The Exchanges

Q How are business models different in this era versus the SaaS era?

A So I would say at the applications and at the teammates layer, SAS took us from the era of perpetual license, which was enterprise software to subscription, where you charge per user per month. Now we are moving into an era of consumption based pricing or outcome based pricing. I don't pay you per month. I pay you for what I use. And that's where this era is going, where I don't charge per user per month. I charge for the activity I do for you. I charge for the task I complete for you. So that's where the business model is changing. And that's a very important component of these AI native companies. They have to move from subscription to consumption based pricing models.

AI assessment note: “Now we are moving into an era of consumption based pricing or outcome based pricing.”

Q So let's talk about the full stack of technology. You invest end-to-end semiconductors to applications and software. Could you lay out all of those layers and how do they work together?

A Absolutely. So first and foremost, I would say, as I mentioned earlier, this is the intelligence era. So our belief is humans and AI will work together And take us into this new era of collaborative intelligence. And in this era, AI will enable humans to get to super humans level. And what I mean by that is if AI can help augment me and free up time so that I can go do things which I really want to do. So they can amplify my capability. AI can amplify my capabilities. Give me time to do creative things or do things I could never do. So to enable that, we believe there's a collaborative intelligence stack which is being created, which starts from the hardware layer, which is the semiconductors. The next layer up is the brain, which is the model companies. Once you have models and you have hardware, this gets packaged by a cloud service provider. Or by an inference company. You need to bring your own data, so the data layer becomes important. Once you have data on top of models, on top of hardware, now you need middleware and tooling to do something with this plumbing layer. Once you have that, you build intelligent applications. And then finally, the most exciting layer of this stack is what we call AI teammates, which are built on agentic technologies. And what are they? They're digital companions that work with humans to take us to superhumans level. And that market is a massi…

AI assessment note: “starts from the hardware layer, which is the semiconductors. The next layer up is the brain”

Q Some people call it also the AI super cycle. Where do you think we are in the evolution of this super cycle?

A Yeah, I would say we're still early on in the super cycle, but the valuations are really high in some of the spaces, and this is the standard hype curve. When new waves come, expectations are ahead of reality. But if I go out four, five, 10 years out, it'll be a reality. The same thing happened with the internet, where things just astronomically were overvalued in three, four years. But the difference today is the reason it's going to be a hundred X bigger, because when the internet happened, people didn't have connectivity. Today, 7,000,000,008 billion people are on the internet using mobile or their PCs. So you can have things grow to a hundred million monthly active users in one month, in two months, in three months. It took 10 years for Amazon to get to that point. Versus when TikTok happened, people were talking about, hey, nine months, six months to get that. Now people talk about like a week to get there. So I think The reality is it's early days, but some of the things have already taken off, whether it was with chat GPT, but let's look at vibe coding. Companies are doing a hundred million in revenue, two hundred million in revenues. They're going to be doing ten billion in revenues. We'll blink and that's what they will be. But on the other side, if I look at enterprises, still early, lot of hype, but very few enterprises. Can actually do anything with this technology,…

AI assessment note: “I would say we're still early on in the super cycle”

Q So one of the themes that we've covered throughout many of the questions and your answers is what happens when we get super intelligence? There's lots of opportunity, but what does that actually mean? For people at their jobs. And when do you think we're, when do you think we're going to get AGI?

A Yeah, I think AGI to me is like far away, but far doesn't mean 1520 years, but it's not one, two years. I would say already we are seeing with machines, they can do certain tasks at the same level as humans, or even better than that. But humans, I believe, Are not dinosaurs. Technologies come and go. They adapt themselves. And that's what is going to happen here. AI is going to be a tool where humans are the jockeys. They'll use this horse to adapt themselves, to endorse this, adopt it, and do things that they should be doing. And this is the same level, right? Like when you're an exec in a company, you have a team. You don't do everything. Similarly, if you're an employee, you don't write by hand. You just type. Now you'll be able to verbally see it. So what does this do? First, intelligence is going to get democratized. Knowledge is going to get democratized. Expertise is going to get democratized. So why should only a technologist or an engineer be able to start a company? So I think the number of creators or number of entrepreneurs is going up by 10 to 20 X. It's going up by 10 to 20 X because now using vibe coding, using AI as a tool, you will be able to create. Something. You don't even need to know marketing. AI can do it for you. So to me, knowledge, expertise is gonna get democratized, and humans will do things that we only dreamt about. Using AI as a tool, using AI as…

AI assessment note: “I think AGI to me is like far away, but far doesn't mean 1520 years”

Q And what does that mean for the AI industry today, for early stage companies?

A Yeah, I think it still remains the same, right? Old is gold. At the end, venture, entrepreneurship is an apprenticeship based business. And in the AI era, overall, it's a hundred X opportunity. I've been lucky to be involved in the web era, the mobile, the cloud, the social, the not so good eras. And then the AI, I think those eras were a 10 X opportunity. This is a hundred X opportunity. So the kind of success that'll get created won't be just ten billion dollar companies. There'll be many hundred billion dollar companies, but again, what will matter the most is the people who are creating these companies. That's not going to change, at least for Mayfield and me at the stage we invest, which is inception and ideation stage primarily.

AI assessment note: “what will matter the most is the people who are creating these companies”

Q How do you and the team at Mayfield value companies? What's the strategy there?

A Uh, I would say at Mayfield, since we are investing at the ideation stage, we are looking for a win-win with the entrepreneur. Clearly from where we were investing in the cloud era or the mobile era, valuations are dramatically up. But for us, a win-win with an entrepreneur is, hey, everybody needs to make money. We respect their business model. They need to respect our business model, which is VCs need to have margin in their business. So we need to have with small checks enough ownership for the risks we are taking. And luckily to succeed, there are enough entrepreneurs who believe in that philosophy that we can still survive and thrive. So that's what we are trying to do. There's no science. It's really an art. You can't be in every, on every cap table. We can't invest in every company. Every founder may not work with us, but can we find seven to eight inception stage companies a year where we have meaningful positions and we can work with these entrepreneurs to help create leaders of tomorrow? There are. So we try to stay away from people who have unrealistic expectations and why? Why get into business with people? And it's vice versa. My standard answer to people is if they ask for something, sorry, we don't have such a product, right? Like I love In-N-Out, but I have their veggie burger. Now people will say In-N-Out only sells one kind of burger. So what is this veggie bu…

AI assessment note: “we need to have with small checks enough ownership for the risks we are taking”