 Kelly from IDC coming on, Cube alumni coming in. Again, we had her at HP Discover. We are down with a final stretch here and our special coverage of Moonshot HP's big announcement today. Michelle, welcome to the Cube again. Good to see you. Great to see you again. Step up to the microphone. How's that? Good day. How are you? Good to see you. We got to come to California to see each other. I know. It's crazy, right? How did you even get here? Do you have any power? No. Left Dev and the kids with a generator going. Good stuff. A bunch of gasoline. Me too. You're out of power as well? Actually, we have power. We're lucky this time. Yeah, Shondra kind of was just talking about power and how to harvest it from the crops. Yes. We were talking to Angelina about how much water you save with these new washing machines. So when you're out of power and you get your water from the oil, you get a health. If you want to flush the toilet, wow. It takes a lot of water to flush the toilet. That's true. OK. We are here inside the cube. I'm John Furrier with my co-host. Dave Vellante of wikibon.org. We're here with Michelle Bailey of IDC, vice president, and guru of the server marketplace. And pretty interesting announcements today, Michelle. I'm very excited about getting your take on this and figuring out what it all means. You were one of the first to really quantify the energy problem. In fact, all that department of energy data really comes from IDC data. And if it comes from IDC data, we know where it comes from. From your laptop. So take us back to some of the early studies that you did where you predicted, basically, that from an operation cost standpoint, energy cost, we're going to outpace any other cost, total cost of ownership. That was like several years ago, you predicted that. Yeah, we did that study back in 2007. And of course, at the time, the server-installed base was growing at an exponential rate. And much has happened since then, of course. We've gone through a lot of consolidation. And visualization has really taken hold. And it's really helped to level out what was becoming an enormous problem for the data center market. And so at the time, roughly. The crappy demand helped with the problem. Partly. Well, so you think about where the market has been. And the x86 market, as it evolved, it became unmanaged, basically. And it was just growing completely along with the demand around processing. So what we were predicting at the time is that the market had kept going the way it was, was that we'd be out of power. I mean, the data centers wouldn't be able to continue to grow. Fundamentally, in the US, you'd see constraints around businesses. And that's very much what was happening. We had a lot of enterprise customers who were out of power, out of space, and were unable to continue to grow their business. And so when you think about how much of enterprise businesses today depend on IT, I mean, it is the business for many of the large financial companies for example, even manufacturing now, you think about what have been traditionally sort of bit-based industries, if you will. So fortunately, there was a lot of really great technology that came along over the last five years. It was really, again, a perfect storm. Virtualization came along at just the right time. And again, a really good technology worked well and enabled a lot of consolidation. But we're in this really interesting period where the x86 market is maturing. You're starting to see lots of really great technology come out around the x86 platform around management and networking. And a lot of all these new fabrics that you're seeing coming out are very much built around that technology. So the market's ready for a new disruption. We see this every 10 or 15 years. And we've sort of been waiting for something like this. And that's why there's so much interest being paid to the announcement today, because it feels like the time is right for this. So x86 has, what do you figure say, 97% of the marketplace for server? At least, right. It is the market, right? So I think what's really interesting is the parallel here where you look back when x86 really took hold in the mid-90s, that was a chip that came out of PCs and was running toys, right? It was running consumer applications and made its way into the data center and became a robust operating system as well around x86 servers. And it became serverized and it had all the things you needed like IO and ECC memory and all those kind of good things. And I think we're seeing a lot of parallels today around this new technology. Yeah, so people have been very careful today about making bold predictions about that this is gonna take over the data center, but if you use that PC analogy back then, I mean, who really would have predicted that Intel would have had 97% of the server market? I think nobody. So today, you do a lot of work on workloads in use cases. So today, it's, I don't know, what percent of the market or even candidates for these low power? Well, not a lot today, right? Less than 10%? Maybe even a little less than that. But I think what we have to keep in mind is this is not really about today's applications, it's about the applications for the future, right? But there is demand though today. I mean, you're seeing specific use cases like big data and cloud. Absolutely, right? And these are really the applications that I'm thinking of as being future applications. So, I mean, it's so exciting what you're seeing going on around big data and even around convergence, right? You're seeing all of these very large specialized systems being built to take on these very large workloads at the top end of x86. So what's fascinating about the point of time that we're in right now is we're seeing maturity and a lot of excitement around the evolution of a market that's been around for a long time. And at the same time, these disruptive new products come along that make us think