 from the globe, it's theCUBE, with digital coverage of AWS re-invent 2020, sponsored by Intel, AWS, and our community partners. Hello everyone, welcome back to theCUBE's live coverage, CUBE live here in Palo Alto, California, with the virtual CUBE this year, because we can't be there in person, I'm your host, John Furrier. We're kicking off day two of the three weeks of re-invent. A lot of great leadership sessions to review, obviously still buzzing from the Andy Jassy three hour keynote, which had so many storylines, it's really hard to unpack. We're going to dig that into that today with Jerry Chen, who's been a CUBE alumni since the beginning of our AWS coverage, going back to 2013, Jerry was wandering the hallways as a in between, you were in between VMware and VC, and then we saw you there, you've been on theCUBE every year at re-invent with us. So a special commentary from you, thanks for coming on. Hey, John, thanks for having me, and a belated happy birthday as well. Everyone out there, John's birthday was yesterday, so the hardest working man in technology, he spent his entire birthday doing live coverage of Amazon re-invent, so happy birthday buddy. Well, I love my work, I love doing this, and re-invent is the biggest event of the year, because it really is, it's become a bellwether, and so super excited to have you on. We've had great conversations by looking back at our conversations over the Thanksgiving weekend, Jerry. The stuff we were talking about, it was very, I propose that Jassy's leaning in with this whole messaging around change and horizontal scalability, he didn't really say that, but he was saying, you can disrupt in these industries and still use machine learning. This was some of the early conversations we were having on theCUBE, now fast forward more mainstream than ever before. So big part of the theme there. Yeah, it's, if you, Amazon re-invent Amazon evolution to your point, right? Because it's both re-inventing what companies are using with the cloud, but also what Amazon's done is they're evolving year after year with their services. So they started a simple infrastructure, you know, S3 and EC2, and now they're building basically a lot of what Andy said, yesterday deconstructed CRM. A lot of the stuff they're doing around the call centers almost going after Salesforce with the kind of deconstructed CRM services, which is super interesting because the day, you know, Amazon announces all those technologies, not to mention the AI stuff, the semiconductor stuff. You have Slack and inquired by Salesforce for $27.7 billion. So a lot of stuff going on in the cloud world these days and it's fun to be part of it. You know, it really is interesting. You look at the Slack acquisition by Salesforce, that's interesting. You know, that kind of takes Slack out of the play here. I mean, they were doing really well, again, message board service, turns into a more collaboration software. They hit the mainstream, they have great revenue. Is that going to really change the landscape of the industry for Salesforce? They got to acquire it. It opens the door up for more innovation. And it's funny you mentioned the contact center because I was pressing Jassy on my exclusive one. I went with them, I'm like, it's an Andy, my daughter and my sons, they don't use the phone. They're not going to call. Is it a call center deal? And he goes, no, it's about the contact. So the thing about that notion of the contact, it's not about the call center. It's the point of contact. Okay, LinkedIn is with Microsoft. You got Slack and Salesforce, contact-driven, collaboration, interesting kind of play for Microsoft to use voice and their data. What's your take on all that? I think it's, you know, I have this framework as you know, I talk about systems of engagement over systems intelligence and systems of record, right? And so you can argue voice, email, Slack, these are all different systems of engagement. And they sit on top of the system of record, like CRM, customer support, ticketing, HR, something like that. Now what Salesforce did by buying Slack is they now own a system of engagement, right? Not only is Slack a system engagement for CRM, but it's also a system engagement for ERP, ServiceNow. It's how you interact with a bunch of applications. And so if you think about Salesforce strategy in this space to compete against Microsoft or ServiceNow or other large ERPs, now they own Slack and system engagement that's a super powerful way to actually compete against rival SaaS companies because if you own the layer engagement layer, you can now just intermediate what's in the background. Likewise, the contact center, your own voice, email, chat, messaging, right? You can just intermediate the stuff in the back. And so they're trying to own the system engagement. And then likewise, Facebook just bought that company customer a week ago for a billion dollars, which is also omnichannel support because it's chat messaging voice. It's again, the system engagement between your end user, which could be a customer or it could be employee. You know, this really kind of makes, enterprise has been so much fun over the past 10 years, I got to say in the past five, you know, it's been even more fun as it becomes more the new fun area, you know. And the impact to enterprises has been interesting because we're