 Hello everyone, welcome to theCUBE's coverage of AWS re-invent 2021. We're on site in person, it's a virtual event, also hybrid events. I'm John Furrier, my host, Dave Vellante. Ninth year, Dave, we've been doing AWS re-invent theCUBE and it's 11th season. We've seen a lot. Yeah, I'll say. I mean, and the show is pretty packed, John. I mean, I think it's surprising folks, over 25,000 people here. I mean, obviously a lot of sponsors, but a lot of customers too. A bad event for AWS in terms of attendance is like record breaking for any other company. People are standing in line for sessions. It's definitely happening. People are here to learn. They're not just all employees. So definitely a successful event in person, as well in the live stream, but so much news to talk about. Andy Jassy is now the CEO of Amazon. That's the top story. Adam Silebski is taking over as CEO of AWS. Long time Amazonian who left Amazon to take the CEO job of Tableau, sold that company Salesforce under Mark Benioff, now back to take the helm from Andy Jassy and quite the pressure cooker here as he takes the stage. A lot of people are asking is, will he do well? Will he fumble on stage? Will he do the right things? And does he have what it takes to take the cloud to the next generation with AWS as their number one clear and far and away than the second competitor at Microsoft? And then a distant third in Google. So Amazon's earned a ton of competitive pressure at least from an industry standpoint, everyone's still trying to catch up. It's the same theme, Dave, every year. Amazon's out front and the lead just gets extended and extended. And again, here, no exception. Well, the Uber story, of course, as you mentioned is Andy Jassy's now taking over as CEO of Amazon. And history would suggest that a lot of times that companies falter when there's a CEO transition. But it feels like it's different this time. Andy Jassy was here since the beginning, launched AWS, which is a profit engine of Amazon, brought back Adam Silebski who has a deep understanding, he's not as technical as Andy, but obviously has a deep understanding of the business. Yeah, he was seemed comfortable up in the keynote. It wasn't on a typical fire hose of announcements. Even though it's a lot of announcements, they didn't shove them down our throat and they didn't in the analyst session as well. Usually in the analyst session, it's hours and hours and hours of fire hose, Kool-Aid injection, not this year. Why do you think that is? Is that a COVID thing? Is that a change in? No, I think Adam Silebski wants to be his own guy as leader. A lot of things were eliminated from the keynote that Andy Jassy did. For instance, Andy Jassy loves music, so he always had the music, walk-up music, kind of like you see in sports, which is very cool. That's an Andy Jassy kind of tweak. Andy is all about announcements and he was just a push in the envelope. Adam was much more laid back. He sees, I think, more of a holistic picture, being more of an app guy, being more of a data guy, less of a, I would say, under-the-covers nerd like Jassy was. Andy was very deep on a lot of the tech stuff, as is Adam, but I think Andy a little bit more proactive on that. So, Adam was very much more about the impact of AWS culturally as a society, as a company, and kind of brought in this kind of think different Apple Vi, which is, the people who are pathfinders, as he takes that Jassy kind of approach of leaders, be a builder, be a change agent, be a game changer. Adam took it to another level by saying, hey, it's okay to be a pathfinder because it's net new disruption with the cloud. And I think that's the story that I see coming out of this where, in talking to Adam one-on-one, Amazon absolutely has a secret weapon in its chips, custom silicon, they're absolutely crushing it with how they're thinking about SaaS and platforms, and they have a huge ecosystem. And I think at the end of the day, and we talked about this in our story, on SiliconANGLE, Amazon could actually wipe out Microsoft. And I think Microsoft's core competitive advantage has always been their ecosystem and their developers. I think right now in the next few years, if Microsoft doesn't match Amazon, they will be decimated. Wait, wait, hold on, okay. Amazon's not going to wipe out Microsoft. Microsoft has too much of a cash cow. Look it, hanging onto Windows couldn't, the mistake in missing mobile, initially missing the cloud didn't wipe out Microsoft. So they've just got too much of a software cash flow. That's not going to happen. Maybe a little bit over the top. But Microsoft has done a great job on it starting to tell it to kind of stay in the game and do more. But if you look at the major inflection points, Dave, where's digital equipment corporation? Where's prime computer? Well, I think this is the point is, again, history would show that those companies when they handed the reins over to new CEO failed. They faltered, they were self-inflicted wounds. It almost happened, you thought it would happen with Microsoft where they became irrelevant under bomber. But when Nadella came in, he reinvigorated it because specifically they had the cash flow to be able to do that. So the big question is okay, what's going to happen? We ran a survey to our community to see what could disrupt Amazon. You know, the US government wants to break them apart, you know, or wants to regulate them. But our survey respondents said there's a 60% plus probability that Amazon will be disrupted by other factors. And that's really- And self-inflicted wounds. And that's Jassy's, that's right. And that's Jassy's big challenge is how to not make those disruptions, how to fight those disruptions. The number one reason why they could be disrupted was self-inflicted wounds, which again, history