Chit Chat Stocks - The State of the Cloud in 2023 with Shawn Wang

Episode Date: January 5, 2023

The "Cloud" in its simplest form refers to servers with data stored on them and those servers can be accessed over the internet. Many companies like Amazon, Microsoft, and Alphabet lead this fast-growing industry. Listen in as Brett and Ryan ask questions about the cloud and what growth will look like over the next few years. Enjoy the show! ***************************** This episode is sponsored by Stratosphere.io, a web-based terminal for financial data, KPIs, and more. Try it out for FREE or use code “CCM” for 15% off any paid plan. Sign up here: https://www.stratosphere.io/ ***************************** Want updates on future shows and projects? Follow us on Twitter: https://twitter.com/chitchatmoney  Subscribe to our Substack to receive free show notes and charts that go along with every episode: https://chitchatmoney.substack.com/ Interested to see more of Shawn's work? Check out their Twitter here: https://twitter.com/swyx?s=20&t=pMnKcZ7lY8rGjzEnSZavYg Contact us: [e-mail] Timestamps Shawn's Cloud Background | (5:54) Who are the Cloud Companies? | (22:07) What about Snowflake (Ticker: SNOW) | (46:00) Disclosure: Chit Chat Money hosts and guests are not financial advisors, and nothing they say on this show is formal advice or a recommendation. Brett Schafer and Ryan Henderson are general partners and portfolio managers at Arch Capital. Arch Capital and its partners may hold securities discussed on this show. Learn more about your ad choices. Visit megaphone.fm/adchoices

Welcome to Chitchat Money. This is our Thursday deep dive interview. And today we have on Sean Wang. This is not our typical kind of interview. It's not any one specific company. It's the cloud industry overall. And you're going to hear us talk about it, but it's, I think, an industry that a lot of people are very curious about, but don't quite grasp the intricacies. Sean clearly does. And you're going to see that. And he does a really, really good job explaining the ins and outs of the industry. And my favorite part was probably when he talked about the differences in the big three cloud providers and how integrated or entrenched they are with their customers. It's really a fascinating conversation, just given how wonderful these business models have proven to be. And you can see that just looking at Amazon and Microsoft and even Google is turning towards there, turning towards that profitability, but you can see it in the financials that they break

out for these businesses. So it was really exciting. Did you have any highlights specifically? Yeah. I think the whole thing was great. The discussion around Cloudflare was really fun where talking about how they're trying to attack them from a different angle. And man, I think just the long-term contract. We don't talk investing specific things, but I think there's a ton of investing takeaways you can have from here and why these businesses do look promising and why they're going to be so important this decade. But first, before we get to the interview, why don't we talk about our new sponsor for 2023, and that is Stratosphere.io. Stratosphere.io is a phenomenal platform, and it is a great way to use research,

get data points, financials, all this good stuff. And it's very accessible. And the best part is it's free. So you can use it. You don't have to spend $20,000 on a Bloomberg terminal. You can do tons of different things with stratosphere.io. You can have up to 35 years of historical financial data and KPI. KPIs, you have company-specific metrics like Netflix or Spotify subscriber counts, Costco locations, tons of other stuff for specific businesses. And the best part of it is you have... How do I even describe this? You can get all the financial filings. You can get all the SEC filings all in a centralized location. So you don't have to use the Edgar service, which is clunky, or go to all the different investor relations websites.

This can be your homepage as an individual investor. And on top of that, they have a premium plan for professionals, where if you want to use our promo code CCM for 15% off, you get 15% off your paid plan. That is promo code CCM. The link will be in the show notes. Check them out. It's an awesome service. And the best part is it's free. Ryan, anything else before we- No, I want to follow up on this because this is a new sponsor and I'm really excited that they're sponsoring us because we're friends with the founder and we actually use the service and it's, I don't know how to say it. It's really awesome. And it's free. It basically takes so much grudge work or grunt work out of your day.

If you're researching stocks, you think about a service like Koi Fin or something like that, where you can research all the financials, you get that. Plus you get the company specific KPIs, like you mentioned it already, the Netflix, the Spotify subscriber counts, which can take so much time to just research yourself. It's all centralized. It's really pretty sweet. So really recommend checking it out, stratosphere.io. And if you're looking for a paid plan, you get 15% off with that code CCM. But without further ado, here's our interview with Sean Wang. Welcome to Chit Chat Money. On this show, hosts Ryan Henderson and Brett Schaefer interview industry experts and riff on the world of investing.

As a quick reminder, Chit Chat Money is a CCM Media Group podcast. Ryan and Brett are also general partners at Arch Capital, and Arch Capital may have positions in the securities discussed in this podcast. Anything discussed on Chit Chat Money by Ryan or Brett or any other podcast guest is not formal advice or recommendation. Now, please enjoy this episode. All right. Today, we are welcomed in by, or we are joined by, Sean Wang.

He's a first-time guest, and I believe he found us by listening to the show. And for our recurring listeners, you know that we talk about cloud all the time, cloud computing, but we're novices. And Sean said, I can provide some helpful context on the cloud industry. And we knew that would hopefully be a great episode for listeners. And just a little bit about Sean. He's the head of developer experience at Airbyte. He's also the author of the Coding Career Handbook. So if you like this interview and you want to read up on anything else, feel free to check that stuff out. We'll link the handbook in the bio. But before we start, I do want to say, I personally am really looking forward to this because I think cloud is a space that, and I'm sure you get these comments all the time, it's an industry that tons of investors are excited about just because of the economic characteristics and that kind of the growth at scale that they're seeing from Amazon and a lot of the other big tech companies as well in the industry. But it's kind of one that you just vaguely know, myself included.

we know we understand the concept but not quite the ins and outs so hopefully this interview can provide some additional context but before we get into that let's talk about you what's your background and how did you i guess what are you doing today yeah um probably my background is more relevant here so thanks for having me on that yeah so i literally i i was a i've been listening for a few months uh and then i heard you guys struggling with cloud stuff and i was like I know cloud stuff. I can come on and chat with you. So my background is useful here because I used to be in finance. I used to be in sales and trading on currency derivatives.