talking about system engagement, this is now the new challenge for the enterprise. So I want to get your thoughts, Jerry, because how you see the CXOs and the CISOs and the architects out there trying to reinvent the enterprise. Jassy saying, look at find the truth, be on the right side of history here. Certainly he's got some self-service interest there, but there is a true band-aid with COVID and with digital acceleration for the enterprises to change. Given all these new opportunities to revolutionize or disrupt or erratically improve, what's the CXOs do? What's your take on, how do you see that? It's increasingly messy for the CXOs, and I don't envy them, right? Because back in the day, they kind of controlled all the IT spend and kind of they had a standard of what technologies they use in the company. And then along came Amazon and cloud and all of a sudden, like your developers can go, hey, let me swipe my credit card, I'm going to access to a bunch of APIs around compute and storage. Likewise, now they can swipe their credit card and use Stripe for billing, right? There's a whole bunch of services now. So it becomes incumbent upon CXOs. They need to use set of management tools, right? So it's not only just like security tools they need, they need also observability tools, understanding what services are being used by the customers when and how. And I would say the following, John, like CXOs is both a challenge for them, but I think if I was a CXO, I'd be pretty excited because now I have a bunch of other weapons and other bunch of services I can offer my end users, my developers, my employees, my customers. And, you know, it's exciting for them, it's not only can they do different things, but they also change how their business is being done. And so I think both interacting with your end users be a chat like Slack or be a phone, like your contact center or Instagram for your kids. It's actually a neat challenge. If I were a CXO, it's time to build again. You know, I think COVID has said it's time to build again and you can build. To kind of take that phrase from the movie Shawshank Redemption, get busy building or get busy dying, kind of rephrase it there. And that's kind of the theme I'm seeing here because COVID kind of forced people saying, look at these things like work at home, who would have thought 100% people would be working at home? Who would have thought that now the workloads got to change differently? So it's an opportunity to deconstruct or disintermediate these services. And I think, you know, in all the trends that I've seen over my career, it's been those inflection points where breaking the monolith or breaking the proprietary piece of it has always been an opportunity for entrepreneurs. So, you know, and for companies, whether you're a CIO or a startup, by decomposing and you can come in and create value. And I think to me, snowflake going public on the back of Amazon, basically. This is interesting. I mean, so you don't have to be, you can kill one feature and nail it and go big. I think we talked in the past, like is Amazon or Google or Microsoft going to win everything? Because it wouldn't take all, wouldn't take most. And you can argue that it's hard to find oxygen as a startup in a broad platform play. But I think snowflake and other companies have done and comes like MongoDB, for example, or Elastic, have shown that if you can pick a service or a problem space and either develop like IP that's super deep or own developer audience, you can actually fight the big guys, the big three cloud vendors, be Amazon, Google or Microsoft in different markets. And I think if you're a startup founder, you should not be afraid of competing with the big cloud vendors because there are success patterns and how you can win and create a lot of value. So as a founder investor, I'm super excited by that because I don't think you're going to find a company to take down Amazon completely because they're just the scale and the network effect is too large. But you can create a lot of value and build valuable comes like snowflake in and around the Amazon, Google, Microsoft ecosystem. Yeah, I want to get your thoughts. You have one portfolio we've covered, Rockset, which does a lot of SQL, one of your investments. Interesting part of the keynote yesterday was Andy Jassy kind of going after Microsoft saying Windows SQL Server, they're targeting that with this new tool that sucks in the database of it. It's called the Babelfish for Aurora for Postgres SQL. Well, what's your take on that? I mean, obviously Microsoft's big, their enterprise sales tactics are looking like more like Oracle, which he was kind of hinting at and commenting on. But SQL is lingua franca for data. Right. You know, I think we went to like kind of a no SQL phase, which was kind of the trendy thing for a while. And that no SQL is still around, not only SQL, like MongoDB, DocumentDB, kind of that interface still holds true. But to your point, the world speaks SQL, all your applications speak SQL, right? So if you want backwards compatibility to your applications, speak SQL. If you want your entire installed base of employees that no SQL, you got to speak SQL. So Rockset, one of the first public conversations about what they're building was on the key with you and me and Venkat, the founder. And what Rockset is