about a very, very different future, right? And so it really does feel like this next wave or this next business cycle of IT is what we really call it. Yeah, we're going to be at Hadoop World next week with theCUBE. Right. You know, I don't know if you're going to be down there. I won't be at Hadoop World. It's too bad because we would have you on again. I'll be watching you. I'll be watching you. So the question then is, are these new emerging applications, are they outliers? Are they sort of the lunatic fringe or are they harbinger of the future of the data center? I think it's the future, right? I think that if you look at our generation and the amount of data that we've created, a lot of it fairly useless right now because we don't do any analytics on it, right? We don't really harvest it. It just becomes a lot of very static data. And I think that if you look to the future and where the excitement is, it's around taking all of that data that we've created and being able to do analytics on that data and using IT in a very, very different way from how we've been really using it in the last 15 years. Are people talking about monetizing that data? We're talking about it. I think it's already being monetized, right? I mean, look at Facebook, for example. Look at Google, for example. I mean, they're absolutely monetizing that data. I mean, what do you think their market cap is? It's not my daughters. So then the follow-up question then, is that the future competitive advantage in our business? Well, I think so. I mean, I think that we've seen it. Faster CPUs, I mean, come on. Well, I think that we've seen it for a long time in the enterprise, right? I think we've seen it in the financial industry, particularly, and I think that that's, you know, you think about a lot of all these new markets around retail, for example, and understanding supply chains, and I think one of the first examples we've actually seen is in the data center itself, being able to take a lot of information about systems in the data center and doing analytics on that and creating better data center footprint out of it. I mean, you see it again and again, I think in many different ways. So what about, we like to talk about horses on the track when we were at Open World, Oracle Open World, about a month ago, and David had us, we had the future of the data center, we had a lot of microprocessor, secondary at winning by 31 lengths, that was the x86, right? That was Intel, we said Intel's good. We had IBM Power and titanium, sort of, you know, we were back in the pack, and then David made us put arm in the race. Yeah, good. And it will fall, right? Smart man. And he said, trust me, it's gonna be in the data center, so that was good, good call, good prediction, but, and we had Pauline Niston from Intel asked her about arm, and she said, well, we have Adam. But you know, you don't see Adam at this announcement today, you see, you know, arm and Calzada, and what's your take on that? Where's Intel and all of this? I think that they're gonna be formidable player here. I mean, I think that this is probably gonna force them to make some pretty significant announcements around Adam, I would believe. You can't just say it's gonna be arm, right? This disruption can come from anywhere, and it can be all range of different products, and early on in disruptive markets, you see a pretty lengthy battle, usually over different technologies, so absolutely they're a player, and there's no doubt about it. A lot of cash, a lot of revenue, a lot of big ecosystem. A lot of ability to make a lot of similar conductors, absolutely. So we can agree that arm has a lead today, I don't know, if we can care to quantify that lead, do you have any sense of it? I have not quantified that today, I'll get on that for you next week. David Flores in the audience, we could ask him. And David might know. What do you think the lead is, arm? 18 months, 24 months? We just want a quick answer. We don't need you to come on the queue, say where you are. We're at 18 months. You think 18 months? Yeah. Yeah, okay, so. Going back to your issue about consumerization, arm is driving that, mobile is now more greater than the number of years. So 18 months, that's pretty good. 18 months, that's good. The volume that's coming from mobile is really the ace in the hole that arm has. So Intel has to repeat that. And that doesn't come from x86. Of course, they got a lot of, they got a big war chest that they can use to subsidize this. Oh, DOG, Jane might be watching, hold on, be careful. But it could even buy some of these players to get in and accelerate. So that 18 month lead, that's not insurmountable, isn't it? It's significant enough. I mean, getting deals done and getting product to the market, it could be green, it could be interesting. So it's interesting times, right? Now HP is leading the charge here. No longer do you just innovate with microprocessors and operating systems, right? What's your take there? I mean, HP's first, right? What does that mean? Is that just continued server lead? I mean, the number one in servers, but how important is that being first and how are they adding value beyond the microprocessor and the operating system? Yeah, I mean, I think what you saw a lot in the announcement today was, servers not just a server anymore, there's a network fabric involved there, there's a management chip involved in that server. It becomes a much more holistic system. You can start to even think about impacting many, many adjacent technologies in the data center, particularly networking, innovations around memory, right? So I think you need to keep in mind, the types of applications are gonna be written to this new type of capability. It's not gonna be everything in the enterprise data center. You're gonna be looking at things that are typically IO constrained where you can have more of a parallel processing capabilities take hold there, and that's not every