And then I switched over to the buy side, trading TMT stocks. So Tech Media Telecom in a hedge fund. I was in Ballyasney for a couple of years. And over there, I, you know, it was kind of quantum mental is what they call it, which is what makes a qualitative of quantitative investing. and we covered a bunch of cloud stuff similarly, right? This was in 2014, 2015. And there was so many uncertainties from the financial side because like I could break down a financial statement

and I could talk about like projections and stuff like that. But really when it came down to the underlying technology, I didn't know a router from like a switch or whatever, right? Like it was just so opaque to me and I felt very uncomfortable with that. And I think my key learning, I think it was 2015. There was a small bout. I don't know if either of you guys were covering

Amazon at the time, but there's a small bout of worry with regards to the economy. We'll be going into recession. We now know that we were not, but back then it was one of those moments where it kind of felt like the end of 2021 where everyone was like, okay, this is the peak. This is it. You don't actually know. And there was a thesis. We were long Amazon and we actually sold our Amazon position because there was a thesis that Amazon AWS division would suffer because it was overexposed to startups. And in 2015, 2016 played out and it turned out that it was complete nonsense and Amazon compounded quite a bit and we fell behind in our benchmark because of that. And I think that spurred me a lot to realize I didn't know what the heck I was

doing. I was talking about talking points that the rest of the street was talking about and had no fundamental understanding of the business. And probably that was going to be important if I wanted to understand this completely. So I switched careers completely. I went from finance to tech. I went into a coding bootcamp and I came out of it as a software engineer and worked my way up. And so, yeah, that's a little bit of my background. I can talk about what I'm doing today, but I just saw Brett on mute. So did you have a question? I was warming up for the next question. If you have anything else, though. I'll give you a bit. I'll fill out the background a little bit. So left

Belyasny, went to Bootcamp, and then was a software engineer for a bit, and then joined Netlify, which is a second layer cloud, which is something we'll talk about, built on top of AWS, and then left Netlify to join AWS itself, worked there for a year, and now moved into data engineering, which is my current role at Airbyte. So I can talk about the breadth of the cloud services and the startup ecosystem, the small players and the big players. I don't have specific numbers in mind because I'm not an industry analyst, but I can talk about it from the developer point of view and answer any questions you may have. All right. Yeah. We're going to move right into that. We're going to get into the details. We're

going to hit the individual big threes, and then we're going to hit stuff like Cloudflare, Snowflake, and then the second layer stuff, like you mentioned, and a few other things. But first, I think just as an overview, some people might not even understand, you know, what's driving or who's even spending money in the cloud today. What does the cloud market look like in late 2022? Geez. See, that's one of those industry analysts questions that I'm that I'm like, you know, I would just refer you to one of those reports. At this point, you know, I think at the high level, if you talk about like Fortune 500 type companies, the CIOs and CTOs, they will talk about things in terms of digital transformation, right? What percent of their workload is in the cloud versus on their own premises? And I think the numbers are somewhere between 30 to 50 percent of all computing and all IT budget is in the cloud.

So there is still significant, there's been a lot of progress made, but then there's still some ways to go. Probably we'll never reach 100%, but maybe the steady state of things would be 70 to 80%, which means there's another doubling from here. I don't, these are numbers

I'm just kind of roughly ballparking for you. I don't actually know the full numbers and obviously individual sources will differ in their statistics. I do think that probably the best way to identify all these sort of very, very high level macro things is to look at the individual cloud keynotes.

So AWS re-invents Google's and Azure's events, respectively, they'll talk a little bit about their runway. Famously, Jeff Bezos was like, this is TAM unconstrained. But I mean, that's not realistic. Like there is a TAM, it's just very big. So who knows?

It's hundreds of billions of dollars. Gotcha. And I think the growth has been phenomenal over the last five years. And I think for context today, Well, listeners would be, you know, it'd be helpful to understand what drove growth over the last five years. And what do you think will drive growth, say from, we're not talking about maybe just revenue or whatever, just from a usage perspective as someone that works in industry today, what will drive growth over the next five? And do you think maybe it's the same exact thing? okay interesting um so again this is another sort of industry-wide thing that like i i would earmark myself as not an expert um the last three years have been basically determined by covid

right uh and then there's sort of two years before that uh fortunately that's exactly the pretty much the amount of time i've been in the industry i would say that a big part of growth has been driven by essentially the programmatic configuration of cloud. You used to have to go, if you think about the early days of Amazon, think about the earlier clouds like Rackspace or DigitalOcean or anything that's sort of older. You would have to go in and sort of configure things yourself to SSH into terminals and then set them up by running terminal commands.

This would be a lot of the job of a sysadmin. That has essentially upgraded into what the profession is now called DevOps, where it's like a developer-driven operational engineer. And you would provision large quantities of cloud resources by running Terraform scripts or configuring Kubernetes. Each of these technologies have risen to essentially control vast swarms of cloud resources on your behalf. And I think that's really scary for some people because it's very hard for one person to understand all of it. And the cloud diagrams are ridiculous. But I think it's pretty optimistic for the capabilities of the individual developer because I can control so much more than an individual machine. If you think about the sort of alternative to cloud, and this is like how people got started with their IT stuff in the 90s.

Like you would have to buy machines, you have to rack and stack them in a data center. and the primary purpose of cloud is to get to eliminate all these right that they would run the data center for you and you would sort of just buy them off the shelf by clicking some buttons and you would have it spun up for you so i think going from that going from sort of clicking buttons in the ui which is what we call click ops to infrastructure as code which is programmatically provisioning vast quantities of these and managing them in a very deterministic fashion that we understand exactly what's going on. That has been a big driver for cloud in the past five years. I would also say something that I've been personally involved in is essentially

the evolution of second layer clouds, like Netlify, the company I used to work at, Vercel, and there's a bunch of others. There's Railway, there's a whole ecosystem of verticalized clouds on top of the big three clouds, which are providing horizontal services as well. So I don't know if I mean, and then so that's the macro sort of fundamental trends that is going on, the evolution of DevOps. It was accelerated by COVID, obviously, because everyone needed to move online, not because developers were basically working online anyway. Like I was remote before COVID. And so like my life didn't change, which is a little bit worrying to me. But I think just the sheer amount of e-commerce, for example, that moves online, the sheer amount of SaaS services that people started paying for. You know, if you think about this, the scaling needs of something like Clubhouse or Zoom that suddenly overnight became business critical stuff.