doing, they're building real-time snowflake to do lack of better term. It's a real-time SQL database in the cloud that's super elastic, just like snowflake is. But unlike snowflake, which is a data warehouse, mostly for dashboards and analytics, Rockset is like millisecond queries for real-time applications. And so think of them as the evolution of where cloud databases are going. It's not only elastic like snowflake and the cloud like snowflake, we're talking 10, 15 millisecond queries versus one to two second queries. And I think what any Jassy did in Amazon with Babelfish is say, hey, SQL is the lingual Frank of the cloud. There's a large installed base of SQL Server developers out there in applications. And we're going to use Babelfish to kind of move those applications from on-premise the cloud or from old workloads to the new workloads. And I think the name of the game for cloud vendors across the board, big and small startups to Google, Microsoft and Amazon is how do you reduce friction? Like how do you reduce friction to try a new service to get your data in the cloud, to move your data from one place to the next? And so Amazon is trying to reduce friction by using Babelfish. And I think it's a great move by them. Yeah, and by the way, not only is it for Aurora Post-Christ SQL, they're also open sourcing it. So that's going to be something that is going to be interesting to play out because once they open source it, essentially that's an escape valve for lock-in. I mean, if you're a Microsoft customer, I mean, it ultimately is, could be that gateway drug. It's like, it is ultimately like, hey, if you don't like the licensing, come here. Now there's going to be some questions on the translation, some, some, some scuttle-butt about that. But we'll see, it's open source. We'll see what goes on. Great stuff on, on Rockset. Great, great startup. Next, next talk track I want to get with you is, you know, over the years, you know, we've talked about your history. We're going to VMware, and now being a venture capitalist successful, we're going to Greylock. You've seen the waves and I would call it the two ways. Pre-cloud, early days of cloud, and now with COVID, we're kind of in the, you know, not just born in the cloud, total cloud scale, cloud operations. This is kind of what Jassy was going after. And I think I tweeted cloud is eating the world and on premise and the edge is what it's hungry for. Kind of goof on Mark Andreessen's quote of software eating the world. This is where it's going. So it's a whole nother chapter coming. You saw the pre-cloud, you saw cloud. Now we got basically global IT, everything else. It's cloud only. I would say, you know, we saw pre-cloud, right? The VMware days, and before that, you called like, you know, data centers. I would say Amazon launch in what, 2006? 2007, the web services. So the past 14, 15 years have been what I've been calling cloud transition, right? And so you had companies, technologies that were either doing a migration from on premise to cloud or hybrid on premise, off premise. And now you're seeing a generation of technologies and companies that are cloud only, John, to your point. And so you can argue that this 15 year transitions were like, you know, to use a bad metaphor, like amphibians, you're half in the water, half on land, you know, and like, you know, you're not purely cloud, you're not purely on premise, but you can do both ways. And that's great, that's great because that's a dominant architecture today. But with companies like Rockset and Snowflake, you're cloud only, right? They're born in the cloud, they're built on the cloud. And now we're seeing a generation of startups and technology companies that are cloud only. And so, you know, unlike, you had this transitionary evolution of like amphibians, land and sea, now we have, I don't know, mammals, whatever, that are only in the cloud, only on land. And because of that, you can take advantage of a whole different set of constraints that are cloud only, that can build different services that you can't have going backwards. And so I think for 2021 forward, we're going to see a bunch of companies that are cloud only and they're going to look very, very different than the previous set of companies the past 15 years. And as an investor and as you covering as an analyst, it's going to be super interesting to see the difference. And if anything, the cloud only companies will accelerate the move of IT spending, the move of more developers to the cloud because the cloud only technologies are going to be so much more compelling than the amphibians, if you will. Yeah, and to see your point, and you saw the news announcement, they had a ton of news, a ton of SageMaker, we got called kind of the democratization layer. But if look at some of the insights that Amazon's getting just as the monster that they are in terms of the size, the scope of what their observation space is, they're seeing all these workloads. They have the DevOps guru, they launched that DevOps guru thing I found interesting. They got data acquisition, right? So when you think about these new, the new data paradigm with cloud only, it opens up new things, new patterns. So I think to me, I think that's to me, I see where this notion of agility moves to a whole nother level where it's not just moving fast, it's new