The alpha, twice a week
That was 5 min.
Get the best new moments in your inbox every Wednesday and Saturday. It's free.
Check your inbox
We sent a link to your address. Click it to finish signing up. Not there in a minute? Look in spam.
or
Sourcery · 5 min · Nov 11, 2025
AI buyers now seek working systems, not proof that software works
Alex Karp says visible AI successes have shifted buyers from skepticism to asking how to implement it well. He describes sales cycles shrinking from years to months, and why Palantir resists process-heavy deals.
Alex Karp is CEO and cofounder of Palantir.
Co-founded Palantir with Peter Thiel and has run it as CEO, selling data and AI software to governments.

Molly O'Shea, Host · Founder, Sourcery
Ex-Upfront Ventures investor; has interviewed Alex Karp and Palmer Luckey.
About this segment
Alex Karp says visible AI successes have shifted buyers from skepticism to asking how to implement it well. He describes sales cycles shrinking from years to months, and why Palantir resists process-heavy deals.
From “Alex Karp, CEO of Palantir: Exclusive Interview Inside PLTR Office”, Nov 11, 2025. Watch the full episode
Thanks. We'll look at and fix what's off.
“it used to be no one believed the software would work. Now everyone believes it should work.”
Alex Karpat 29:22 · Sourcery29:22Takeaways
- Buyers increasingly arrive convinced that AI software can work.
- Implementations have compressed from years to months.
- Palantir turns down sales processes that consume scarce delivery capacity.
In this episode
Full episode- 22:38Palantir builds for what clients need, not what traps them4:52
27:50AI buyers now seek working systems, not proof that software works4:36
- 3 more moments in the full episode
Transcript
Transcribed by machine (Whisper) from the episode’s audio, so a word may be off here and there.
- 27:53
So on AIP, it seemed like that was a big pivotal moment for the company and getting into this next round. of AI, which is something in transformation that you've embraced full, wholeheartedly versus maybe some other people to the extent that you go into different corporations and then you make the products for them. On top of our current products or we build new ones. Yeah, that's true. Would you say that that was a major component to why sales cycles have compressed and why it's been easier to sell?
- 28:23
You know, it's a really good question. It is that, but also the zeitgeist has changed. Although there have been a lot of projects in AI that haven't worked, most leaders in the corporate world and government know of projects that have worked. A great portion of those are in Palantir. So in the government, there's a deep understanding of the government side of Maven and other things, Hivemind. And then on the commercial context, we have all these very public customers and people begin to see their unit economics changing.
- 28:58
And particularly important, they're able to change them while being in the public sphere. So you can do what you had to do in private. But you used to have to buy the company, take it offline for like five years and then scrap the costs and we sell. And then you could do this while remaining public. And there's lots of economic advantages to that. And so you can also do it in private context. And so the thing that you see as a common through line in AI on the positive side is it used to be no one believed the software would work.
- 29:28
Now everyone believes it should work. And they wonder, well, if mine isn't working, where can I find something that is? And more and more and more people, honestly, in the U.S., it's not true in Europe. In Europe, they're still like, it doesn't work, but hopefully this will fail because it's all American products. But in this country, people are like, I know this should work. I turned on TV and there was someone I know or someone I've talked to at a conference and said, yeah, OK, good.
- 29:57
It is a little weird when they show up. The person showing up is going to be carp. That's not everyone's cup of tea. And someone who's like looks 18, but is actually 25. And it'll work. I mean, I think that's what's like going around the country. and that means both like from introduction to sale it's quicker and as importantly from sell to actual output we have a lot more authority like we used to basically get in these enterprises and i have to spend a half my day explaining to people look i know the person down the hall is your tech person but they used to be a school teacher in music this is actual real example Okay. Yeah. And what, you know, I like the idea of a music, but it's like, and what they're saying about the tech stack, you know, and then you could do five years or you could do at the prime, I used to say five years versus nine months. Now it's like five years versus three months or two months. And there's actually much more of a, please just tell us how to do this. Not always.
- 31:01
Yeah. And I would say, interestingly, our newer partners, customers are much more wanting to hear, OK, I know this works. It's gone from I'm not sure this works, but my wife is on the line in commercial and government like five years ago to I know this. I've heard and know this will work. How do I do this optimally? And then, you know, the bottlenecking we have currently means like, you know, I mean, I tell I'm telling governments all over the world.
- 31:29
look, we're not showing up to do a sales call for Maven. You know it works. We know it works. Show up at my office, explain to us how you're going to make this easy for us because we don't have huge bandwidth. I mean, I'm not talking about the US government. That's a completely different thing there. We will bend over backwards and do whatever it takes. But like allied countries in Europe was like, yeah, you know, they're like, oh, we need five sales meetings and a PowerPoint.
- 31:54
We're like, great. I can recommend some other company. It won't work, but you'll get all the sales reps. All that cycle, all the forms, everything you want. And I was the other day talking to a very important European leader. I was like, look, you have an intel service, you have a military. They should evaluate what works. They're going to tell you this works. Nothing else really does. You can buy, you want the thing that doesn't work but sounds good.
- 32:22
They'll do lots of meetings with you. We can't do them. We just can't.