 Live from New York, it's theCUBE. Covering the IBM Machine Learning Launch Event, brought to you by IBM. Now, here are your hosts, Dave Vellante and Stu Miniman. Welcome back to New York City, everybody. This is theCUBE, the leader in live tech coverage. We've been covering all morning the IBM Machine Learning announcement. Well, essentially what IBM did is they brought machine learning to the Z platform and my co-host and I, Stu Miniman, have been talking to a number of guests and we're going to do a quick wrap here. You know, Stu, my take is when we first heard about this and the world first heard about it, they were like, hey, okay, that's nice, that's interesting. And it's, but what it underscores is IBM's relentless effort to continue to keep Z relevant. You know, we saw it with the early Linux stuff. We're now seeing it with all the open source Spark tooling. You're seeing IBM make big positioning efforts to bring analytics and transactions together and the simple point is that a lot of the world's really important data runs on mainframes. You were just quoting some stats, which were pretty interesting. Yeah, I mean Dave, you look at, one of the biggest challenges we know in IT is migrating, moving from one thing to another is really tough. I love the comment from Barry Baker was, well, if I need to change my platform, by the time I've moved it, that whole digital transformation, we missed that window, it's there. We know how long that takes, months, quarters. Yeah, I was actually watching Twitter and it looks like Chris Mattern is here. Chris was the architect of Venmo, which my younger sisters, all the millennials that I know, everybody uses Venmo. He's here and he was like, almost all the banks, airlines and retailers still run on mainframes in 2017 and it's growing, who knew? And he said, you know, you've got a guy here that's developing really cool apps that was like finding this interesting and that's an angle I've been looking at today, Dave, is how do you make it easy for developers to leverage these platforms that are already there? The developers aren't going to need to care whether it's a mainframe or a cloud or x86 underneath, IBM is giving you the options and as a number of our guests said, they're not looking to solve all the problems here, it's that here's taking this really great new type of application, using machine learning and making it available on that platform that so many of their customers already use. Right, so we heard the, so it's a little bit of roadmap here. The ML for Z goes GA and Q1 and then we don't have specific timeframes but we're going to see Power Platform pick this up. We heard from Jean-Francois Pouget that they'll have an x86 version and then obviously a cloud version. It's unclear what that hybrid cloud will look like. It's a little fuzzy right now but that's something that we're watching obviously because a lot of the model development and training is going to live in the cloud but the scoring is going to be done locally is how the data scientists like to think about these things. So again, let's do more mainframe relevance. Trying to, we've got another cycle coming soon for the mainframe, we're two years into the Z13 and so when IBM has mainframe cycles, it tends to give a little bump to earnings. Now granted, smaller and smaller portion of the company's business is mainframe but still mainframe drags a lot of other software with it. So it remains a strategic component. So one of the questions we get a lot is what's IBM doing in so-called hardware? Now of course IBM say, well, it's all software but we know it's still selling boxes, right? So all the hardware guys, EMC, Dell, IBM, HPE, et cetera. A lot of software content but it's still a hardware business. So there's really two platforms there. There's the Z and there's the power and those are both strategic to IBM. It's sold, it's x86 business because it didn't see it as strategic. They just put Bob Picciano in charge of the power business. So there's obviously real commitments to those platforms. Will they make a dent in the market share numbers? Unclear, it looks like it's steady as she goes, not dramatic increase in share. Yeah, and Dave, I didn't hear anybody come in here and say, well, this offering is going to say, well, let me dump x86 and go by mainframe. That's not the target that I heard here. I would have loved to hear a little bit more as to where this fits into kind of the broader IoT strategy. We talked a little bit on the intro, Dave. There's a lot of reasons why data's going to stick at the edge when we look at the numbers. There's for the huge growth of public cloud, the amount of data in public cloud hasn't caught up to kind of the equivalent of what it would be in data centers itself. What I mean by that is, we usually spend say 30% on average for the storage cost inside a data center. If we look at public cloud, it's more around 10%. So we had at AWS re-invent, I talked to a number of the ecosystem partners that started to see things like data like starting to appear in the cloud. This solution isn't in the data like family, but it's with the analytics and everything that's happening with streaming and machine learning. It's a large repositories of data and huge transactions of data that are happening in the mainframe and just trying to squint through where all the data lives and the new waves of technologies coming in. We heard how this can tie into some of the mobile and streaming activities that aren't on the mainframe so that I can pull them into the other decisions. But some broader picture that I'm sure IBM will be able to give in the future. Well I mean you would think normally you would expect a platform that is however many decades old the mainframe is and after the whole mainframe downsizing trend you would expect there would be a managed decline in that business, meaning I mean you're seeing a lot of places. Now we've talked about this with things like symmetrics, it's you minimize and focus the R&D investments and you try to manage cost, you manage the decline in the business. IBM has almost sort of flipped that. They say okay we've got DB2, we're going to continue to invest in that platform, we've got our major subsystems, we're going to enhance the platform with open source technologies and we've got a big enough base that we can continue to mine perpetually. Okay great, the more interesting thing to me about this announcement is it underscores how IBM is leveraging its analytics platform. So we saw the announcement of the Watson data platform in last September, which was a sort of this end to end data pipeline collaboration between different persona engine, which is quite unique in the marketplace, a lot of differentiation there. Still some services, I've talked to last week at Spark Summit, I talked to some of the users and some of the partners of the Watson data platform and they said it's great, we love it, it's probably the most robust in the marketplace but it's still a heavy lift. It still requires a fair amount of services and IBM's still pushing those services. And so IBM still is a large portion of the company, still a services company. So not surprising there but as I've said many, many times the challenge that IBM has is to really drive that software business, simplify the deployment and management of that software for its customers, which is something that it's, I think working hard on doing and then the other thing is you see IBM leverage those platforms, those analytic platforms into different hardware segments, whether it's hardware slash cloud segments, whether it's Blue Mix, Z Power. So pushing it out through the organization and developing those, IBM still has a stack, like Oracle has a stack. So wherever it can push its own stack, it's going to do that because the margins are better. At the same time, I think it understands very well it's got to have open source choice. Yeah, absolutely. And that's something we heard loud and clear here, Dave, which is what we expect from IBM, choice of language, choice of framework. When I hear the public cloud guys, it's like, oh well here's kind of the main focus we have and maybe we'll have a little bit of choice there. Absolutely likes of Google and Amazon are working with open source but at least first blush when I look at things, it looks like once IBM flushes this out and as we said, it's the spark to start and others that they're adding on but IBM could have a broader offering than I expect to see from some of the public cloud guys. We'll see, as you know, Dave, we've got, Google's got their cloud event in a couple of weeks in San Francisco. We'll be covering that and of course Amazon, you expect their regular cadence of announcements that they'll make. So definitely a new front in the cloud wars, as it were, for machine learning. Excellent, all right, Stu, we got to wrap because we're broadcasting the live stream. We got to go set up for that. Thanks, really appreciate you coming down here, co-hosting with me, good event. Always happy to come down to the Big Apple. All right, good. Okay, thanks for watching everybody. So check out obviously siliconangle.com. You'll get all the news from this event and around the world. Check out siliconangle.tv for this and other CUBE activities where we're going to be next. We've got a big spring coming up, end of winter, a big spring coming in this season and check out wikibon.com for all the research. Thanks guys, good job today. That's a wrap, we'll see you next time. This is theCUBE, we're out.