capabilities. So how do you see that happening? Because this is where I think the new generation is going to come in and it's be like servers, Lambda, like you guys actually provisioned EC2 instances before it was servers on data centers. Now you got EC2, what, Lambda? So you start to see smaller compute, new learnings, all these historical data insights feeding into the development process and to the application. I think it's interesting. So I think if you really want to take the next evolution, how do you make the cloud programmable for everybody? I think you mentioned SageMaker, machine learning, data scientists. The SageMaker user of the data scientists, for example, does not allow provision containers and code YAML files and understand Kubernetes, right? Just like the developers today don't want to rack servers. Like, oh my God, Jerry, you had to rack servers in data center and install VMware. The generation beyond us doesn't want to think about the underlying infrastructure. They want to think about how do I just program my app and program the cloud writ large? And so I think where you can see going forward is two things. One, people who call themselves developers, that definition has expanded the past 10, 15 years. It's only growing. So everyone is going to be a developer right now, from your white collar knowledge worker to your hardcore infrastructure developer, but the population developer is expanding, especially around machine learning and kind of the SageMaker audience for sure. And then what's going to happen is, a lot of this audience doesn't want to care about the stuff you just mentioned, John, in terms of the online plumbing. So what Amazon, Google, and Azure would do is make that stuff easy, right? Or a startup can make it easy. And I think that the move towards Lambda and serverless is that move specifically that, don't think about the underlying plumbing. We're going to make it easy for you, just program your app and then either a startup will abstract away all the underlying infrastructure bits or the big three cloud vendors will say, you know, all this stuff would do in a serverless fashion. So I think serverless as a paradigm and kind of quite frankly, a battle front for the big three clouds and for the startups is probably one of the front lines of the next generation. Whoever owns the kind of programmable cloud model, the programming the internet, programming the cloud will be maybe the next platform for the next 10 or 15 years. And I think it's still up for grabs. Yeah, I think that is so insightful. And I think that's worth calling out. And I think that's going to be a multi-year effort. I mean, look at just how containers now with EKS anywhere and you got the container service of the control plane built in. You got, you know, real-time analytics coming in from Rockset and Amazon. You have Pennora, Panorama appliance that does machine learning and computer vision with sensors. I mean, this is just a whole new level of purpose of built stuff. Yeah. Software powered, software operated. So, you know, you have this notion of DevOps going to hand in the gloves, software and operations, kind of how do you operate this stuff? So I think the whole new next question was, is, okay, this is all great, but Amazon's always had this problem. It's just so hard, like there's so much good stuff. Like, who do you hire to operate it? It's not yet programmable. This has been a big problem for them. Your thoughts on that? You know, I think that the data illusion around DevOps, et cetera, is the solution. So all of a sudden you're going to have information from Amazon or from startups that are going to automate a bunch of the operations. And so, you know, I'm involved in a kind of chronosphere that we've talked about in the past. Team kind of Uber, they built something called M3 that's basically kind of a next generation data dog, next generation of this reliability platform. They're going to collect all the data from your applications. And once they have your data, they're going to know how to operate and automate scaling up, scaling down, and the base remediation for you. So you're going to see a bunch of tools take the information from running your application infrastructure and automate exactly how to scale and manage your app. And so AI and machine learning at large, John, is going to basically make a lot of this plumbing go away or maybe not completely, but let's you scale better. So you as a single system admin or use a single SRE site reliability engineer can scale and manage a bigger application. And it's all going to be around automation. And to your point, you said earlier, if you have the data, that's a powerful situation. Because once you have the data, you can build models on it and you can start building solutions on the data. And so I think what happens is when we build this programmable cloud for your broad developer population, automating all this stuff becomes important. And so that's why I say serverless or this automation of infrastructure is the next battleground for the cloud because whoever does that for you is going to be your virtualized backend, virtualized data center, virtualized SRE. And if whoever owns that is going to be a very, very strategic decision. Yeah, that's great stuff. And this is back to the theme of this notion of virtualizations now gone beyond