Episode Show Notes & Transcript
Charles Fitzgerald is a Seattle-based angel investor, with a focus on developer platforms and infrastructure. Previously, he spent 20+ years working on platform businesses at Microsoft and VMware. He can see the cloud from his house.
- Main company site: https://www.platformonomics.com/
- Twitter: https://twitter.com/charlesfitz/
- LinkedIn: https://www.linkedin.com/in/charlesfitz
Announcer: Hello, and welcome to Screaming in the Cloud with your host, Cloud Economist Corey Quinn. This weekly show features conversations with people doing interesting work in the world of cloud, thoughtful commentary on the state of the technical world, and ridiculous titles for which Corey refuses to apologize. This is Screaming in the Cloud.
Corey: This episode is brought to you by Trend Micro Cloud One, a security services platform for organizations building in the cloud. I know you're thinking that that's a mouthful because it is, but what's easier to say? "I'm glad we have Trend Micro Cloud One, a security services platform for organizations building in the cloud" or "Hey, bad news. It's gonna be a few more weeks. I kind of forgot about that security thing." I thought so.
Trend Micro Cloud One is an automated, flexible, all-in-one solution that protects your workflows and containers with cloud native security. Identify and resolve security issues earlier in the pipeline and access your cloud environment sooner, with full visibility, so you can get back to what you do best, which is generally building great applications.
Discover Trend Micro Cloud One, a security services platform for organizations building in the cloud. Whew. At trendmicro.com/screaming.
Corey: Welcome to Screaming in the Cloud. I'm Corey Quinn. I'm joined this week by Charles Fitzgerald, who's a Seattle-based angel investor, historically with a focus on developer platforms and infrastructure, but before that in the depths of prehistory, he spent over two decades working on platform businesses at Microsoft and VMware and, as his bio states, he is very proud that he can see the Cloud from his house. Charles, welcome to the show.
Charles: Thanks for having me.
Corey: Thanks for being had. So, you run the platformonomics.com website, which it’s appropriately named after your company, Platformonomics. What do you do?
Charles: Well, my usual answer to that question is, “Hey, I'm on Twitter,” which leaves people dumbfounded. It does leave me some spare time, and it's probably worth going through my background. I joined Microsoft out of school a long time ago, and basically worked on every platform effort there from 16-bit Windows, to 32-bit Windows, to COM, OLE, ActiveX, served in the trenches during the browser wars and was one of the very first people to work on .NET. So, I've seen a lot of iterations of platform gambits in the marketplace.
Left Microsoft end of 2007, kind of dorked around with a startup, ended up at VMware for about two and a half years, where I helped launch Cloud Foundry, and by about 2013, I was kind of done working for the man. And since then, I have been angel investing primarily in Seattle, my nominal focus is on developer infrastructure and platforms, but done lots of investments in other stuff from FinTech to crypto, to VR, a lot of boring SaaS companies will pretty much look at anything. And during that time, I've also, kind of, couldn't but help myself watch the evolution of the cloud race. I joke I can see the Cloud from my house because I live in Seattle and there are plenty of clouds to be seen. You know, if I twist my head and open the blinds, I could probably see Amazon right now, and continue to have flashbacks from Microsoft. So, I have a pretty good perspective on what's going on in the Cloud. So, to the degree I do anything, that's what I do.
Corey: So, the way that I feel became familiar with your work is when I started my consulting company three and a half years ago it was, “All right, I have a background as an AWS architect/systems engineer type. What expensive problem can I align this around?” And the AWS bill was the easy answer. So, I started doing some research, as some people apparently don't do.
And I was, “What do I call myself? ‘Cloud Economist’ sounds like the right direction to go in because it's two words no one understands.’ ‘cloud’ is a bunch of other people's computers, kind of, and ‘economist’ is someone who claims to know about money, but dresses like a flood victim. Smash it together, and no one knows what the hell I'm talking about.” But when I was doing some research to see, does this actually exist? Platformonomics came across my radar, and it was one of those things where, “Huh, I’d better stay away from this guy. He might actually be able to call me on my nonsense.” So, I was scared of you for the first couple of years until I started reading more of what you've done. And the more I read, the more I like.
Charles: Yeah, familiarity just undermines that intimidation factor completely.
