The AI Investment Landscape: Opportunities and Challenges

Welcome to the Venture Capital Podcast. I'm John Bradshaw, this is Peter Harris.

Howdy.

He's a vc, I'm a founder. We're here to fight.

There we go. Round one, round one.

Different perspectives. The Wall Street Journal in the last this week talked about last quarter, 31% of all investments were deemed specifically AI investments over SAS investments.

All those people are going to lose their shirt.

So as we navigate, as we navigate this, I've got my perspective on this space in particular, but let's hear it.

What's your perspective, John?

I think most VCs and founders that are specifically choosing to say I'm an AI company are going to flop.

Why?

I think that AI in most cases, let's look at a Zendesk. It's customer support software. AI is a fairly inexpensive product feature versus being its own set apart system. So I think it's something of hey, it's an add on versus a standalone product in most scenarios unless you're like an OpenAI.

So here's the big question. When mobile came out, I would argue so you've had a lot of big waves, right? Internet big wave, right? Mobile big wave, social big wave, right. I think a lot of companies looked at mobile similarly to the way you just described AI like oh, it's just a feature, we just need to add mobile and then boom, we're there. And there was a lot of talk around like no, that's not, that's like not how it works. If it's not like in your DNA, if you are not mobile first as we called it back then, right? Like you were not going to be as successful. Okay, so I'm curious, what do you think? Do you think like short term or longer you need to be AI AI first and that that gives you an advantage over incumbents who just aren't built from the ground up to implement AI. I think is it the same or is it different?

I think there's a short and long term play the short term, I think there's a massive amount of cash that that VCs will write to zero. Specifically that are quote unquote, AI investments, which is very. And specifically it's a, hey, we are AI customer support versus we're adding AI into our platform and not just raising to say hey, we want to go build these features. Like one of the companies we're talking to with right now, that's what their round is specifically for is to build AI in their existing app. I do think those that have an AI first perspective long term over the next five to 10 years are going to create significantly better products. I think the thought leaders will go there and in markets depending on what they are. I mean a lot of current SaaS applications, I think they'll just still be, they'll be gobbled up and acquired. SAP has been here for a long time and has survived the mobile trend. They've survived.

But they're not trying to be like mobile. Right.

Well, just like given it. So let's give a better example. What do you think?

Like what company is a better example with an AI?

Because I think a lot like customer support, a lot of SaaS, a lot of it is there are exist a lot of these SaaS companies are just, they're just workflows and AI think helps optimize it. The work, Some of the workflows.

Yeah. I mean my perspective honestly hasn't changed much from last time we talked about my overall thesis on AI. And if you, you want to listen to that, you know, go find the podcast where I do kind of a deeper dive. But on a high level, my belief is every business breaks down to the three components. You build a product, you distribute, you, you market and sell the product and then you distribute the product. You get the product in people's hands, right? That's it. Those three, those three components. The Internet drove the cost of distribution basically to zero. I can now like set up stores, I can sell globally. Right. I don't have to have physical locations everywhere. And when that happened, the price of, you know, marketing and sales went up and the price of product creation went up. Right. You look at like who are the most highly paid people at any startup it sells and dev, right. And now AI is really driving down the cost to create products. In fact, I think most, like most companies are starting to see, and the reports are starting to implement that or indicate this, that the biggest gains that they're seeing from AI implementation is on dev and customer support and stuff like that. That's all product related. So the cost of product is driving down to zero. And so what is that going to do? It's going to drive the cost of marketing sales through the roof, in my opinion. And this makes intuitive sense because you're just like, if I can now go on and within like a word prompt and have AI spit out a ton of code and write an app for me, right? There's going to be a lot more apps. There just are. And so now Discoverability is going to be this huge thing, right? How do you solve for that? And so if you believe that that is, that is accurate. Then you know, you look at like, who already has all the customers? Well, Zendesk already has all the customers.

Correct.

And so then the question is, what's harder? Is it harder to implement AI into your existing product or is it harder to acquire customers? And my thesis has been that it, it's going to get incrementally harder and harder and harder to acquire customers. Whereas again, I just said like implementing and building product is getting easier and easier and easier. And so implementing AI is going to be easier. So it's, it's like, can I become the Zendesk faster than Zendesk can become me, the little AI, you know, customer support app. And my guess is that Zendesk is going to be able to acquire or is able to become the AI powered customer support app faster than that new company, that new entrant is going to be able to gain enough market share to actually be a threat to them. I do think, though, talking out of the other side of my mouth, that there are a bunch of companies that come from it from a AI first perspective. Just like mobile. There are a bunch of companies that were just, they just weren't possible before mobile, right? So think about like Uber and Lyft, okay? Like fundamentally not possible until mobile takes off. Right? Once mobile takes off, then you can build an insanely large business. But if you're going to do it, you don't build it on the desktop and then eventually write an app to do like, you know, peer to peer car share. Right? You start mobile, you are mobile first. Uber was mobile first. And I think there are a bunch of applications where like there are no incumbents because it wasn't possible to do that until AI rolled around.