enterprise application today. Are you tracking the app impact to this? I mean, honestly, one thing we've heard on the IT consumerization trend is you got all these kind of Ruby framework type developers that aren't really hardware geeks. So what's the impact to that big spend that's going on on the app side? Everyone wants an app store for the enterprise. We heard SAP talking about it, others, and now they got homegrown apps so they do it internally to the outsource of kind of these new dynamics. What's your angle on that? So this is what I think is interesting about HP's announcement and any other announcements we're gonna see here is that you've got this whole new revolution of application development taking place at the same time. You're seeing a lot of the change in the systems come to market. And if you really look at a lot of the new applications that are being built out there on these new frameworks, they're actually right for more of these parallel types of technologies. So again, it's this perfect storm of a change in application development along with the change in the systems could really lead to a very, very different data center future than what we see today. At the same time, everybody's got 30 years invested in what's going on in the enterprise data center. There's a lot of Rip and Replace coming down the pike, don't you think? Well, Rip and Replace takes a long time, right? So again, it's another 10, 15 year cycle and we're a year one or 18 months, as David might say. It's like a land server to go back to our PC now as we were talking about PC, spun the wind tail server. And you know, I was laughed at, I was, you know, who's gonna file sharing big deal? That morphed into blades, right? So here we got mobile driving now this product. So it's massive. Absolutely. And you know, if you think about a lot of these mobile applications are written on an ARM platform as well, right? So that has a very big impact there. Yeah, code portability, we're open source. Yep. Hadoop, for example. Absolutely. Has a growing Apache project ever? Yep. I mean, it's just, it's fantastic. What's going on in the market? So you gotta think about, you know, it's not just these large providers that are building these applications. You know, if you think about what the impact is on the enterprise data center, there would be a lot of rewrite that would have to be involved in going back to some of those legacy applications. This is really much more about, you know, if I want Hadoop within my own data center, perhaps this is something I think about rather than traditional platform. So I wonder if we can go back to the horses on the track, cause we'd love to talk about that. I know nothing about horse racing, but go ahead. So HP's got the leading share of the server market, about a third, 30%. Is that about right? Yep, they're usually in a pretty tight race with Dell, the other horse in the race. Yeah, so Dell, IBM, Oracle, I guess Cisco's on the track, right? They're definitely on the track. Okay. What are the, so again, in this world of, you don't necessarily have to own the operating system or own the microprocessor, what are the characteristics that have allowed HP, really without ownership of those core elements, even though it's got some sort of, what are the elements, the characteristics that have allowed them to achieve that 30% that lead and how do you see that shaping up going forward? What are gonna be the attributes that determine success and leadership in the next 10 years? Well, in terms of their lead, they were aggressive in the market for a very, very long time, right? And they were very, very competitive, particularly against Dell in the x86 marketplace. And I think you saw a shift in, or mine shifted, if you will, around HP, where they would get the kinder, gentler company and then they became pretty aggressive and really drove a leadership position in terms of the number of shipments that went into the data center. And I think you've also seen that they've made a number of acquisitions over the years as well that have done very well for them, mainly around the system space. So particularly around their storage environment and being able to integrate their storage into another networking environment as well. And that's really, I think, changed things for them. So that system view, actually, that the whole converged infrastructure? Well, you know, it converges, again, really the maturity of the x86 platform. And if they hadn't done a lot of those things, they wouldn't have evolved along with the market. And so it was actually very necessary, a lot of those moves that they, absolutely competitive advantage in the marketplace. And you're seeing a lot of the other vendors in the marketplace do similar things, right? So Oracle, for example, is another great example. Dell is making a lot of these acquisitions as well. But that's about maturity or evolution, right? This new disruption, this is something entirely different. This is going off on a completely different plane and starting again. So they've got to be very careful to manage that whole portfolio in terms of managing their existing assets and making sure they capture value out of what's been a very important market for them while at the same time managing these new disruptive products. So basically, do they cannibalize themselves for the growth market? Well, that's what they're holding on to. They're holding on there. So I'm not so sure if it's cannibalization at the moment as it is making sure they're in place for the future and future applications. They're certainly painting this as incremental, but there's got to be some kind of overlap there, right? Yeah, I mean, at some point, once, you know, products like iron-based products become, you know, more mature, there becomes overlap, right? It becomes very difficult to decide what you sell where. You want to be in the position to cannibalize your own as opposed to having to