Corey: Oh, absolutely. It's bred an awful lot of contempt. It's been spectacular, but one of the things that I found that you talk about, along with basically nobody else, has been a recurring series that you're calling Follow the CAPEX. What is that?
Charles: CAPEX is capital expenditures. And the origin story, I was—I mean, this was probably 15 years ago, I was still at Microsoft, I was sitting in a meeting where somebody was reviewing Microsoft's efforts to build data centers. And if you grew up in the software business, and ‘grew up’ may be a strong term in my case, CAPEX was, “Hey, we have some buildings, we have some desks, some new laptops.” That was about it. There really was very—it's a very capital-like kind of business.
And so, sitting in this presentation, and listening to people talk about there's concrete, there’s steel, there are vast racks of hardware. This is before the big cloud companies started doing their own transoceanic cables, but I was just fascinated because it was so counter to what a software company had traditionally done. And then in the early days of the Cloud, I'm sitting here watching everybody posture and tell us how great they're going to be, and I would just look at the CAPEX numbers. And the numbers are just, they're obscene. I mean, I have been updating these every year for a few years now, and if you look at Amazon, Microsoft, Google, they've spent almost $350 billion since the year 2000. We've gone from this incredibly asset-light industry to the cloud vendors are now on par with the biggest CAPEX vendors in the world: the energy companies, the telcos.
And that's not all cloud, for sure. Google is mostly search and YouTube. GCP turns out to be a very small tail on what is a very large infrastructure dog. I used to joke that there must be a space elevator in there as well. That was, kind of, the acceptable societal trade-off for accepting the ad surveillance network. Doesn't look like we're going to get that now. If you look at Amazon, and Amazon's numbers have just gone absolutely ballistic in the last couple of years.
They spend on AWS, it's probably a fraction of their overall spend. They spend on robots, and trucks, and planes, and warehouses. And now, they really do have a space program as they start to do this internet satellite constellation. Microsoft has probably the cleanest CAPEX in that they don't have a whole bunch of those other extracurriculars, but it's just an obscene amount of money and turns out to be a huge barrier to entry. I mean, imagine somebody gave you $100 billion and said, “Go try to catch these companies. Good luck.”
Corey: That wouldn't be enough.
Charles: It would take a long time. So—
Corey: All three of these companies had other businesses that financed the build-out of their cloud environments. No one has done a, ‘we're going to build a cloud,’ and succeeded at a hyper-scale way. There's always been something else, be ad sales, e-commerce or giant software ecosystem that has financed this build-out.
Charles: Yeah. You definitely had to have a big high margin, big balance sheet in order to go do that. So, anyway, I'm sitting here watching this stuff, and then you start to look at the rhetoric coming out of different companies and some companies, the CAPEX was up and to the right, and you could be pretty optimistic about them. Some companies were talking a big game, but their CAPEX was actually declining. And at some point, I decided I would separate these different vendors into, we have the ‘clouds’ and the ‘clowns’ and that resulted in a bunch of interesting interactions with some of the clowns.
Corey: Oh, yeah. And one of the hard parts about this too is all right, you're going to throw enormous piles of CAPEX at your cloud buildout. Great; that's not going to pay dividends for three to five years at the earliest. It isn't something that you can flip a switch and you're good to go. It takes a lot of time to build these things.
Charles: I mean, one of the other lenses on these companies is some of them are engineering engineering companies, some of them turn out to be financial engineering companies and either weren't willing to spend the money on concrete and servers and fiber and everything else, or they really didn't have the money. You know, the dividend was far more important to them than investing in the next round. And IBM in particular is the poster child there. I wrote a post in 2013, arguing it doesn't look like IBM was going to make the cloud transition, and the post was pretty well-timed. I think I missed IBM's all-time stock high by about 16 days, and they proceeded to reel off—what was it 24—straight quarters of revenue declines.
But I had all these IBM people coming at me waves, they kind of made three arguments. One, we're IBM; we always make technology transitions. And my response to that was, “Well, you actually have to do something if you want to make that transition.”
Argument number two was, “Hey, Wall Street loves us.” Well, it got to the point where even Wall Street was telling IBM not to focus on the financial goals, they need to focus on their survival.