server virtualization. It's media virtualization with theCUBE, my big joke here with theCUBE virtual, but it's to your point, everything can now be replicated in software and scaled at cloud scale. So it's super big opportunity for entrepreneurs and companies to pivot and differentiate. The question I have for you next is on that thread, huge edge discussion going on, right? So I think I said it two years ago or three years ago. The data center is just an edge. It's just a big, fat edge. Jassy kind of said that on his keynote. He looks at that as just an edge point with his, from his standpoint, but you have data center, you have real edges. You've got 5G with the wavelength, this local zone concept, which is, you know, Amazon in these metro areas reminds me of a built wireless point of presence kind of vibe. And then you've got just purpose built devices like cameras and factories. So huge industrial innovation, robotics meet software. I mean, a whole huge edge development exploding. Which, what's your view of this and how do you look at that from as an investor and industry? I think edge is both the opportunity for startups and companies as well as a threat to Amazon, right? To the reason why they have outposts and all this stuff in the edge is if you think about, you know, decentralizing your application and moving into the edge from my wearable to my home to my car to my city block, edge is actually super interesting. And so a couple of things. One, companies like Cloudflare and Fastly and a company I'm involved with called Kato Networks that does Sassy secure access service edge, right? The names and the edges in the category definition. Sassy is about how do you like get compute to the edge securely for your developers, for your customers, for your workers, for your end users. And what, you know, companies like Cloudflare and Kato have done is they built out a network of pops across the world, their own infrastructure. So they're not dependent upon, you know, the big cloud providers and the telco providers, you know, they're partnering with the big cloud, they're partnering with the telcos, but they have their own kind of system or own kind of platform to get to the edge. And so companies like Kato Networks and Cloudflare that have a presence on the edge and their own infrastructure more or less, I think are going to be in a strategic position. And so Kato has seen benefits in the past year of COVID and lockdown because more remote access, more developers, I think Edge is going to be a super great area of development going forward. I think if you're Amazon, you're pushing to the edge aggressively with Outpost, I think your developer startup, you know, creating your own infrastructure and riding this edge wave could be a great way to build a boat against a big cloud guy. So I'm super excited if you think Edge and this whole idea of your own infrastructure, like what Kato has done, it's going to be super useful going forward. And you're going to see more and more companies spend the money to try to copy kind of a Cloudflare, Kato presence around the world because once you own your own kind of infrastructure, own set of pops and you're less dependent upon them, a cloud provider, you're in a good position because there was an Amazon outage last week and I think like Twilio and a bunch of other services went down for a few hours. If you own your own set of pops, you're independent of that. It is actually really secure. Yeah, and if they go down too, it's on you, but that was the Kinesis outage that they had they before Thanksgiving. Yeah, that's the problem. So I guess the question for you on that is is it better to partner with Amazon or try to get a position on the Edge and have them either buy you or compete or create value or co-exist? How do you see that strategy move? Do you co-exist? Do you play with them? I think you have to co-exist. I think you have to partner co-exist, right? I think like all things you compete with Amazon, Amazon is so broad that there'll be parts of Amazon you're going to compete with and that's fair game. You know, like snowflake competes against Redshift, but they also partner Amazon because they run an Amazon. So I think if you're a startup trying to fight the Edge, you have to co-exist with Amazon because they're so big, big cloud, right? The big three cloud, Amazon, Google, Azure, they're not going anywhere. So if you're a startup founder, you definitely co-exist, leverage the good things of cloud, but then you got to invest in your own Edge, both figuratively what your Edge and literally the Edge, right? And I think you compliment your Edge presence, be it the home, the car, the city block, the zip code with using Amazon strategically because Amazon is going to help you get to different countries, different regions. You can't build a company without touching Amazon in some form of fashion these days, but if you're a startup founder doing strategically how you use Amazon and picking how you differentiate is going to be key. And the differentiation might be small, John, but it could be super valuable, right? So maybe it's only 10 or 15%, but that could be a whole ton of value that you're building on top of it. Yeah, and there's a little bit of a growth hack too with Amazon, if you know how it works, if you compete directly against a core building