These were only made possible because in the past three to four years, we actually had all the technology advances made to increase the provisioning of cloud in a very quick fashion. So another thing I like to bring out now that ChatGPT has happened is ChatGPT scaled from zero to one million users in five days. And that sort of thing typically doesn't happen without programmatic cloud provisioning. Okay. I've got like a million follow-ups, but I think I'm going to start with, I'll start with, I guess, the legacy system or what it used to be, and maybe still is to some extent today. Are there any advantages from a startup's perspective of going with the traditional model of, I think the term you used was rack and stack or on-premise servers, as opposed to- Bare metal is another word. yeah yeah are there any advantages to that or is it just too costly cost uh well cost and control right um so uh i have friends who work at equinix and uh and packet and uh other bare metal providers they are very specialized in in terms of uh those those costs and um i think basically

whenever you make these kinds of clouds cloud decisions you can always go down one layer of abstraction in order to gain more control and to decrease costs as well. But there are trade-offs for that. In exchange for that, you do take on more maintenance. You would have to have the people who know how to do that on staff. And when things go wrong, you can't blame anyone else. You have to go fix it yourself, that sort of thing. And to some extent in enterprise SaaS, which is most of the businesses that I'm involved in, there is an acceptable level of failure where if AWS is down, you can just blame it on AWS and no one will blame you. Whereas if you took over your own data centers and your data centers catch fire or go down, it's on you. And there's a little bit

of trade off there. I will also offer up the idea that specialized cloud services do exist and are valid. So for example, right now in the AI field, there is a nuclear arms race for NVIDIA A100 your chips. And all those are essentially GPU server farms that are independently run because Amazon just would not invest in that itself. You want to control that yourself, right? So that amount of control, specialized compute, it's not a cost decision. It's just simply those general big three cloud services don't have your use case in mind. And there's also with that, security is important. I don't know if you were mentioning that in kind of a... Security is important. There's a very famous Google Cloud video by Stephanie Wong, their head developer advocate, of like five layers of security to just get into a data center in Google, which is pretty fascinating if you want to check that out.

All right. Another question. You mentioned, I believe the term was second layer cloud. Can you describe what that is just broadly and then maybe some of the examples, like an example use case? Yeah. So a cloud built on top of clouds. And it is, it sounds ridiculous until you try to use Amazon and you realize that you need to go to, you know, Amazon cloud school for nine months in order to get something productive done. Or you could use a service that's built on top of Amazon that is specialized for your use case. So I came from a background of primarily JavaScript developers. So let's say overall population of developers in the world is about 70, 80 million developers. About 30, 40 million of them are JavaScript. Another 30, 40 million are Python. And the rest are a fairly long tail of C Sharp and Swift and what have you.

So JavaScript is a huge chunk of developers. And these guys do not know cloud. They do not have the patience to go learn AWS cloud formation. They don't have the patience to wrangle a bunch of different things. So essentially, what these second layer clouds are, are cloud distributions that predefined, preset number of things. I will pick five of these things for you and set them up in a way such that you don't have to think about them. You as a non-specialist in cloud can actually use them. And I think that is basically emerging as a huge trend. I mean, emerging is understating it because this was an investment thesis like five years ago that has largely played out. So, I mean, it's pretty much validated. And so Vercel, Netlify, Begin, Railway, these are all second-tier clouds,

all built on top of AWS or Azure or GCP to verticalize to serve a certain segment of the market that is underserved by definition because the big clouds just cannot optimize for those things. And so if you build a better developer experience, they will migrate to you. And essentially they become aggregators of that audience

for the first layer clouds. And I think it's a nice symbiotic relationship, but it's also a little bit uncomfortable because at some point, If you want to gain more control over your margins, then yes, you do build your own data centers and you start leaving Amazon. But that is intended to be an abstraction layer and most companies haven't had to reach there. I mean, it's essentially getting to the point where like the AWS and GCPs and Azure's of the world, the first layer clouds become commodity providers of infrastructure. And there's pretty much no reason why you would ever try to compete with them. You should try to compete on user experience, on developer experience, on features, on things that they care about, rather than the sheer operation of machines, which the AWS data center people are perfectly capable of doing.

does the second layer cloud change the this might not be something you know but since you worked in the industry you might does it change the unit economics for the big cloud providers are they getting say lower margin do you believe um or what's the revenue share like for a second layer cloud versus aws so as far as it is as i'm aware aws is concerned uh it was doesn't care at all these these these companies appear like regular customers to them they just they're basically aggregators of demand right like uh they're they're sort of resellers uh it's almost all it's almost like a drop shipper of uh of compute services um as far as they as far as they're concerned obviously you know the second layer clouds are very incentivized to provide

value-added services on top of them and they do um and so it's intended to be higher margin but lower volume right and you sort of try to make it up from there um uh but i think you know in a in a in a in a weird way it kind of makes sense that there's symbiotic in the sense that like um amazon you know it's focused on on margins focused on large scale whereas the the smaller clouds uh focus on uh higher margins but uh but maybe slightly smaller tam but i mean like i just told you like the time is still pretty huge um the fact that you can build uh multiple unicorns on top of AWS kind of speaks for itself. All right. We're going to talk about each of the big cloud providers individually, but kind of just looking at them as a whole, AWS, Azure,

and GCP account for the majority of the market right now. I believe I saw one of those industry analyst reports that was like 75%. In that ballpark, yeah. why goal there's ibm yeah okay why would that change over the future is there any way that it's not changing there's there's not even worth discussing it's not changing there's so i guess what are the what are the advantages then yeah well why is it so entrenched why is it so entrenched uh for one thing uh big gets bigger in tech that's definitely something that you see a lot, like the sheer network advantages of I serve more customers, therefore I know more customers and I can build more for them because I have more resources because

from the same money that I got, big just tends to get bigger in tech. And being one of the big three, the fact that it feels like an arbitrary cutoff, but I think it's quite a big gap between them and everyone else. That's the default 90% of the time that gets bigger in tech. You have to come up with something really disruptive, and we're going to talk about Cloudflare later on. You have to come up with really something different in order to change those rankings. And the other is Oracle, IBM, Tencent even. No one's talking about them in any sort of capacity as being able to take the others on.