And then the third argument was, “Well, at least we're not HP,” and I had to give them that. But I've sort of instituted a policy where if you don't say ridiculous things, I'll leave you alone, but if you keep saying ridiculous things, I'm going to call you on it. So, it's led to a very interesting set of interactions with IBM over the years, where they continue to say preposterous things about their relevance to the cloud space. And that's in contrast to somebody like Oracle, where Oracle talked some ridiculous things, but their CAPEX is—I mean, it got to the point where the big cloud guys were spending more on CAPEX in a quarter than Oracle had—or maybe was more in a year than Oracle had in the entire history of the company. But they've kind of stopped saying ridiculous things. Larry occasionally slips out, but they've dialed it way down.
Corey: They hunt him down rather effective. They've gotten much more efficient at it.
Charles: Yeah. You could tell they're there about three straight quarterly earnings calls where they did not mention cloud. Was very clear they'd gotten the discipline about what they were going to say and what they weren't going to say. So, it's turned out to be a super useful lens where you just—who's spending big on data centers, and transoceanic cables and millions and millions of servers, they're probably going to do all right if they can meld that to software, and go to market, and everything else. And the people who are talking a good game and issuing press releases and not spending any money, they may have a couple of closets here and there with a couple of racks in them, but that's the furthest thing from a cloud. So, it turned out to be a great lens. I mean, I karmically shorted IBM; I wish I'd shorted them because they dropped a third of their value and have just been a financial disaster over the last 10 years.
Corey: One of the hard things for me is, as difficult as it is to interpret all of AWS’s 200-some-odd services and put them in their proper place and keep track at all, that is a trivial easy exercise compared to reading their financial statements and making sense of them, just because you want to talk about real engineering versus financial engineering, it feels like they're heavily involved in both where they're building out data centers by financing it through both capital leases as well as these bespoke custom lease instruments that really obscure what's going on and how. It does all kinds of weird things to their free cash flow that really lets them plow money into this at an unprecedented rate. And it's also super hard to disambiguate this across different business units as well, in some cases. So, it’s like, “Is that an office building that they're spending all that money on? Is it data center regions? How does that break down?” And at the end of it, I basically shrug, give up, and leave it to you.
Charles: Don't forget the big biospheres. I mean, the things they're spending money on—I mean, they have a very clear strategy to take all that free cash flow and reinvest it in something, and they've kind of run out of things to invest in. I mean, a satellite constellation may be the last gasp after hundreds of thousands of trucks, and a fleet of airplanes, and distribution centers in every city, and convenience stores on every corner. But yeah, I mean, on one hand, I think they've been smarter about leveraging their balance sheets, they actually had to call out and explain to people that they were using those capital leases because people were not aware of them.
Charles: It was just another complicated footnote in the financial statements. But I'm pretty sure that those capital leases are the server investment behind AWS, if you look at the lifespan, it really correlates to the useful life of the server.
Corey: And they've also been talking on earnings that they've been increasing those lately.
Charles: Yeah, they just bumped it up to four years, and the amount of new capital lease has actually flattened out pretty dramatically, right about the same time. So—
Corey: Yeah, you take a look at their two generations back systems, no one in their right mind is going to be building net new workloads on top of those, but if you look CPU capabilities and do a one-to-one comparison, ah, this is almost definitely what things like Fargate are running on top of and presumably Lambda as well, so repurposing things like that from an engineering perspective where you don't guarantee specific performance characteristics is a great way of doing that, and it's clearly working. Those services are growing like wildfire. It's a brilliant play.
Charles: Yeah. Being able to reuse that stuff and get more value out of it for as long as you can, is obviously super key. One of my projects for the summer—we'll see if I get around to it—is to try to put some scope around just how big AWS is. So, from a how many servers, how many cores, how much storage? I’ve sort of got some ideas on ways to go after that. So, if I get motivated, I'm hoping to tease that apart because AWS is definitely a small part of the overall Amazon CAPEX spend, but I have some theories on how to triangulate it.
Corey: I would be very interested in hearing what you come up with on that front.
Charles: As would I. [laughs].
