Access to capital will profoundly impact the performance of AI models: MongoDB CEO Dev Ittycheria

The New York City's tech landscape has thrived in recent years. One company that's played a huge part is MongoDB. It transformed from a small startup, I remember when it all began, to a key player in the global tech ecosystem, powering applications for some of the world's largest enterprises. Joining us right now is David Cherry. He is MongoDB's President and CEO. He is celebrating his 10th year as CEO. So, first of all, congratulations on 10 years. Thank you very much. In the job. Good to be here. You're not just a gig worker. I was going to say in the gig, but this is not really just a gig anymore. I want to actually get into what's going on at MongoDB, but in terms of the larger tech scene in New York, there was always this view that Silicon Valley was going to turn into Silicon Alley. Did it ever really happen? Well, actually, interesting enough, a lot of people were skeptical that you could build a company like MongoDB in New York. It was a deep technology company. It was a lot of, I would say, consumer and advertising companies, advertising technology companies built in New York. So, when MongoDB was started, there was some skepticism that could you really build a... Well, your founder was a former advertising man who sold possibly the most important advertising company to Google. Correct. There were three founders from DoubleClick who started MongoDB, and I wouldn't be here if it wasn't for them. And then they brought me in 10 years ago to help build a business. I joined when we were roughly about 40 million revenue, 250 employees, and today we've grown... And the view was you couldn't do this in New York because? Access to talent, access to capital was limited. You think the access to capital would be everywhere. Well, in the venture business, it was a different era 10, 15 years ago. I tried to start a company in the late 90s here and then moved it to Boston. And, but I think things have radically changed. OneTalent, Amazon, Google, Meta all have huge offices here in New York. Obviously, access to capital is second to none, as well as customers. You walk out the front door, I stare at my customers, some of the largest customers are 10, 15 minute subway ride or walk. And so it's amazing. And from New York, I can take a flight and get anywhere in the world nonstop. Here's a New York question. Did you ever think about moving to Miami? Because as we've been talking about COVID, and there was a whole view that the whole world was going to change and nobody was going to go to the office anymore, or they just live from, live wherever they wanted to. And then New York was a super expensive place to do business. No. So I felt the opportunity here was too big. I recognize that New York went through some issues and challenges during COVID. A lot of us work from home or remote locations. I did have employees move all over the country and in Europe, even different parts of Europe. And some people did move to Florida, but we never contemplated. So what are you doing now, by the way, since I don't know if you, if you were listening to the story that we had just before about forcing people at Stellantis back into the office. We were also talking about Amazon, obviously has now gotten everybody back to the office five days a week, but it's complicated for some people who've moved farther away. Yeah. So we actually believe in hybrid work. And I think it's a misnomer. A lot of people conflate working hard with being in the office. What we found, and we've actually done studies on this, is that our top performers are top performers, irrespective of where they work, you know, in the office, hybrid or fully remote. Where there are issues is our lower performers. We do see more lower performers that are kind of in the remote category. And so the onus is on the leadership to really make sure when someone is not performing to the levels expected, that they take quick action either to remediate the problem or ultimately decide that they're not going to work out. Can I talk just to you about your own business and the transformation of it? As AI becomes a major component of all of this, how do you think of what you're able to do and the data you have given your database business relative to some of these other companies that may have some really strong sort of frontier-level models? I'm thinking Microsoft and the like. But you know, you've had people like Mark Benioff at Salesforce say, look, you actually don't need these large language models to actually get to the promised land. In fact, what you need is the data married with models that actually just work on your data. Well, the way I think about AI, and I think most people should think about it this way, is there's three layers to stack. There's the chip layer, which obviously NVIDIA's done great and probably generated 140% of the profits in the industry. There's the model layer, and you just mentioned a few. There's obviously OpenAI just raised $6.6 billion and seems to be separating themselves from the pack. They have now 250 million weekly average users. And then there's the application layer, and that's the layer that we play in. And when you think about modern AI apps, you need a platform that can query rich and complex data structures, and we're perfectly designed to do that. So we see a lot of startups and enterprises building the next generation of AI workloads on top of MongoDB. So that's where we're focused. And value will accrues first at the lowest layer, which we've seen with NVIDIA, but over time, value starts accruing at the higher layer. But you think it accrues at the higher layer? And what happens to the middle layer? Because some people think that the middle layer gets commoditized, and some people think the middle layer is hugely valuable. I think it's still to be determined. There's obviously a debate about open versus closed. So open AI is obviously a closed model, but then you have Meta with their Lama models and Mistral with their open source models. I think in the world, there will be a bit of both. But it's clear open AI seems to be separating themselves from the pack, and I think access to capital really matters. But are you a believer that everyone's going to ultimately use a large language model to query all of this content and data that you have? Or are they going to be using these sort of smaller, more specific models? And that's why I'm thinking... The reason I asked you about Mark Benioff and Salesforce and his agents, if you will, is it's a different model. I think there's no question agents are going to work, and the jury's still out. What we see is that the performance of agents is still kind of inconsistent. I think the real key comes down to the reasoning value of the model. And I think the more reasoning power you have, which is directly tied to compute, which is also tied to capital, really differentiates the different models. And my sense is that access to capital will profoundly impact the performance of these models. So my bet is that over time, you won't have this many model providers. You may come down to one or two, one closed and one open. And obviously, open AI is in the driver's seat today. You beat on the top and bottom lines for Q3, but the stock has not performed in reaction to that. What are the investors telling you? Investors, obviously, we've had a little bit of a hiccup in Q1. We took guidance down. We recovered in Q2. I think investors are trying to understand, what is the growth opportunity at MongoDB? We think we have a massive runway for growth. And obviously, they're taking a little bit of wait and see attitude. Our multiple is still fairly high. What do you think you need to do to satisfy them? I think we need to continue to produce consistent results, a quarter of a quarter. Obviously, we priced our IPO seven years ago, when I was actually here, at 24. The stock's now around 290. So generally, investors have done well in the stock. But we think we have a big opportunity in front of us. And I think we just have to deliver quarter in, quarter out.