firsthand of what it looks like. I went through the transformation from legacy systems to Hadoop to cloud-based; I've gone through that my entire career. And so, I see their pain. I actually didn't even have time to really dive into the ontology in that call yesterday. I only had 30 minutes, and so I didn't even bring up the differentiation of the ontology to these people. I called it out; I kind of gave a high level of what it was. But really, the real focus I started to pick up on when I was reading the room was they have the cloud, they have all these supposed best-in-class tools from these cloud providers, but they can't get value out of them, right? They're spending no time. And that's where I started to hone in on things: you have to actually be focused not on the tools but the value, and the tools should be frictionless to help you add that value.

Question: Why aren't they getting value? Did they say why they can't make this operational for them?

Answer: Well, they just said, "Man, we're trying pieces everywhere. We're doing all the latest co-pilots and tools," and they named off, you know, 10 different AI companies and tools, and none of them work out. And they came back and said, like, you know, "We're trying these chatbots, but it's really just a natural language filter on a report I already have. It doesn't help me. It's not transformational." Like, yeah, it's kind of nice, it's a natural way to do it, but it doesn't add value. And those were the key words I started to hear: they've tried the co-pilots, they've tried all these things, but they're not actually getting value out of them.

Question: Okay, that's where I hone in on, like, why. And they like, "Okay, the co-pilot example here, but also the services. When you go to a cloud, everyone thinks you have all these services in one cloud, they all just work together." In reality, no. I mean, I worked at Amazon; I can tell you that the teams do not work together to make sure that product A talks to product B natively. And so that's a huge frustration for them, that just in the AI space alone with the co-pilots, but then just in cloud in general, kind of more in the Foundry space, they couldn't get anything done. So, okay, this is incredibly interesting. So, when they are asking for the answer, when they agree with you on the problem in the first 15 minutes of the call that they're fragmented, they can't get things done, they're using these chatbots, it doesn't mean anything. But then they've heard a quote from Shyam or Karp that says "Go beyond chat." What does that mean? What does it actually mean to go beyond chat, and how do you sell that as a vision to them?

Answer: So, like I said, chat is really focused on, like, how do I query my data. And so, I'm just having a natural language thing that's going to essentially write a SQL query. And that's actually what I see a lot of these tools do: I have a data warehouse that you can access through SQL, the chatbot essentially interprets the intent and writes a SQL statement against the data warehouse, returns an answer, put via BI, that's a report. And so, when we say "move beyond chat," it's how do I actually operationalize things? How do I use AI to automate workflows to access unstructured data that I couldn't access before? Maybe it's sitting on a SharePoint site, it's in images somewhere, we see a lot of that. And so, Gen AI can help us do that natively. So, for example, in the procurement use case, like that video I did, there's nowhere that I'm, like, chatting with the data. Gen AI is extracting information from RFPs, it's reasoning across them, it's looking for those across invoices and matching those up. That reasoning function is what we're focused on. Gen AI brings reasoning to things that normally had deterministic logic in humans. Now, the human-machine reasoning is what we focus on with LLMs.

Question: And that reasoning can't actually derive value in the context of the large language models unless it's connected to an ontology, correct?

Answer: Yeah, ontology is the digital representation of your business and your business process, how it actually operates. And so, the focus there is, if I have a digital twin of how my business operates, great, the LLM has no knowledge of your business. It has no knowledge of what I call the dark matter, the things that are on whiteboards and spreadsheets and hallway conversations. So, the ontology gives you that backbone to be able to capture those things and then do that in human language, human knowledge that was trained to train these LLMs. And so, basically, the idea here is the ontology is the representation of your business, the LLM is the machine reasoning. It can use tools to navigate across the ontology, navigate across the digital twin of your business, and then find answers and ground that in the truth and reference back, "Where did I get this answer? How did I use these tools to get it?" And then also take action. So, it's not just the data piece. We say data, logic, and action. So, it's cool, now I found an answer, what do I do about it? What's the optimal, what's the optimal three responses I should do? And then pick the most optimal one and actually take action. All that is facilitated through the ontology. Without that, you just have a bunch of data piles that the LLM tries to reason across and has no clue how they're structured or what it means or how your business actually operates or the stuff that's in the hallway conversations.

Question: My next question to that, and I think this is going to be a question that everyone wants an answer to, because I feel like everyone's been trying to figure out the answer, but we don't explicitly know, so I think someone like you could really tell us what the answer is. Why can't Snowflake do what you just said?

Answer: So, Snowflake, they've got their own LLM as well. They've trained it on business data, but it's not your business data. It doesn't have the logic about how you operate or anything like that. The ontology is something we've built over 20 years, hardened experience to make it scale, perform securely, like all these things have been tested, battle-tested with Gotham through Foundry and all the big work we did at BP and at Tyson and all these other places. That system is very unique. I don't see anyone else in the marketplace. Snowflake is focused on a data model, a dimensional data warehouse, a star schema. These things are just models that try to represent a little bit more business-friendly or user-friendly way of the system speak. And so, the thing is, everyone's focused on the wrong problem. They're focused on, "How do I sell my widget? I'm going to create a semantic model that a BI tool can sit on and it's going to tell you an answer of some metric." That's what they're focused on. No one has even focused on what we're doing. That's the problem I see with all these other players.

Question: No, I think that's so true, and I think that's a very important thing to think about. That's what you said in your tweet, that Palantir can do a lot of these things that some of the Snowflakes and Datadogs can do in terms of being a data lakehouse, but taking it a step further and actually operationalizing all that data to make it create value, which is what you keep talking about for the business.