Corey: One thing that I've noticed among cloud providers, everyone looks for superlatives to some extent. And Microsoft’s strategy, which I can't help but see as a blunder, has been to focus on having the most global regions, and very proud of that fact. They’re trumpeting that during Build this year: Satya Nadella called that out explicitly. And yeah, it's important to have regions near where your customers are, both for performance reasons, as well as for regulatory purposes, but by doing that, rather than focusing on fewer, more robust regions, instead going with a more far-flung approach, the capacity of each one of those regions is not able to absorb sudden influxes, such as, oh, I don't know a global pandemic changes the way that everyone interacts with their cloud services. And there were serious capacity shortfalls for weeks as a direct result of this. That is my analysis of the situation. Do you think that that's spot-on, completely misses the mark, or something else?
Charles: They’ve definitely had some capacity issues, and I think all the big cloud vendors I'm sure are furiously scurrying, given the disruptions on the supply chain side, combined with increases in demand. I think it's a good indication that Microsoft has a different set of priorities than Amazon or Google. Microsoft is much closer to those enterprise customers, it's those regulatory demands where customers say, “For data sovereignty reasons, we've got to store our stuff in the same country.” And Microsoft has been more responsive to that. Contrast that to the Google model, which is far fewer data centers that are of enormous capacity, but don't align well with those sovereignty needs. Microsoft's probably at the other end of that extreme, and AWS is somewhere in between, but all of those companies are going to continue to build out their local capacity over time and hopefully, the pandemic will recede and it'll be easier to do your capacity planning going forward.
Corey: Oh yeah. At some level, I'm reminded, for some reason, of the dot-com bust where everyone was building up in the dot-com bubble, they're convinced that this is going to be the way and the light of the future. And it was, just not for a bunch of those companies, and one thing that we saw happening was the telcos laying enormous quantities of fiber, assuming it was going to go up and up forever, and what they inadvertently wound up doing—from a certain point of view—is investing their own way out of a business model, where suddenly there is so much capacity for high-speed transmission and the rest, long-distance rates plunged to effectively zero. And suddenly you're seeing this whole world of, “Oh. We’re the provider of pipe, and suddenly there's way more pipe than anyone is going to directly need. We can't charge a premium for that anymore.” I'm wondering if we're going to see something like that this time around.
Charles: You didn't have Moore's Law type phenomenon on the telco side, right? I mean, a quality phone call takes about the same amount of bandwidth, and so it's really kind of more Moore’s Law and—
Corey: But it takes way more bandwidth to get a crappy one through a whole bunch of things. WebEx looking at you.
Charles: Yeah. It's been fun dealing with the network issues in the last month or two, I think we're going to continue to see the demand curve is different. The fact that you had lots and lots of fiber didn't result in you making a lot more phone calls. But as compute gets more and more pervasive, people can do more and more things. What we've seen is that demand has continued to go up.
The other thing that's going on with Cloud is we still have a huge amount of substitution to do. I mean, one of the things I'm hearing in the last couple months is a lot of the last holdouts, the people who were very proud of their on-prem data centers and continued to talk about their security concerns with the Cloud, a lot of them have suddenly decided maybe it's time to start to move, and to the degree that sub 10 percent of workloads are in the Cloud or whatever your favorite number for that particular metric is, there's still a phenomenal amount of upside. And I think we're going to continue to find new things to do with compute cycles for the foreseeable future. So, I'm not too worried about that. I think it's a different dynamic.
Corey: I think that it's easy to predict the future after it's already happened.
Corey: It's one of those stories of, “Oh, well, that's what happened last time, so, well, obviously, that's going to happen again.” As they say, “Economists have successfully predicted five out of the last three recessions.” The challenge, of course, is that history doesn't repeat. It rhymes, but it doesn't repeat. What do you think's going to be potentially different this time around?
Charles: Yeah, I've spent a bunch of time thinking about that. And we really want to go back to the models of what happened after the global financial crisis, or what happened after the dot-com bubble collapse. And some of those elements—I mean, I tend to focus on early-stage investment ecosystems, and we've been living in this world for the last five-plus years where zero interest rates have pushed people to look for more yield. It's pushed a lot of money into other asset classes, more speculative asset classes, including venture.
So, we've had all kinds of crazy money floating around. I mean, it's just insane that I'm a little angel investor in Seattle, and I'm running into sovereign wealth funds that are traipsing around in the same spaces that I am. So, one of the big questions is, a lot of this money's been allocated big venture funds, sovereign wealth funds, all the hedge funds, or many of the hedge funds have gotten into venture, you've got corporate venture, what's going to happen to these big pools of money? Are they going to dry up and go away? Are they still allocated?