blocks like EC2, S3, you're going to get killed, right? They're going to kill you. If the white space is interesting, in the old days in Microsoft, if you had a white space, they'd give it to you or they'd roll you over and level you out. Amazon, if you're a customer and you're in a white space and do better than them, they're cool with that. They're basically like, hey, if you can innovate on behalf of the customer, they let you do that as long as you have a big bill. Yeah, I love it. Snowflake's paying a lot of money to Amazon, sure. But they also are doing a good job. So again, Amazon's been very clear on that. If you do a better job than us for the customer, we'll do it. But if they want Amazon Redshift, they want Amazon only, they can choose that. So kind of the playbook. I think it's absolutely right, John. It sets from any jassy and then the Amazon culture, the customer comes first, right? And so whatever's best for the customer, that's like their mission statement. So whatever they do, they do it for the customer. And if you build a value for the customer and you're on top of Amazon, they'll be happy. You might compete with some Amazon services, which know the GM of that business may not be happy. But overall, net-net, Amazon's getting a share of those dollars that you're charging the customer, getting a share of the value you're creating. They're happy, right? Because what's the line? Rising tide floats all the boats. So the more cloud usage is going to only benefit the big three cloud providers, Amazon in particular, and because they're the biggest of the three, but more and more dollars go to the cloud. If you're helping move more apps to the cloud, you're helping build more solutions to the cloud, Amazon's going to be happy because they know that regardless of what you're doing, they will get a fraction of those dollars. Now, the key for a startup founder and what I'm looking for is, how do we get more than a sliver of the dollars? How to get a bigger slice of the pie, if you will. So I think edge and serverless are two areas I'm thinking about because I think there are two areas where you can actually invest, own some IP, own some surface area and capture more of the value to use a startup founder and our belt lasts to you to Amazon. Yeah, great thesis. Jerry, it's always been great. You've been with theCUBE since the beginning on our first reinvent in 2013. And we're now on our eighth year. Great to see your success, great investments you make. You're a world-class investor at a great firm, great luck, great to have you on from your perspective. Final take on this year, what's your view of Jassy's keynote? Just in general, what's the vibe? What's the quick sound bite for you? First, I'm so impressed, Andy, can you feel like a three-hour keynote more or less by himself, right? That is, that's a one-man show. And I'll hold that because I don't think I can pull that off, number one. Number two is the ability for Amazon to execute at so many different levels of stack from semiconductors, right? They're AI chips to high level of services around healthcare solutions and logistics solutions. It's amazing. So I would say both I'm impressed by Amazon's ability to go so broad up and down the stack. But also I think the theme from Annie Jassy's like it's just acceleration. It's, now that we will have things unique to the cloud and that could be just the AI chips unique to the cloud or the services that are cloud only, you're going to see a tipping point. We saw acceleration the past 15 years, John, he called like this cloud transition. But I think you and I are talking about 2021 beyond, you'll see a tipping point where now you can only get certain things in the cloud, right? And that could be the underlying inference instances or training instances that Amazon's giving. So all of a sudden you as a founder or a developer says, look, I guess so much more in the cloud, there's no reason for me to do this hybrid thing. Hybrid's not going to go away, on-prem's not going away, but for sure we're going to see an increasing acceleration of cloud only services or edge only services or things are only on functions that serve like serverless that'll be defined in the next 10 years to compute. And so that for you and I is going to be a space to watch. Jerry Chen, always pleasure, great insight. Great to have you on theCUBE again. Great to see you. Thanks for coming on. Congrats to you guys in theCUBE, seven years growing, it's amazing to see all the content you put on. So you think it's interesting, just last point is you see the growth of the curve, growth curves of the cloud. I'd be curious, John, to the growth curve of the CUBE content. I would say you guys are also going exponential as well. So super impressed with what you guys have done. So congratulations there. Thank you so much, Jerry. CUBE virtual, we've been virtualized. Virtualization is coming here. CUBE virtual, we're not in person this year because of the pandemic, but we'll be hybrid as soon as events come back. I'm John Furrier, your host for AWS, re-invent coverage with theCUBE. Thanks for watching. Stay tuned for more coverage all day. Next three weeks, stay with us. From around the globe, it's theCUBE with digital coverage of AWS re-invent 2020. Sponsored by Intel and AWS. Welcome back here to our coverage here on theCUBE of AWS.