It's pretty much me-too services. If you're in the Oracle ecosystem, if you're in the IBM ecosystem, you use their stuff. But otherwise, there's no reason to use them. And, yeah, I mean, so I do also think that basically developer mindshare leads to market share. And as far as the mindshare goes, it's pretty much those big three. Yeah, it totally makes sense. And we're going to hit Cloudflare. But first, I think it'd be very interesting because from our perspective, and I think a lot of investors' perspectives, they look at AWS, GCP, and Azure, and they are the exact same thing.

so i want to look at each and look yeah i know that they're probably not the exact same thing so i want to look at each and say like why if you're a startup if you're a mid-sized company if you're a legacy company like say walmart trying to transition somewhere we'll start with aws why would someone choose aws say all costs being equal or maybe the cost there's a difference there um over a GCP or an Azure? Yeah, I'll say from the bat, we should probably ignore cost because cost is a very complicated discussion

and probably all of them are approximately equal. There will be some things that are cheaper in the other clouds because AWS surprisingly does charge a premium on some things. But other than that, they're pretty much comparable. So you're really talking about features

and market focus. So AWS was essentially the first cloud There's some claim that GCP might be first, but really pretty much everyone regards AWS as the first cloud. The very strong reasons for adopting them is essentially that they have the deepest talent pool. There are entire industries and a horde of people specialized to help you optimize your AWS cost and to get you set up

and to administer AWS carefully. So just based on the ecosystem of existing partners and people that you can hire and administrate and run these things, AWS just is the default choice for most people, including, by the way, with startups. Because, for example, AWS will do things like if you get accepted into Y Combinator, you get $150K in AWS credits right off the bat. Don't quote me on that exact number. It's in the ballpark of $100K, $150K. And this is the same amount that you get from YC itself. So they give you a lot of incentives to get you started on the cloud. And once you started, you're kind of sticky.

You're not going to migrate just from your cost reasons, particularly if the costs are comparable. So a big reason to choose AWS is just entrenched existing ecosystem. They have a really good startup program. I'll talk about the other companies in a bit. But and they have probably the widest array of offerings. That's a pro and a con because it's extremely confusing to choose between 17 different types of compute on AWS. But the fact is that they have the largest customers. If they're, you know, a lot of times the thinking will go, well, hey, I mean, if Netflix tried to build the cloud and then they gave up and they went back to Amazon, you know, can you do better than that?

And the answer is probably no. And if it's good enough for them, it's good enough for you. What about Azure? Because it seems like they've been gaining share a bit. How would they, why would someone choose them over GCP or AWS? Yeah, so I was listening to you guys, so the episode where I DMed you guys, I heard that you guys were talking about Azure as like the last choice pick because they're more serving legacy companies. Tell us if that is completely wrong. We're dead wrong. So that's not the way that most people would phrase it because everyone serves legacy at some point.

And especially like, it's kind of a goal if you work in infrastructure to serve legacy companies because they have a lot of money. The way that you've talked about it in terms of Azure is that they have the best distribution. They have the Microsoft's ecosystem is extremely strong. The partners, the selling is extremely strong. If you already bought into the rest of the Microsoft stack, you might as well adopt Azure.

It's kind of the way to think about it. That's the one thing. The second thing is now with the OpenAI stuff. They might actually be the preferred partner of choice because they have made a very strategic bet on AI and it is starting to pay off. They invested a billion dollars into OpenAI. They have always been the first cloud to offer credits and a lot of user support when it comes to starting up AI businesses. And so I would say that they are the friend of AI companies right now, like even over Amazon and Google, which is surprising because Google, you know, we'll talk about Google in a bit. But yeah, I would not describe them as serving legacy.

like they serve as a very specific enterprise part of the business that maybe like is less visible because they don't have like the Netflix's of the world. But they actually do quite well there. I think they're in number two position to put numbers. I think I have numbers actually. So AWS has, cause I pulled this up. I only have this, this one number. AWS is 62 billion in revenue. Azure is 52 billion and GCP is at 28 billion. So a firm third place there. yeah and azure they don't even it's so hard like they don't report that on a um i don't know there's there's they're they're a little coy with reporting that number every quarter but yeah that's all of them are yeah aws put off um breaking out uh aws numbers for for many years and i think that

i so i remember i was interviewing for a hedge fund job the day that they stood it out and like the stock itself jumped five percent which for for amazon size is pretty significant yeah no i i've uh i've i've read about that day being uh kind of a crazy one for the industry when they finally broke out the numbers and the margins you know it's a big deal yeah i mean like we kind of knew it was a big deal but like we didn't know it's like profitable we didn't know like how big it was um and yeah you've got you guys have all had that discussion already so i will repeat it but uh so so i think azure you know um it's kind of a weird thing to say that distribution matters but if distribution matters and microsoft is really really good at it because

They also happen to sell, you know, Windows and Office and LinkedIn and, you know, whatever else have you. And they do like to bundle that and they do like to make sure that if you're in the Microsoft ecosystem, you adopt the entire Microsoft stack, including, by the way, programming languages. So if you are in the C-sharp stack, that is a that is you are automatically captive to Microsoft, which is huge. C-sharp, like most people, like people of my kind will never touch them, but we respect them in a sense that it is a very productive stack. It is completely designed end-to-end by Microsoft. And I think it's doing very well in its own way. Gotcha. All right. And last one, third place, I think they have about 10% market share would be GCP, Google Cloud. Why would someone choose them over AWS and Azure? And you had a follow-up when I sent this over about them having, you know,

creating kubernetes and transformers and how that gave them maybe an advantage so what what's holding them back i guess yeah being as big as the other guys so before this show i was commenting to you that um google has had two very big i mean and this is not exhaustive just just two of the big uh innovations in tech in the past 10 years which is one is just kubernetes they they could they which is an external version of the internal version that they develop for themselves called Borg. And that is the behemoth cloud management runtime that essentially has won the cloud orchestration wars. And anytime you see like KubeCon, CloudNativeCon, any of those words that come up, that is a sign that they're part of the Kubernetes ecosystem. And by now, most