If you've raised a venture fund, you've got the money, you probably need to go allocate it. And one of the things that I noticed in 2000 is everybody assumed things would dry up immediately. And what we saw is that people continued to invest like it was 1999 for the first year. So, I think we're going to have to wait until some of these investors need to go back and raise their next fund, and I think it's going to be a lot tougher for them to raise it. I think we still are going to see a lot of money sloshing around in the near term.
You continue to see all these headlines about companies raising big rounds, tens of millions of dollar rounds. So, the money's still out there for companies that have the right story. How many of those deals were done, pre-pandemic and are just now getting announced so that people can look like they're still active, I don't know. But figuring out what happens next, which segments, which opportunities face very real headwinds, which pick up tailwinds is a super interesting question, and everybody's trying to figure that out right now. Generally, I'm not a consumer investor.
So, I think I've got probably one company of the set of companies I've invested in that has had real headwinds, and they've done a phenomenal job pivoting, but most of the others companies I've got are either net neutral, or in some cases, they've picked up a tailwind in this environment, which is great to see. But I think it's going to be a long trek to get back to normal. So, we'll see if—you look at the stock market today, and it's kind of oblivious to the idea that there's anything going on in the public health world. To the degree it takes us a really long time to get to the other side of that, things may get a lot more dire.
Corey: In what you might be forgiven for mistaking for a blast from the past, today I want to talk about New Relic. They seem to be a relatively legacy monitoring company, and I would have agreed with that assessment up until relatively recently. But they did something a little out there: they reworked everything. They went open source, they made it so you can monitor your whole stack in one place and, most notably from my perspective, they simplified their pricing into something that is much more affordable for almost everyone. There's even a free tier with one user and 100 gigs per month, totally free. Check it out at newrelic.com.
Corey: Speaking of the healthcare world, which is always something interesting to cover, one announcement at Build this year was that Azure was going to have a offering focused specifically for the healthcare market. If AWS were doing something like that—targeting a specific industry vertical, I would make fun of that until I was blue in the face because it would be ham-fisted and the complete wrong direction to go in because AWS is terrible at this. What stops me from making that assessment about Microsoft doing this is that they have 40 years of history in being phenomenally successful at targeting specific industry niches and verticals. What's your take on that move?
Charles: Yeah, I saw that headline. And I thought, “Oh boy, this is going to be another niche cloud.” IBM and Bank of America announced a financial services cloud and never got beyond the press release. It doesn't look like it's a separate cloud. It just looks like it's a set of solutions that run on top of Azure, which makes perfect sense given that 40 years of experience working with different enterprises, different industries, and they’re probably are lots of things that have a greater urgency in healthcare to get to the Cloud than they did three months ago: telemedicine, drug discovery, whatever it may be, so I think that makes sense.
But I'm surprised we don't see more niche clouds. As I think about where the opportunities are going forward, you've got these, sometimes I describe it as the three big hyper-clouds, sometimes I talk about the two and a half big hyper-clouds—and maybe we can talk about GCP and whether they're serious about this in a minute. But normally, you'd expect to see a power-law distribution. And so we've got these three huge players. And then there's kind of nothing, right?
You've got a few little companies that they’re regional hosters who may be a little bit ambitious in describing themselves as clouds, you've got some people who cater to cost-conscious developers, and that's never a great way to build enough margin to invest. But I would expect to see more niche cloud offerings, and they really just don't exist today. So, it's something that I’m—definitely have my ear to the ground on as I look at what's happening next in the cloud world.
Corey: One thing that is interesting, too, that I see in these global market share analyses is folks are saying that Alibaba has surpassed Google for number three. In fact, they just announced this year a $28 billion investment in their cloud offering. But by everything I can see, all of that is focused on Mainland China and servicing businesses with significant presences there, I tried to sign up for their cloud and give it a test run, and I could not get that far because it wanted me to upload a whole bunch of documentation in order to qualify for a couple hundred bucks in credits to test stuff out, and all I really wanted to do was spin-off a Linux VM for half an hour and play around with it, and kick the tires. It was very clearly not aimed at a business that is not used to the Chinese market. And I'm curious as to see, do you think that that is likely to change? Do you think that there's going to be a breakout of Alibaba Cloud into the broader cloud ecosystem, or is it going to remain a—I hesitate to say ‘niche player’ because the Chinese market is still over a billion people, but I'm curious to hear your take.