would show would happen. But one of the things we talked about is that normally happens when companies stop innovating, when they rest on their laurels, right? And you kind of, you kind of saw that with those companies that you mentioned. But you mentioned their secret weapon. We wrote about that in our article, the chip. So we heard no secret. Everybody knew Graviton 3 was coming, right? And so that is Amazon's secret weapon. You know, I've been thinking about this, John. Amazon makes a lot of money on x86 instances that they've deployed years ago, right? And they charge a lot for. I was wondering, are the old x86 instances actually more profitable than Graviton? Maybe at this point in time, but long-term, Graviton, they control their own destiny because they control the hardware and software stack. And I bet you it allows them to get better negotiating leverage with AMD and Intel. Of course. I mean, Pat Gelsinger, we should talk about this all the time, but about as days in Intel, if you're not out in the next wave, you're driftwood. And I think Intel and AMD and others, they have purpose-built general-purpose chips. They're probably going to be for the lift and shift stuff. But if you're actually seriously writing software as an owner on the cloud and you want specific advantages of speed and performance, you're going to want the custom silicon that's purpose-built for your application and write code to that stack. So I think there's a whole another level of platform as a service, Dave, that's kind of coming out of this reinvent that I think could be a multi-generational trend, which is, hey, the cloud is a super cloud or platform. Look at the rise of Snowflake and Databricks. Those guys are on Amazon. Like, they're super clouds and themselves, they're platforms, they're not a point-sass solution. I think Microsoft in my analysis says, yeah, they got Office 365, okay, word processing stuff. But what other SaaS apps do they have assigned SQL Server and other things that are actually being built on there? And if I'm a developer, you're going to want to go to the platform, that's the highest performance for the application. Office 365, right? Yeah, it's a cash cow. But how long is that going to last? A long time. I mean, Microsoft's got a major momentum. We can argue about that later, but I want to touch on Graviton 3 because I think that was the big announcement of the day. 25% faster than Graviton 2, at least. Twice the floating point performance, twice the cryptographic performance, and three times for machine learning workloads and very importantly, 60% less power. So at Amazon scale, Adam said this in our meeting. He said, the economics really favor us because of our scale. And so, and they've also announced new trainium instances and what having custom silicon allows Amazon to do is release on a much, much faster cadence than traditional x86. And they could do really cool things. Nitro is their smart nick, which is the basis, their new hypervisor, if you will. So it allows them to bring in x86 NVIDIA NPUs, some of their own, or NVIDIA GPUs, some of their own silicon. So optionality is really the key there. You heard them announce an SAP instance, so that's a memory intensive instance. They can dial things up, dial things down, they've got full control of the stack. And by the way, copying them, Google's copying them, Microsoft is copying them. And who's leading this charge in custom silicon? AWS obviously, Tesla, Apple. I mean, these are leading companies that I don't think they all got it wrong. I think they got it right. The silicon angle is to have your own custom silicon. That is clearly the advantage, as it's vertically integrated. But the other thing that's coming out of this reinvents the purpose-built software concept, where they're not copying Microsoft's playbook as the Wall Street Journal was saying, and some are saying, Microsoft's copying Amazon. Amazon has always been this horizontally scalable resource. That's cloud. But with machine learning and AI, you now have this purpose-built capability from software into the app itself, where data has to be addressable. And I think the people in the data business kind of know this, but as the rest of the world comes out architecturally, having that horizontal observation space and data that's vertically tied to machine learning is a huge architectural shift. This is a complete rethinking of how software's built. And that's going to be a game changer. I think Amazon's well out in front of that. And I think that's going to be a huge architectural shift. Well, let's quantify this a little bit, because you're making it the point that Amazon is the number one cloud, which I would agree with. We're talking here about IaaS, infrastructure as a service, and the PaaS layer that sits on top of that. Microsoft defines the cloud as, we'll put in an Office 365, Google will put in its Google apps, Amazon's pure infrastructure as a service. And if you just look at that space, that's about $120 billion business when you add up AWS, Azure, Alibaba and GCP, which I would contender the only four hyperscalers out there. I don't include Oracle as a hyperscaler. I don't include IBM. I get a lot of crap for that sometimes. But we're talking big scaler, $120 billion. So actually relatively small compared to the trillion dollar opportunity that they have, but it's growing at 35% a year. Amazon will do more than $60 billion this year, $62 billion. Just to quantify it in that IaaS space, Microsoft will be about $38, $39 billion, okay? So pretty substantial. Those two are far ahead of the others. Everybody else is, you know, Google's still under $10 billion, Alibaba's right around there. So those two, it's really a two-horse race in IaaS. Microsoft, using its software estate, Amazon's got to be