companies, especially if you're infrastructure companies, you run on Kubernetes. So Google released this and then failed to capture market share despite being the creators of Kubernetes, despite being the first consumers of Borg. And that's a shame. I would say that Amazon Kubernetes service and Azure Kubernetes service are probably bigger than GCP, which is sad. And then Google also invented the transformer in 2017, which is the core architecture for all the large language models and now azure is the ai company um they're really missing the ball there uh large language models that's the open ai stuff right and all the stuff that the chat box with yeah aligned with uh microsoft yes gotcha just wanted to have

that but continue yeah absolutely yeah yeah so uh so uh uh quite a bit like look like i don't know the people involved i'm sure they're very smart very very well very capable uh but there's uh you There's a perspective that it's been mismanaged. They had some churn in the CEO. I forget, was it Diane something? It was a former CEO of one of the companies that they acquired. She was in and out for like three years, and it was just kind of mismanaged from the top. I also wanted to cover briefly Steve Yegge, who is a very prominent developer, early Amazon and early Google.

So has seen AWS and then seen GCP form from the very beginning. He has a very famous platforms rant that I would refer everyone to. There was an original one and then there was a follow-up that people should check out that describes why GCP is in third place, that they have not been able

to maintain their products for developers in a way that they would trust. So they would randomly deprecate things overnight. So the thing that you build your entire business on, they'll send you a notice and say like, oh, we have a new version of this. The thing that you're using is going to be deprecated at the end of the year.

Please move. And unscheduled IT spends is never a good thing with a platform that you want to build on. So they have this reputation of, you know, discontinuing stuff. Yes, in the consumer division, obviously the Stadia and like, you know, the 10 different chat apps that they have. Consumer division is very different from GCP, right? So the GCP people will tell you like,

yeah, we're a very different org than the rest of Google. But they have this stain and reputation and it's something that they're working very hard to overcome. So if you ever want to talk to someone from GCP, I think their head of marketing, I forget their name, their head of marketing recently came from Amazon and is trying to figure out how do we fix the image. Obviously, Google is itself very good at cloud. They don't get platforms is the assertion there. And it's something that they have to overcome. All right.

Last one, I guess, on the individual cloud players, Cloudflare. And we did that interview a while back. discussing the business, but people may not have listened to that. So could you talk about, I guess, what Cloudflare is? And then you mentioned that they're beating AWS at their own game. What do you mean by that? So that's a little bit of a provocative title. Obviously, AWS is something like something in the order of 50 to 60 times the Cloudflare size. So it's kind of a pin in the neck for them. So Cloudflare is a, first and foremost, a content delivery network, which is, I actually went back and listened to your interview previously, and I thought it was very well

done. So it competes with Akamai, it competes with other CDNs I can't name, because there's so many of them. AWS itself has a CDN called CloudFront. And there are some market share statistics out there. I would say CloudFront and CloudFlare have about the same market position, both sort of tied for number one. And so what a CDN does is it accelerates delivery of your content. Whenever you are requesting data for watching your videos

or reading something from a website, downloading any image, it doesn't necessarily come from an origin server, which is, let's say, for most of Amazon, it would come from US East 1, which is in Virginia, some data center in Virginia. Like the concept of a physical location for your data, for your code, for anything that's being run or downloaded goes away with the CDN because then it would be distributed to the edge.

And Cloudflare has something like on the order of 10,000 points of presence around the world. And that is about a couple of orders of magnitude more than the points of presence that Amazon has. And any CDN provider would basically provide something like this. that caches your stuff at the edge, has a cache expiry, basically accelerates the speed of your experience. And that's how they started out. I would say that the other thing that they provide is also security services. So Cloudflare's innovative approach to the market was essentially to take a firewall

that most people would charge for and give it away for free. Just give it away. And that was extremely disruptive because everyone else was charging for it. And they were like, Nope, we'll do it for you for free. And I think the thing that people don't understand is that there's a huge growth. Sorry, there's a huge gain in value when you do things at scale. And for them giving away for free gave them the most data about DDoS attacks, which is something that you guys covered in your previous interview as well. And by that point, making them Cloudflare become the world experts in DDoS. And so it's come to a point where I've heard startup founders themselves just go like, just put Cloudflare in front of it. Like it always solves DDoS. And whenever any founders, I've heard a founder basically say like they are the sort of pain button.

Whenever they're experiencing a pretty bad security attack, there's a button that you can push for premium Cloudflare support. You pay them like 60 grand or something, they'll come in and fix it for you. And they are extremely capable of that. And so it's very interesting and disruptive. So from the security side, from the content side, that's good. By the way, for those who don't know cloud, there's sort of three main categories of monetization, right? There's compute, storage, and network. So primarily, they are monetizing network, which is how much data goes back and forth

over your network, or bandwidth, as you call it. Compute and storage are less important for Cloudflare, but that's something that they're going into. So probably their newest, hottest thing is essentially the Edge Worker platform, or Cloudflare Workers is what they call it. So Cloudflare Workers launched as a smart edge functionality where you could run compute at the edge. And this would be good for things like redirects and authentication and rendering and a bunch of other smaller use cases.

But that is now increasingly growing because they've also added storage solutions. So now they have the full suite of compute, storage, and bandwidth, except that they've done it in a decentralized fashion compared to AWS. AWS. And it's pretty interesting because they are now the technology leaders in edge working to the point where they have a three times cost advantage to AWS Lambda at edge, like the equivalent competing solution from AWS. And so most people would pick them over AWS at this point. And so my blog post on this, when I realized this, I was like, holy crap, this is huge. I called it eating AWS from the outside in, right? You do not go after AWS itself to compete with AWS, you surround AWS with all your services. I would run my business on AWS, but on top of it,

I would add Cloudflare for the DDoS protection, for the caching speed. And then I'll move some of my compute and storage over to the Edge Workers platform. And then pretty much, I don't need AWS anymore because I've just moved increasingly more and more pieces of my business to Cloudflare. And that's essentially the strategy to becoming the fourth cloud. So that's pretty a bulk bull pitch on a cloud flare. Although again, none of this is an investing advice. Everyone should do their own research and go listen to that. We did. I think it was a couple hours.