Charles: I really don't think we're going to see any new hyper-scale challengers to AWS, Azure, and GCP. They have such huge moats, they have such huge investment, and if you look at companies that have the resources and the skills to do it—maybe an Apple, maybe a Facebook—they really don't have any interest in serving an enterprise customer and you can't play in this space if you're not going to go after the enterprise customer because that's where the majority of the dollars are. So, I don't think we're going to see generally new hyper-cloud competitors. I don't think we're going to see the Chinese hyper-clouds get much traction outside of China or places where China has a lot of sway. This deglobalization trend that was happening prior to the pandemic has really only been accelerated by it.
So, I think we're going to see a world where the big hyper-clouds from the west are one set of vendors, and then we're going to see a different set of vendors inside China. I don't think the Western companies are going to make progress inside of China and vice versa. I tried to dig into Alibaba’s CAPEX spend a couple years ago, and I discovered that if you're a Chinese company and you're listed on American stock exchanges, you don't actually have to report any financial numbers. I mean, it's really crazy.
Corey: That sounds like an interesting edge case that is not broadly known.
Charles: Yeah. So, if only American companies could get the same treatment. Financials? Eh, who needs them?
Corey: Yeah, just take our word for it. It's fine.
Charles: Yeah. It's pretty tough to see what's going on there. Somebody told me that they have recently gone back and started to report—they’re unaudited numbers, but they're more detailed numbers, so I was going to go take a look at those and try to get a sense of which of their CAPEX expenses might be much cheaper. They're probably paying less for concrete and labor, their Intel processors, I'm sure they're still paying the same prices as anybody else. So, that's another thing that's kind of on my list to go look at, but last time I looked at it, it was very, very difficult to make any headway on what's really going on there and how big they are.
Corey: It's sort of a giant open question mark. And I see these analyst reports saying that they're number three there, but I've got to say, I've never seen them in the wild. As I wind up walking through the world and talking to big customers. There was one company I talked to that was global and did a small workload there, but they were running on everything. I mean, I certainly see Azure, I see GCP a lot, too, and then there's a couple of weird exception edge cases.
I’m increasingly starting to see Oracle Cloud, of all things, around data transfer deals, where maybe having a retail price that is not eight times as expensive is starting to win those deals over. And yes, of course, everyone at size pays discounted rates on every provider, but when the list price is an order of magnitude or so larger than someone else's, people aren't even going to call and request a quote past a certain point.
Charles: Yeah, it's funny to think of Oracle as the low-cost leader.
Corey: Oh my god, I've always operated under the assumption that if I recommend—“You should pick Oracle,” for an economic story, great, that would get me thrown out of the room, possibly via the window. But it is economical in this case.
Charles: They're in their ‘we try harder’ phase. We'll see how long that lasts.
Corey: Yeah, number two tries harder. And yeah, Oracle's has a lot of number two, historically.
Charles: Number four, or five. But, who's counting?
Corey: One thing that is fascinating to me is that it's clear that they're making some aggressive moves in the place. They've hired some amazing executives out of AWS—and other places—and given them free rein. The biggest weakness I see an Oracle Cloud today is the word Oracle at the front of it, but I would not necessarily count them out. The problem is there's so much bad blood around Oracle that there has to be a massive reputational and perception change. I would say it's impossible except for the fact that Microsoft did it. Where’s their Satya Nadella moment? It [unintelligible] require having Larry Ellison sent to a home, but that's neither here nor there.
Charles: It's more than Larry. If you're a distant challenger in the market, you really need customers who are willing to take a chance on you and make a bet on you, and it's hard to see a lot of people are willing to do that with Oracle. They have hired a lot of people up here in Seattle. This is their second, third, fourth go around on that. I'm not sure that strategy has been super successful for them, the kind of people they've hired, they've been more motivated by big salaries and big titles, necessarily, than building something great. And so I don't think they're going to be able to hire up a bunch of people and successfully challenge the big clouds because it's just too late in the game. They needed to do it 10 years ago when Larry was still poo-pooing Cloud.