Like what?

So you know, education I think is like an interesting example.

Okay, like school AI locally.

Yeah, maybe, yeah. So let's talk about school AI locally. AI should allow teachers now, for the first time ever, to create a more personalized teaching experience with their students, with agents. With agents, right? That wasn't like fundamentally possible before. Right? So there are no big incumbents that are, that have built tools, you know, to do that.

What about robotics? Same. Same scenario. Or do you think, or do you.

See it differently, maybe robotics. I honestly am not a robotics investor, so I haven't thought a lot about that space. But I do think that there are other applications where, yeah, there isn't a big incumbent yet. And I've seen a handful of these startups popping up and those are kind of interesting to me.

Which startups in Particular trying to think.

Like what are some applications where it's like, yeah, that wasn't even really possible before actually. Okay, so I'll give you kind of an example. I mean there's, there's a bit of a threat that, that Google is kind of the incumbent, but I think the real incumbent is people. So like I'm looking at an app, an AI app right now, just not as an investor, just personally that goes through and automatically cleans up my inbox for me and like sorts everything, highlights the stuff that's important, drafts responses automatically, stuff like that, where it's like in the past I would have had to have like an executive assistant, a human being go through and do all of those things and like that human being, because they're not like, you know, they're limited by the number of hours that they can work, they're not scalable. It's not like they're an incumbent. Right, they are an incumbent but they don't have the market share. So you know, I look at apps like that, you know, maybe that's not a fair one because maybe you could argue Google's the true incumbent. They'll implement these AI functionality and displace those companies. So that makes it a little more nuanced. But I think at its core that idea of are you building an application that historically required a human being to do it in a non scalable way and if so, and there isn't an incumbent that really controls your distribution channel like a Google with email, then maybe that's actually a pretty interesting place to play and maybe there are going to be some really interesting businesses to arise from that.

The agent. The agent marketplace.

Yeah, yeah. But you know, I think there's a bunch of other stuff where it's like, and I think the majority of it is the case where yeah, I think the incumbents are going to win. They're going to be able to implement AI faster than new entrants because I just think fundamentally incumbents are going to be able to implement AI faster than new entrants are able to acquire market share and be a significant threat to them.

I'd agree, except in the agent space.

Why is that?

I think there'll be a huge shift and I think a lot of people won't know how to train agents on their own. And so I think we're going to see more and more apps emerge.

So you think the training, the agent is very challenging.

For example, there's a ton of tiktokers that will show you how to automate Your life. And most people, I don't think are able or willing or want to change. And I think if I were to invest in the AI space, I'd be investing, I think particularly in specified uses of AI for AI agents.

But isn't that kind of what I was saying? Because you use an agent to replace a human being. Right. But I do think, like, there is value to having being, to being built from scratch with AI as your core.

But I think, I think under the Trump administration, there will be a lot more acquisitions than there were before. And that's, I think what will happen primarily.

Yeah, could be. Well, I mean, a lot of these incumbents are going to need their own AI teams. Right. And some of them are just going to go buy these companies. Right. And they may end up being expensive. They may not at the same time. Right. It might be like, hey, there could be a lot of acqui hires, which is fine.

I think that's what I was talking about at the beginning is I think an AI customer support first tool will be much better than Zendesk. But Zendesk will acquire them.

Yeah. Or they'll, or they'll just figure it out themselves and like, they might be slightly better than Zendesk, but it won't matter because Zendesk already has a brand, which is going to be the really hard thing to replicate.

I also think a lot of people forget that a bundle that a lot of times inferior products, superior products will lose to inferior products that are sold as part of a bundle.

True. Although you could also look at disruptive innovation, which basically almost not entirely, but kind of says the opposite, which is you have a company that has a product that's really good at one thing and terrible at everything else, but that one thing is the thing that matters. And eventually they displace the other companies that are better at everything but that one thing.

I think the winners in the space are the power companies. I think the winners in the space, I think power to run just a simple prompt versus an AI prompt, the power usage is 33 times more. Yeah, I think the hardware companies like Nvidia are the winners and I think LLMs are going to become a commodity.