We did a two-part interview with Joe from over at non capital. Yeah. He knows quite well. He also mentioned the R2 thing, which is again, the most bold faced rapper service strategy I've ever seen. So literally, so AWS, I think they're second, the third ever service, the most popular service by far. is AWS S3, Simple Storage Service. And that is the default store of all data on the web.

Anytime you upload a file, you look at the URL. If it says S3 on it, that's AWS. Cloudflare is saying, right, we'll wrap around you like every service and we'll reduce your bandwidth cost to zero because AWS charges for all downloads. We'll just say, Cloudflare just like, we'll wrap it for you and make the downloads free. And that's something that Cloudflare has just pursued for years

and AWS just has not. AWS is not interested in being friends of any other cloud providers, whereas Cloudflare has a bandwidth alliance that they've established where they have peering agreements that make data transfer free for everyone who's not naming AWS, which is extremely, extremely aggressive. But it's working. Gotcha. And so I think that leads into this next part pretty smoothly. People talk about multi-cloud. You know, Snowflake is the biggest one. I think that was a hot topic.

no no no those those are different uh okay well correct i guess correct me if i'm wrong then or define multi-cloud i guess because maybe we don't understand it but exactly does that turn the industry into a commodity uh and why or why not um no it does not does the turning the industry is turning itself into a commodity it doesn't need to help a multi-cloud to do that just because of race to the bottom of prices um And so multi-cloud, the reason I had such an instinctive no when you mentioned Snowflake is because that's not usually the way that people talk about multi-cloud. The typical way that people talk about multi-cloud is if, for example, you are using both Azure and Amazon in the same company for different services, different features that they have. So typically, that seems like a waste.

That seems like you want to centralize your budget. But realistically, each of these companies would have different offerings that might appeal to different parts of your company. And who are you as Mr. CTO or Mrs. CIO to dictate what I use, right? I, as the developer, I'm the expert. And I should make the best technical choices that serves my product and matches my expertise, right? So there's this back and forth between the top down decision making in tech versus the bottom up decision making that's going on.

And I think a lot of companies rightly are just saying, well, I don't I don't care. Like, you know, I just managed to spend. But, you know, as long as I as long as my developers are happy and productive, I don't actually care what tech you use. And that's probably the right approach. But there is something to be said about centralizing costs in a way that you can negotiate at a company wide level. Right. So I think there's a back and forth between the bottom up and top down thing. There are multi-cloud companies that are targeted at serving this market. You know, probably one of the more interesting one is HashiCorp. And one of the ways I always talk about the multi-cloud strategy is most people are multi-cloud by accident. Like they're not multi-cloud by choice.

Usually, for example, when it's decided at the corporate level to make an acquisition, Like a large company buys a medium-sized company. And then at the integration stage, when all the acquisition details have worked out, they discover they're working on different clouds. Do you force a rewrite? No, that's stupid. Like that's a huge cost expense

that you don't have to avoid. So you can just avoid it. So most companies are essentially multi-cloud by accident and that is completely fine. What are the feature offerings from the different providers? Is there a lot of overlap? Do they do a lot of them like already provide a lot of the same features and it's more just like one's a better product or are they more like specialized?

Yeah, there's a long tail. So there is, yes, there is a lot of overlap, but all you need is one feature you can't get from someone and you're just going to go after that one. Right. Particularly if it's some AI thing from Azure or some Kubernetes thing from GCP that Amazon just doesn't have, you're going to go for it because, you know, Amazon doesn't offer it. You're going to go for it. But yeah, like I said, right, like most of the universe can be consolidated into, you know, 17 different types of compute, 17 different types of storage, 17 different types of networking. and there's the Amazon Basics version of this, there's the Google Basics version of this. They all have like slightly different takes on it. And companies like HashiCorp

are helping to commoditize between them, but not in a way that is seriously damaging because all they do is help pave over the APIs. There will be Terraform providers that you can sort of adapt to each of these. So I don't think that is the main concern. Everyone's trying to pursue sort of the value-added services

and the big three clouds are no special in that, right? All of them are trying to pursue like different ways in which they can capture your mindshare so you stay within a particular cloud. I will highlight Amazon again as a cloud that particularly incentivizes you to stay within their cloud by charging what they call the data egress fees, right?

So you can check in, but you can't check out. Or if you want to check out, they'll charge you for it. So if you want to operate your workloads across clouds, You're going to have to pay it for the bandwidth if you want to do that. Or you can take the Amazon Basics version of it and stay within the cloud and not get charged for your bandwidth. So there's all these sort of incentives for you to centralize your workloads. But if there's a feature that you just plain do not have, then people will go for that. Where does Snowflake fit into all this?

Exactly. So why I kind of reject the concept of Snowflake being part of the multi-cloud story is that Snowflake is an independent data cloud, right, that has really carved out a niche for itself. Probably, you know, in the discussion earlier that we have about talking about driving growth over the past five years, that's one thing I missed. I'll go back and add that. that the growth in data warehousing in general is a huge, huge, huge driver of IT spend and has changed my own career as well. I used to be on the compute side and networking side

and now I'm in data. And another data point I can offer you is that AWS Redshift, which is the Amazon in-house version of Snowflake, is the fastest growing Amazon service ever. and since it was introduced 10 years ago. So there's a lot of pointers towards data being a central force in how people run companies.

I mean, it's kind of a meme to say data is the new oil, but data is the new oil. And Snowflake has really carved out a niche for itself as the best data warehouse in the space, in the space filled with, Google has BigQuery, which is probably number two. Azure has Synapse, which is a distant number four, number five, and Amazon has Redshift. So Snowflake is in that mix, but has somehow managed to carve itself out as the best independent data cloud, which makes it play really nice with the other clouds because it has all its own independent infrastructure.