Corey: I don't know, I mostly try and stay out of the predict the future market, which I know makes me something of a rarity, but the reason behind it was always that my first year doing this stuff in a ‘predict what's going to be released at re:Invent’ capacity was also the last year I did it because I remember the day before that newsletter went out that I had a customer who, through their TAM was getting roadmap emails of selected things that were coming out, and I double-checked that, and it turns out, I’d gotten three or four things, right, so I had to quickly go through and remove it, and I realized the risk was too high because I had access to inside information from time to time, and when I got it right, no one would have believed that I didn't break NDA. So, it just easier for me to stay away from forward-looking speculative things around specific services or products.
Charles: Yeah. I mean, you don't have to look at the specifics. You just have to look at the magnitude of the gap that exists between—I mean, this gap is true. AWS is probably twice the size of Azure, Azure is growing 50 percent plus faster. The gap between Azure and GCP is probably 4x, and we don't know what the GCP growth numbers are, but they're probably growing marginally faster than Azure, and arithmetically it's extraordinarily hard to paint a scenario where they can catch up in any interesting timeframe.
If you then go look at some of the also-rans, the Oracles, the IBMs, they need to be growing many, many, many times faster than the existing leaders in order to catch up because they're starting from such a small base, and it's just impossible to pencil out the math that makes that work. And you start to look at cumulative CAPEX spend, you start to look at number of employees, you start to look at portfolio of services, it's really, really hard to catch up. I mean, even today, GCP is super challenged. I mean, they have lots of talent, they have lots of money, they have phenomenal infrastructure, and they're really having a hard time keeping up with AWS and Azure. And in the end, I think it comes down to the fact that GCP is a hobby, and for the other two companies, AWS and Azure are existential for those two companies.
Corey: The challenge, too, is that there's a very much a virtuous flywheel effect here, where when you're one of the big providers you are drawing customers in, who are A) asking for new and creative things that give you ideas to explore next in ways that satisfy their needs, and they're paying you an absolute mountain of money, so you can afford to make those investments in building those things out. Whereas, if you're trying to enter the stage as a cloud provider, you're still struggling with operational concerns that the big players have already mastered. Things like, “How do we run Linux VMs in various parts of the world reliably?” There's a gulf that's only getting larger. I'm sort of worried on some level that Google is completely going to screw this up, and Microsoft is going to remain relegated to the enterprise world leaving us, more or less, in a world of an Amazonian monoculture, and I don't want that.
Charles: Yeah, I think we're going to see Amazon continues to be the leader, I think Microsoft's going to be a strong challenger. I mean, the thing to remember is, as the enterprise opens up, an awful lot of the available money for the cloud market is on the enterprise side, and that explains why Microsoft continues to outgrow AWS. They're just a lot more mature on the enterprise side. The challenge for Google is Sundar wakes up in the morning and his Google Cloud, all up—not even just GCP, but combination of G Suite and GCP is, at best, the fifth-largest business.
Corey: And it's the first largest business, as far as something that is not directly or indirectly driven by advertising.
Charles: Yeah. The other four businesses above it on the list are nicely intertwined in terms of who the customer is, the skill sets, the talent inside the company. I described GCP as it’s a hobby, and that annoys my friends at Google, but that's the reality. And we'll see what they decide to do with it.
Corey: Yeah, I'm curious to see what happens. Thank you so much for taking the time to speak with me today.
Charles: It's been fun.
Corey: It really has. If people want to hear more about what you have to say and how you say it, where can they find you?
Charles: Well, I'm on Twitter at @Charlesfitz, charles-F-I-T-Z. And my website is www.platformonomics.com, which I sort of treat as long-form Twitter. You'll get really long posts from me, very sporadically.
Corey: Well, thank you so much, once again, for taking the time to speak with me. I've really enjoyed this.
Charles: You bet. It's been fun.
Corey: Charles Fitzgerald, managing director at Platformonomics. I'm Cloud Economist Corey Quinn, and this is Screaming in the Cloud. If you've enjoyed this podcast, please leave a five-star review on Apple Podcasts, whereas if you've hated this podcast, please leave a five-star review on Apple Podcasts along with a comment detailing the exact CAPEX breakdown of AWS’s last earnings release.
Announcer: This has been this week’s episode of Screaming in the Cloud. You can also find more Corey at ScreamingintheCloud.com, or wherever fine snark is sold.
This has been a HumblePod production. Stay humble.