sell your own. You better eat your own dog food, right? You're too young for someone else to do that. Yeah, no, it's very true. It's an exciting, exciting time. I was just saying before, this is one of the most exciting announcements in the server industry in 15 years, right? We really haven't seen this kind of change. Pretty incremental, kind of boring. I hate to say it. Good stuff, though, right? You did a good job. The blades were great. I mean, the blades hit that mark, you know, where it's like that glass ceiling. And then this brush... The blades were kind of cool, right? Makes sense. It wasn't really... That was the beginning of what you're saying, right? You agree? Absolutely. It just feels like a major, major disruption and a big new wave. I mean, we heard this morning, you know, the four big ones, X86, blades, pods, and now this. So this just feels much, much bigger, in my view, anyway, than blades and pods. Blades, big, you're right. But a lot of packaging, this is the step function. Order of magnitude, more exciting. I think so. And I think managing that ecosystem and building it out over time is going to be really critical. And, you know, HP has big pockets to help enable that. And for me, that's what's really exciting. When you see some of the very larger vendors getting involved, then you know that the ecosystem around it can begin to develop. And as a customer, you want to make sure you can get service and that, you know, you're going to know the product's going to be there a couple of years from now and all those kind of things. The energy story and the performance story is just too compelling. You just cannot deny it. Is that the big execution risk in your view? Is the ecosystem and building that out? I mean, there's lots of them. Undoubtedly. I mean, if you don't get that right, then nothing really works. And if you do, does that become a sustainable competitive advantage or are these guys just arms dealers that are going to go wherever the business goes? No pun intended. Well, you know, they're going to have to be careful about how they build that out and make sure that they're not pushing out at the players, right? You want to grow the ecosystem. You want to enable everybody to win there to get some volume behind it. So that's a fine line to play to. You want to own it, but you don't want to own it so much that nobody else can play there. Yeah. Well, I mean, you see, you know, what VMware, for instance, is doing in this ecosystem. It's good to see HP playing that card again. Yeah. You know, in a fairly big way. Yeah. I think so. What's your take on the disruption to just basic industry metrics? I mean, her pulmarits at VMworlds saying things like virtual machines are out shipping physical servers. That was my day. You're numbers again. And someone, someone really smart. And so now you have, I mean, how do you count servers? I mean, they got four, I mean, 280 processors. So you get a collapsing of it. Now, remember, the network things, ports and this and that, I mean, how do you measure markets here now when you got 280 servers? It's a great question. For you, Box. So we're looking at being able to count the number of VMs, which is difficult, of course, because a VM can live and die within the matter of an hour, right? So it's very, very difficult. It becomes a market that isn't static in the way it's been over the last 15 years. So it requires a really different set of metrics. But what we do look at really now is how many VMs does an enterprise environment host? And what's the associated management cost with that? So one of the things you've got to be concerned about with ARM is it proliferates and then suddenly you've got this huge management problem on your hands again, right? You've got this again new, you know, massive environment that you haven't probably put in place the right management tools around. So that's a huge concern. Partho was talking about the ensemble concept where, you know, bringing in a new management layer, things like that. Embedded within the technology, right? Absolutely. So you basically put a frame around the market in a different way and kind of try to extrapolate the market share. So when they get into the who owns what market share, it's going to be interesting. It's a great opportunity to restate numbers this virtualization thing, isn't it? It's always a great opportunity. New survey. Things change very dramatically. You know, the market is different, though. You know, I think back when I started at IDC almost 15 years ago and what we had to do then, you know, we were just counting boxes, right, when they were coming out of the back of the door. So easy. How easy was that? We thought it was so hard. Yeah, well, of course. It's very different. And then, of course, you know, you've got server devices that look like storage devices. So, you know, when does it become a server and when's it really a storage device? And nano store coming around the corner, you got crazy stuff. And convergence complicates things, you know? Convergence, absolutely. With the UCS and the whole war and all that fun stuff. Yeah, I mean, it's really pretty simple. It's really, you know, what's happening on sort of the maturity of the x86 market and convergence and private clouds and they all really fit on that axis. And then you've got this new axis starting around these new disruptive products. And, you know, there'll be different metrics to measure this market. It won't be the way we've measured the old market. The value proposition is compelling when I saw, you know, four servers on a small little card. That's compelling. It's pretty wild. It's very compelling when you run in these kind of big data things we're doing. Like we're doing. Very data driven. I mean, yeah, I mean, I just love to have that. I can have 16 servers. That would, that could power our entire site for two years with our big data project. Right. Just, I mean, two little cards. Right. That's, you know, and then they go to scale. They'll look at Facebook and these guys. So, you know, they want cost and they want performance. Well, you know, it's some, you know, when you're an enterprise customer and you're looking to that next technology, we always find this magic number of 25 or 30 percent. You've got to be saving me at least 25 or 30 percent. 