But it's a really interesting non-consensus view that someone who is not AWS or GCP or Azure can come from nowhere and just completely dominate. I think at their peak, they were like a $120 billion business. Gotcha. And this might be, tell me if this is a dumb question, but are the cloud providers trying to vertically integrate and provide their own software tools? because I kind of see, you know, sometimes maybe Redshift is the wrong one, but you see a lot of them, you know, come up with a lot of products that maybe they're not purely software, but they're trying to compete with, say, I guess an example I know of is a MongoDB. Do you think that this strategy will work in the long run?

Or I guess it's a hard question, but any thoughts on the vertical integration? Between Snowflake and Mongo? No, it's just, yeah, sorry, I should have. For, say, the, I guess, you know, AWS, GCP and Azure, you might have, I forget GCPs, it's BigQuery, or, you know, some of the other stuff that might be, it might be software, it's, I guess it's not pure software, but the, you know, they might have their own tools that compete with some of the products that are the customers of them. do you think that strategy can work or is it kind of it seems like there's going to be a lot of conflict is what i'm saying it seems like it but like just this sheer existence of snowflake is actually giving me a lot of hope that you can uh build independent clouds you don't have to be a second layer cloud uh which is exciting uh it means there there is still white space left in in the cloud ecosystem uh but like you have to be a snowflake and i feel like they are they're

almost like the exception to the norm. So I don't know how much to generalize from that one example. You did mention MongoDB as an alternative data company, but I would not place them in the same class as Snowflake. MongoDB just happens to be a very successful database company with a huge licensing business that is only slowly becoming its own independent cloud business.

But yeah, most database companies do not reach the scale of cloud business that Snowflake has. So they've really carved out a niche for themselves as providers of warehousing and particularly all the services that come with warehousing, including compute. Okay. And we talked about this next question before we recorded. So as generalists, I guess, or anyone that's not in the industry, which I think a lot of, you know, what's a lot of our industry as investors, we think of the cloud as just, all right, you took your servers, you had them at your office, and then you moved them to whatever, the place in Oregon on the West Coast, the cloud. You moved it to Virginia, right? And AWS does all the work. Is there anything that is important, I guess, maybe from investment or product perspective that generalists like ourselves, investors, analysts on Wall Street misunderstand about the cloud today that's very important? Yeah.

So I don't talk to journalists all day long. So it's not like I have a huge grasp of where people are off. I would say do not underestimate the change in the way that people are using cloud within the cloud category itself, because there's very huge structural changes in how that works. So my favorite statistic coming from inside Amazon is that Amazon itself used to run on EC2, which is Elastic Compute Cloud, which is pretty much the first compute service that was launched in Amazon. today 50 or more than 50 percent of amazon.com's internal services that are internal customers of aws 50 of them use aws lambda which is the serverless equivalent of ec2 and that has shifted dramatically towards more flexible computing faster computing more scalable computing as well and so that is going to drive sort of the next wave in cloud adoption in cloud

spends in in volume and i think to miss that structural shift is going you're you're going to miss um you know changes in in who's who's leading who and who's uh falling behind um the role of data is something that um i i would particularly call out as well like the um who's winning and uh falling behind in the data engineering space um which is uh has led to my choice of uh company to be at um and then maybe uh you know second layer clouds as well the the The concept that the big three have become so dominant that no one's trying to compete with them. Everyone's trying to compete on value-added stuff

by verticalizing on top of the first layer clouds. Okay, tailoring off that then, if we were talking on, all right, what's today? December 19th, 2022. If we're talking on December 19th, 2027, what does the industry look like then? How do you think the next five years kind of work out? is it still going to be largely dominated by the big three in terms of market

share? Or do you think, I think you, the term you said was eating them from the outside in, do you think more workload or spend is going to be with those other ancillary or add value added companies? Yeah. That's interesting. Five years. So, you know, I think, I think you covered this as well in the, in previous podcasts, the existing commitments to AWS is very strong and the existing clouds.

The way that you do cloud spend is you lock people in for three to five years, right? Like, so pretty much people like Andy Jassy, who runs Amazon now already knows what earnings are going to be in five years. So the only thing that's not determined in the, in the five-year window is how much the others will grow. And it's not as, it's not a fixed pie, you know, We're also dealing with a moving target.

And it's very typical in growing industries like this that the incumbent players will grow, but at a slower pace, and the up-and-coming players will take up the majority of the growth. And so as an investor, you're like, where should you put your money? Should you put your money in the slower-growing but bigger companies, or should you put your money in the faster-growing ones? And that's obviously a factor of interest rates and growth and what have you. um so typically like it's funny to say that five years is short term but it kind of is because the the the bigger changes will be seen over the 10 to 20 year

range but that's that's the order of that's how downstream uh the financials are from the tech right like the tech is so far ahead that you have you kind of have to go like all right on principle this is a fundamentally better technology therefore this will win in a 10 to 20 yearish timeframe, which is a really, really hard thing to try to interpret, which is why I'm no longer in active public market investing, by the way. But I love just getting deep into the fundamentals of the tech and trying to determine that. So yes, I do think that Cloudflare does have a very good shot. I don't know if it's within a five-year window, but they have, from a fundamental basis, constructed a market position for themselves where AWS cannot fight back. That is a very,

very good place to be in. Snowflake has done that. Databricks has done that, by the way. Databricks is an ML version of Snowflake. And there are a bunch of these other smaller companies. Every company that you guys are dealing with, if the market cap was a billion dollars, it'd be too small for you. If it's in the $5 billion, it'd be too small for you. And that's a struggle that I had in my hedge fund. But there are so many of us in the startup and tech ecosystem where we deal in the 10 to $100 million market cap range, like valuation range, where we're betting on the future of tech, that will be the dominant players in the distant future. So I don't know how to phrase that to you apart from like, I feel like if you're looking for investing opportunities,

the world is probably wider than you think once you look outside of public markets. Beautiful. Yeah. And so it would be very surprising to see the cloud market from a spending perspective. It doesn't matter who has the market share. That's, you know, That's a little bit of a harder question. It'd be hard to see it not being, say, 50%, 100% larger five years from now, just because you have those existing long-term contracts. Yes. Yeah.