So I actually don't believe that LLMs are going to be commodity. Okay, so here's my thesis on that. I used to worry that LLMs could be a commodity. However, you do have like this virtuous cycle. More data, more engagement, results in better models over time. But if you think about LLMs as a, as a parallel to AWS, I would argue that AWS is way more commoditized, like way less proprietary than an LLM. And yet AWS dominates the market. Right. Why do they dominate the market? Well, because they've got the full ecosystem of all the tools and other apps that are built on it. They have the brand, they are the verb for moving into the cloud. They own the space. And that doesn't mean that Microsoft Azure isn't successful and Google Cloud isn't successful and don't generate lots of revenue and profits. But, but on a comparative basis, AWS is king. And I think from an LLM perspective, it's not like I can't just spin up my own server at home today. Any company could and be in the cloud and connect to the Internet with their own server. But nobody does that because it's so much better. Well, I mean they do, but no big tech company does. Nobody that's horribly successful.

They do.

Who? Who? Tell me who does that at scale across their entire IT infrastructure.

Fintech is a bigger as a big space that will do that more.

They're doing some of it more.

So I can go give you the list.

Okay, okay, okay. You're gonna hold me to like being absolutely specific with my words.

I mean no, I think right now.

But my point is like AWS is a business that is worth probably north of a trillion dollars to Amazon today, despite the fact that it is arguably more commoditized than an LLM and doesn't have the barriers. So I think the big LLMs, OpenAI, Anthropic and others are going to be successful because it's not, it's not even about the underlying lm. It's about all of the other ecosystem and infrastructure and brand that matters.

So if we trust Google Gemini, only 50% of the world is using a public cloud today, which is higher than I thought it was going to be. Which would mean roughly one out of every two enterprise companies you think is using Google, Google, gcp, Google Cloud, aws, Azure is self hosted.

Great. That only reinforces my point that despite the fact that it is so commoditized that half the world doesn't even have to be on an AWS or comparable platform, there are still like multi trillion dollar markets being spent in that space.

Oh for sure. But I think, I'm not saying that LMS won't be super valuable, but I think it's the energy markets that powers it. The hardware markets maybe.

I agree. I think, I think infrastructure, I think the big LLMs, I think the incumbents, I think apps that are AI first for AI, like things that just weren't possible before AI. I think companies with exceptional user experiences will always do well. That won't really have anything to do with AI, but you know, may masquerade. Masquerade as an AI company when really, really they're a UX company. Yeah. Infrastructure. Yeah. And probably power companies, except for power is kind of a commodity, so. And we basically pay nothing for power. And it's. They're all monopolies, natural monopolies. So I don't know that I would be plowing a ton of money there, but Microsoft is. Yeah, well, they need to. Right.

Google is.

They need the power.

They're buying nuclear power plants. So anyways, anyways, I was just curious how you are navigating as if you see the early AI space.

Yep, that's how we're doing it.

And if you were a founder, talk about the founder side real fast. I think if you're working on the AI version of Zendesk first, but it's AI first, I think you could build a very good business, especially if you focused. You can get more specific and focus on a subset of, of a customer base and then expand up if you want to.

Yeah, yeah. But I'll tell you, as a vc, it doesn't get me very excited.

Right, but you can bootstrap that.

Yeah, I think there will be. So, so the flip side of this is that's where I think the venture opportunities are. If I'm like an individual and I just want to like profit off the AI wave. Yeah, go, go start like go pick a like super now niche. Be the AI customer driven, customer support driven tool for that little niche where you can train your agents to be like phenomenal in that niche. Right. Which is no different frankly than like starting like before AI starting a consumer support or like a customer support app that plays like very, very well in a very, very tight, specific niche. Same thing. You know, we're investors in a company called UAI and they have a really interesting platform that's like, you know, squarespace for AI tools and apps where you can just go in drag and drop without having to know a lot of coding and build an AI specific app. And so like there's an opportunity for people just like in the early days of the Internet to go around door to door and sell websites to businesses to sell AI apps. It's like, hey, you have a plumbing business, Let me build you a custom AI, you know, app for your plumbing business or your retailing business or whatever. It is like I think those opportunities will also exist they're not venture backable individually but like you couldn't that doesn't mean you can't make a lot of money doing them and do very well.

For sure so I think the AI space will be very interesting I think there'll be a lot of churn a lot of burnt I think a few winners outside of the incumbents I think a lot of people still just don't know what the future looks like it's too too early in the game.

Well Skynet is just going to take over and we're all going to work for the AI supreme overlords eventually so it's.

Fine that's a possibility Elon OpenAI talked about it being 2025.

When when the singularity arrives when Genai arrives we'll see.

All right well thanks for watching go to VC FM or comment below on these vid on these videos and we'd love to know what you think on how would you navigate the world of AI either as a founder or as a vc.

Thanks so.