10 percent, it's not worth my while to have to go out and make the change. You know, if you start looking at price differentials around 50 percent, then it's very hard not to look at that. Yeah, well, what's the numbers today? No, you know, he said 10 data centers down to a half. You've got to dig into this, but still 89 percent less energy. That ain't 10 percent. Yeah, I mean, for very large data centers, you know, if you're thinking like a host or someone like that, that's huge. Yeah, 94 percent less space and then 63 percent less cost. I presume that number's lower than the others because you still got, you know, operational costs associated with people, but... That's hard not to take a look at. Yeah, I mean, if you put that in front of the CIO, they're going to go, I'll listen. Yeah, sales guys got me a meeting. No problem. Sales, we could save you 12 or 13. Particularly, well, particularly if it's for newer applications, right? When it's an existing application, you've got all the migration costs and all of the risk that goes along with that, right? Then you've always got to factor that in. When it's for a new application anyway that you're going to build, then it's something you really should take a look at. Yeah, and the Clodera and the Hortonworks, these guys out there, punched in the big data. Their whole pitch is, you don't need compute. You've got plenty of compute. Your issues IO. So years come out of the year. So this fits right into that wheelhouse because now they have the ability to manage processors up and down and manage the IO. So I think HP will capture a lot of that business. It was interesting to see Cantor Fitzgerald on the panel this morning, right? I mean, you know, they're a real-time application. I mean, they're data attention. Yeah, I feel like he did a nice job to describing how high-frequency trading isn't just one application, right? It's many different kinds of applications within that. And even if it isn't the platform for the entire app, you can carve off pieces of it and drive efficiency there. Yeah, it's the big data piece that's driving their interest in lower energy processing. And so that's, I mean, it seems to be a trait of this marketplace is that you got maybe a single or maybe a couple of applications that are really driving. Look at Facebook, it's really one app to a million people. Yeah, it's a very different approach than the classic enterprise, it's very true. As we know, our days of application portfolio. Absolutely, very different approach. It's hundreds and thousands of apps for maybe thousands of users. Yeah, well, if you have that kind of a data center, then you're always looking for the next disruptive thing, right? You have to, I mean, do your power and cooling costs or affect your bottom line? You have to constantly be thinking about that. Data center's an ATM, I've been saying all the time. As opposed to a cost center. Okay, I'm getting it. It's the holy grail of IT value. Yeah. And it's printing money. Yeah, it's true. It's the profit engine, it's absolutely true. And the car was full of crap. IT does matter. Sorry. It worked out for you though. He's got a lot of books. He's got a lot of publicity out of that. He became wealthy with a failed premise. Very much, it's not so much the IT, it's what you do with it, right? And again, for Can of Fitzgerald, it wasn't so much that they had this really cheap platform they could put in places that they could do very different things with a very important application for their company. And that's what's really driving the value there. Well, I think that's a really important point. It's what a customer is doing with IT that's more important than the new operating system or even the new chip. Absolutely. The application is really where the innovation is today. Yeah, I mean, I think what he was saying was, they had reached a point where they couldn't continue with the existing technology the way it was and make it cost competitive. And so that's why they're constantly looking for something like this, right? Because that drives new business for them. Very important for them. All right, Michelle. Well, listen, really appreciate it. Michelle, great. Thanks for having me. Thank you. Always a pleasure. Did I close it up? I know. Okay. Great guest, as always. We are here at a historic place behind us as the offices of the founders of Eulah Packard here in Building One in Palo Alto, California. We're live special coverage of HP's Moonshot, Project Moonshot Announcement. I'm here with my co-host. I'm Dave Vellante, wikibon.org. Thanks very much everybody for watching today and stay tuned. We'll be at Hadoop World next week. Look for... New York City. Yeah, we're in New York City. Tuesday, Wednesday of next week. Look for the coverage of this event. It'll be up on our sites on wikibon.org on siliconangle.com. Look for that. It'll be up this evening and all these years. Videos on siliconangle.tv. And again, Hadoop World next week should be a big event. We'll be on the ground, Twitter, Facebook, stumble upon Google. Everyone's gonna be there. A lot of financial services. A lot of government talking big data. We'll have two days of eight hours a day coverage from New York City. So siliconangle.tv. So that's a wrap from our special coverage of HP's big moonshot announcement, game changing. Thanks for the great sound bites from Michelle Bailey IDC calling it a once in a lifetime generational shift that we're seeing, bold move, big growth, new metrics. Dave, thanks so much. All right, pleasure, John. That's a wrap. See you soon, everybody.