These things are megatrends, right? They are existing, like the DoD moving over to cloud, the fact that scale just locked in a 10-year contract for $2 billion. dollars like it's these kinds of things take uh take on the order of of a decade to play out uh and they've already been predetermined today like we we know they're going to happen they're just going to take some time to move over um and i mean that's encouraging and some expect but also like as investors you're looking for volatility right you want the 10 bagger in one year maybe i don't know but like that's where that's those are the hero trades those are the

things you talk about at bars but really like when you build a business you're you're building for sustainable revenue and predictable growth and like those things don't don't don't happen because that's not how you get the big the big fish right the big fish hunt slowly and then there's a lot of them to go around once once they've been hunted uh so so i don't know how to how to advise uh on that um i will say you know i i think maybe in in five years given the trajectory of AI progress and spend, the sort of dark horse in this would be Azure overtaking AWS purely on AI. That is massively projecting from today's hype.

So I'm definitely talking, obviously, after every hype, there's a disappointment phase. So who knows? But if any were to win or to significantly disrupt the current path of things, it is that AI is a much bigger deal and much bigger, broader opportunity than most people expect today. And therefore, you should invest in any company that is much further along in the AI space. And I'd probably say that Azure is much further along. Okay, I think that's all the questions we have, Brad. Yeah, he's giving me the thumbs up here. um shot this is this is a lot of fun where if listeners want to like keep up with you what's

the best place to do that god i used to say twitter but like now twitter is a complete like chaos and uh i i have no idea um so you can find me on twitter um or you can find me on my site six.io um i do have a couple of uh blogs one on developer tooling and then one on ai um which is something i'm tracking pretty pretty closely so yeah find me on six.io and it's and that's S-W-Y-X dot I-L? Yeah, those are my initials. And yeah, basically, I love talking about the business implications of tech, but it's hard to make forecasts. That's where the analysts come in.

But I think there's a lot to share from finance people going deeper into tech. And I hope to encourage more people to understand the fundamentals because that is what will last. It's not the quarterly sort of missed at miss the street, you know, beat the street. Like I don't, I no longer care about that. And that is a great benefit to my mental health. Yeah. It doesn't matter if AWS grew 28% versus 30%, like the stock's down 5%, you know, is that a, is that a big deal? I love, I love when people talk about like percent IT spend, like,

like the really, really big needle moving things. Those, those are the things that i wish that more people um looked after and i i you know if i if i could go back to my former finance analyst days i would have asked i would have tried to get into a vc or long-only funds that invested that way instead of what i ended up with which is essentially a chop shop like we we had a portfolio turnover of 12 times which means that once a month we turned over our entire portfolio uh which is which is uh very very stressful yeah that is uh And it's hard to, I think, focus on bigger trends in longer term when you're turning it over that much. But I think that's going to do it. Thank you, Sean, for doing this. We should throw a disclosure on here. Brett and I are not financial advisors. Anything we say or discuss is not formal advice or a recommendation. We are, however, general partners at Arch Capital, so clients may have positions in the securities discussed on this podcast. Thank you all for tuning in. Thank you again, Sean, for coming on the show. This was a blast and we will see you all next time.

Okay. I'm welcomed by the founder of our exclusive sponsor, Stratosphere.io, Brayden Dennis. Brayden, welcome. I wanted to basically give listeners that are interested in Stratosphere, more context around what the platform is. So let's start there. What is Stratosphere and then why did you decide to start it? Yeah. Thanks for having me. I appreciate it. And I'm glad to be sponsoring the podcast as a listener myself. I like the deep dives. I like the different guests, the different perspectives on some interesting companies. So I think it's a good concept for a podcast, which is kind of what led me down to making Stratosphere in the first which was I was making content online and frustrated with the tools that were available

to me. So I started building a very scrappy version of the product just for free, just to figure out how can I overlay 10 years of financial side-by-side up to 35 years we have now? And how can I actually build out a proper database of company KPIs that are not just revenue, But if you're looking at Costco, how many warehouses do they have? How many paid members are in our Costco members? Or if I want to do a comp against the streaming, how many Netflix subs versus HBO Plus, Discovery Plus, Disney Plus? How do I build up proper comps of those? Because those are the metrics that actually move the business. Those are the ones that actually move the needle more than any gap financial metric you'll find.

And so it started off as just purely a passion project. And I figured, let's just make the leap into entrepreneurship and see where it goes. And, you know, it brought us here today. Yeah. And like you mentioned, it is the stuff that you can't find anywhere else, at least not now. I mean, you could find it page by page on their financials, but you can go through 35 PDF filings and find it, be my guest.

And that's basically what we did for a long time. So what do, I guess, maybe describe the pricing model so people know, but you're going to say there's a free platform. What do free users get? Yeah. Good thing. Because our mission was to always build a free platform. And so we really kept true to our mission and give an amazing platform for free, which gives you 10 years of financial statements on 40,000 global securities. So we don't list you just to US securities, it's on global stocks. We give you a watch list, the screener, comparisons on competitors, fundamental charting up to 10 years, filings, transcripts. You can look at the press releases right inside the app, news, ETFs, funds,

super investors, hedge fund letters, investor holdings, and financial calendars. Those are all the features you'll get on the free tier. Now, on the middle tier, the personal tier, you're going to unlock up to 35 years of financials and just nice to have, quality of life, notifications being built in, price targets for building models, business owner mode where you can hide prices, like kind of like just that next level for individual investors who want to level up. And then the top tier is for like investment teams and professionals who want to unlock that KPI data and request KPI coverage as well. Like a firm will be like, here, we want these 10 names in our coverage and in your coverage. And then you'll have basically our entire universe that

we're looking at, which is great, right? Because like earning season comes around and we have it updated within 15 minutes. When Netflix comes out with their net subscriber ads, it's right there in one place, especially easy to handle around the peak of earnings season. That matters a lot for these people. And so we have a premium tier for that as well. That's the three plans that are available today. And now a perfect time to shameless plug our code. If you use CCM, you get 15% off any of the paid plans, but I think that covers it pretty well. Uh, if you're interested, please go ahead and check out stratosphere.io. We'll, we'll have a link in the description as well, but, uh, thank you, Brayden for joining us.

Ryan, keep it up. I really like what you and Brett are doing and, uh, I'll be listening along.