Answer: And that's what I put in the tweet, because yes, you need all the data in one place. That's great, that's the first step. But when you talk to a business executive, like you go talk to a CEO at one of these companies, and they're like, "I just went through this cloud transformation. I spent three or four years, tens of millions of dollars. I have all my data sitting in a data lake now, but my business hasn't changed, right? My two reports are now one, they run in the cloud, and they cost me two or three times as much. How is that transformational?" That's what they see out of the value of those things, because people stop at putting stuff in a data lake.

Question: And these people are upset at this stuff, right? I'm assuming they're annoyed.

Answer: Yeah, they're annoyed. I mean, if you just spent tens of millions of dollars in two or three years of all these promises about how getting all your data in a data lake is going to solve the world's problems, that's one step, but that's not the end goal. And that's where a lot of IT practitioners fall short because they're focused on getting data in one place that will solve all their reporting problems. That's where they're focused. The business is like, "Yeah, that's not transformational."

Question: Now, when you say in this tweet, "I know it sounds like black magic, but our connected tool chain is a single secure model," is that basically what you're talking about, the ontology at the end of the day?

Answer: No, so like I said, I wasn't actually hitting on ontology here. I only had 30 minutes to do this whole bit. I had maybe 10 minutes. So, my focus here was actually the friction between tools. So, you go to the cloud, and you stand up whatever cloud tool you want, Redshift, BigQuery, all these things. It's great. Now, okay, I want to create a workflow and an application on top of it. Well, how do you secure that? How do you do that with security and governance across all those things? How do I run a simulation on top of those things? Each one of those is a new product in one of these clouds, and none of them work together. And that's where the false dichotomy has been sold. All these cloud providers say, "We are going to take away the non-differentiated heavy lifting." That's AWS's kind of tagline, right? The problem is, when I go into their cloud now, I still have to stitch their own tools together to get them to work. And so, you're spending all this time stitching tools together, security models are different between tools. And wait a second, how is that any different than what I was doing in my data center? How is that any different than what I would do if I bought a bunch of SaaS products? It's not, right? You just get one bill.

Question: This was really, really helpful. I think a lot of people are getting value. In the context of AI, once you have that ontology, I'm assuming these AI agents, that's where it goes to a different level in terms of sueding value, right?

Answer: Well, imagine this. So, imagine you have a digital twin of your business process. We call this kind of like north of the ontology. Once you have the ontology, it unlocks simulations, it unlocks agents, it unlocks workflows, it unlocks SDKs. All that just becomes native once you have an ontology. So, all of those other business practices, hey, you want to go do process mining on how your orders are flowing through? Cool. That's like a widget. That's not another tool you got to go buy. It's not another thing you got to go stand up on a cloud. Great. It's literally a widget in a Workshop app, and an hour later, you can have a process mining workflow in your application. That is the difference. First, I'm going to go spend, go do an evaluation, an RFP to go find a process mining tool. I'm going to plug it in. I'm going to spend months installing it. Or even if it's already existing, I've got to go spend time integrating it, figure out security, go through my architecture review. It takes weeks, months to go do all that. And in the tool, we can do it in a few minutes. That differentiator is huge.

Question: Right. This is awesome. This is really cool. I see a lot of comments in the chat. People are getting a lot of good info. Last question for you, Chad. How cool is it working at Palantir?

Answer: Coolest thing I've ever done.

Question: Because you've been in big tech for a long time, but Palantir, I think, is a little different for you.

Answer: We just operate differently, and the only thing at the end of the day that we care about is winning. I like my stock-based compensation.

Question: Going to people who say that. Well, actually, one final question. How does it feel for Gary Tan to follow you and you turning into a little tech celebrity? What's going on here?

Answer: I'm excited just that we can start to get the message out to everybody now. Gary knows better than anyone what we do, but his connections and everybody else, the lack of communication for so many years about what we do and how we operate was tough for people to understand. What is Palantir? When Palantir showed up on my doorstep, I had never heard of the company, didn't know what they did. And that's always been the problem, right? It's a pseudo-Black Box. And it's really the unwalled garden. It's the place that is very open. Everything has APIs, SDKs, the ontology SDK. It's interoperable in ways that other platforms are not. But we very much get this black box because it's CIA, it's secretive, it's this and that thing. In reality, our platform is actually very extensible and open. So, I love that Gary and others start to follow me because they can help spread and show the message about what it is we actually do and how we're different. Finding new people, and I've already had people connect and start to talk to me about different work they're doing in ML engineering and all these different things, like how can I try out Palantir. All of that is super exciting to see people build with a platform. That, to me, is the most enriching thing: something that I'm working on, and people can go build their company on it, make real transformation based on it. That's exciting.

Question: It's really exciting, and the thing I think about you that's incredible and just impressed me and Arny when we met you beyond imagination is that you work in big tech data, you're an engineer, sometimes you're in sales meetings, and then you're also growing this Twitter brand, like doing marketing and getting the message out. So, you kind of do everything, and I think it embodies that winner mentality at Palantir. It's super exciting to see it all play out.

Answer: The thing I say, you know, when I used to have teams of people and new employees, there's no job that is below you. Not that Twitter's below me, but there's no job that you should look at that you shouldn't be a part of doing if it's going to help you win.

That’s it for today - see you tomorrow!

This newsletter will always be free and never have paywalled content. To support the newsletter, you can give a gift subscription below. Thank you for reading daily!