the innovator and has to have the best cloud to win, and it does. Well, it's also a platform. Let's go back to a little history lesson for the younger folks out there. When Microsoft had a monopoly, they had Windows operating system, which had DOS under the coverage, but Windows was the operating system and Office was a suite of applications. They encouraged software developers to build on top of Windows. And they had other servers, soft SQL server all came out, all that's all history. So their bread and butter was to have developers build on top of Windows, hence the monopoly. Of course they had the application and the system software, hence the monopoly, hence the Microsoft breakup by the government in 1997. Now today, cloud is essentially one big kind of PC concept. It's like Windows, it's the Windows equivalent. So cloud is essentially an environment platform that has apps that run on top of it, okay? In that world, Amazon by far is the number one. Windows model at Amazon, I mean Microsoft's used to is, okay, I got Azure and I got Office 365. That keeps them in business. That keeps them from losing. So that's a placeholder. So what I'm looking at is what is Amazon, I mean Amazon versus Azure doing relative to ISV and uptake for developers. And I'm suggesting that this continuing trend of Amazon will go, if it goes uncontested by Azure, they'll wipe the table on ISVs and software developers. If you're an owner of a software, you're not going to write software that's going to be sub-optimized for a platform that's not going to be performing. Unless you're a Microsoft developer, nearly all dot net days, and there are a lot of those. And that's what Microsoft is doing there. They've shifted to cloud. They've gone everything into cloud. So Azure is their platform for innovation and acceleration. So those developers going to build a sub-optimized application versus going over here on AWS. Well, that's the story with Microsoft. Good enough. Again, we're speculating, but we're going to watch that. But that is to me, it will be the battlefield of what will determine Azure versus AWS. And I think everything else is smoking mirrors. Amazon's webs are way ahead of Azure, but the tell sign is going to be, does AWS attract those developers on their cloud with the custom silicon, with the integrated stack, and with the purpose-built software? I mean, it's looking really good. I think they got a really compelling story. I do think it's less about Azure versus AWS. I mean, that's an interesting storyline, and I love to talk about it. But I think they'll go back to 120 billion out of 4 trillion. That's really the larger opportunity for both Microsoft and AWS to continue to grow. Because you look at Dell with Apex, you look at HPE with GreenLake, Lenovo, Cisco, they've all got their own clouds. One of the things that didn't get into our article, but Adam Solipski, when you asked him about hybrid, is that hybrid cloud? When we were talking about some of the stuff they're doing, he said, look, that's not cloud, what those guys are doing. That's not what we did. And he talked today about Edge, it has to be AWS, not like AWS, that was the quote he used, talk about private 5G, bringing out posts, and he gave some examples of that. The point is, AWS is bringing its system, its architecture to the Edge, its programming model, infrastructure is coded to the Edge. Now, Kubernetes doesn't moderate that a little bit, but his point was, that's not AWS, that's not the cloud. Yeah, I think in summary, Dave, I have to wrap up what's the big trend this week, is that Amazon Web Services is a heaven environment for a developer, for the elite people who want to roll their own. For the folks in IT and these other environments, you can have prefabricated, purpose-built software platform to build on top of, and I think that is going to address the whole ease of rollout. So if I'm a SaaS developer, I don't want to rebuild that over again, I don't want to roll my own, I'll take what you got. And connect's a good example. If you want a call center, you can take it, use it, and then build on top of it, and iterate on it. So I think it's more of, here's a platform for you and take it, so I think that to me is the big story. And that's not, and think about it, how many people out there are rolling their own Amazon? You got to be pretty strong at Amazon primitives to roll your own. A couple of other quick points that, he barely emphasized the primitives, the APIs, that multiple databases, right tool for the right job, took a shot at Oracle without mentioning Oracle, because there's sort of one database. But I will say this, mission critical, Oracle still owns that. They talked about a mainframe migration tooling and runtime for mainframe compatible runtime. That's going to allow them to nip at the edges of those mainframe workloads and Oracle workloads, if they're not going to get to the core anytime soon. They also talked about roll level and cell level security. We think that's the squirrel acquisition from years ago. And then he made a statement, we have three X with Redshift price performance, better than any cloud data warehouse. Sort of interesting shot at Snowflake. And Databricks. And Databricks, so anyway. Yeah, I mean, I think overall, I thought Adam did a good job. I think he didn't disappoint. Okay, but I think his goal was to get through this and not have people go, well, it's not Andy Jassy. I thought he did an awesome job and he did a good job and he got what he needed to do. He was comfortable and he obviously leaned on some of his Pathfinder customers. NASDAQ, I thought was very impressive. United Airlines dish, so. Okay, CUBE coverage, ninth year of the CUBE here at AWS re-invent 2021. It's the CUBE you're watching, the leader in high tech coverage at CUBE.