 from San Francisco, it's theCUBE. Covering Micron Insight 2019, brought to you by Micron. Hi everybody, welcome to Pier 27 in San Francisco. My name is Dave Vellante and I'm with my co-host David Floyer and you're watching theCUBE, the leader in live tech coverage. This is our coverage of Micron Insight 2019 hashtag, Micron Insight. David, I love this show because, well, of course we're going to talk about Micron and Memories and DRAMs and NANDs and all that techy stuff. We're also going to sort of set the tone in this day. It's a really thought leadership day and we talk a lot about AI and Edge and the big mega trends and superpowers, the cloud, mobile that are really affecting demand and it all starts with data. So Micron is a company that we're going to talk about and talk about in detail. But what are you seeing, David, is the big trends that are driving demand for bits? For bits? Well, let's start with the Edge that you were talking about. The Edge is growing and it's going to grow very, very strongly indeed. It's going to grow with smaller processes. It's the ARM processes at the Edge doing inference processing, capturing the data and wanting to do that capturing of the data and the processing of that data as close to the origin of that data as possible. So memory and all of the NAND is moving out to the Edge itself and it's going to be lots of smaller processes as opposed to the lots of big processes. So there's your question. So we've been following these markets for many, many years and of course when we started in the business it was all mainframe and that was really what drove the consumption of data and then the PC changed that. And then that, you used to count markets, we used to do that all the time and there was much more data going to the laptops and desktops. The internet began to change that and of course cloud sort of re-centralized a lot of the spending and a lot of the buying power. Do you see, is it a pendulum swing again? Is it that dramatic or do you see it as different? Like all big trends, the center still remains. So the center now is cloud, still mainframes as part of that cloud. That has to remain and that is just much more economical for large scale processing. That's the most economical. However, also the economics of it is that moving data is very expensive. It's very expensive in terms of the effort and it also, when you move data you lose context. So if you want the best context and if you want to do things in real time you want to process that data in real time as close to where it was produced as possible. So yes, there will be a very big swing on in the amount of processing and the amount of important processing that happens at the edge. So from the standpoint of things like NAND and flash Steve Jobs changed everything when they decided to put flash inside of the iPhone and iPod actually. And that drove massive, massive, that was the beginning of the dam breaking and what happened is that volumes went to the roof, costs went down and that's really when you first predicted way back in the early part of this decade that NAND and flash would affect spinning disk and it clearly has, pricing maybe hasn't come down as fast as we thought because of supply constraints. But nonetheless, it's happening and now the prices are coming down more. You've seen somewhat of an oversupply in NAND but prices have come down pretty substantially and there's elasticity. I mean ever since we've been following this market you've seen when prices drop people buy more. You know at the same time you saw like Pure Storage last quarter said, well the prices drop faster than we thought and actually hurt our revenue because it just happened so fast in the middle of the quarter that it hurt pricing overall for the subsystems but nonetheless that's the trend that we see happening. It feels like there's a new wave or a new step-function of consumption going on with regard to flash. What are you seeing? Yes, flash was always about performance before and there were two constraints to flash in terms of its impact on the whole industry. The first was that the protocols that were used in flash were the old-fashioned protocols that were used for HDD. Now those have improved enormously with NVM, NVME, et cetera and those have got much, much better. So that increases the demand for that flash. The usefulness of flash is now much better. And the second is in terms of that's high performance. There's high capacity flash and now flash is growing in two dimensions. It's growing in the number of layers but it's growing from SLC to MLC to TLC to QLC in terms of the number of bits that it can pack into it. So those all have cost implications on the cost per bit, obviously. Sure, so both of those are reducing the cost per bit and making it available for different markets. So the capacity market now as the prices come down mean that it's going to take a bigger bite into the HDDs. And in data center it's going to become the norm just to have flash. Now Micron's a little bit late to NVME but they're now hopping on board. Actually you've made the comment to me in previous discussions that they've actually timed things pretty well. You kind of didn't want to over rotate to NVME. I mean I know Pure was first but Pure's a relatively small part of the marketplace. It seems like now everybody's going to NVME and basically what this does is that you pointed out it eliminates a lot of the sort of older slow overhead chatty protocols and now it's like a bat phone right to the data. So what are you seeing from terms of NVME adoption? Is it now mainstream? Yes, we're predicting that in 2019 a 50% of the drives will be NVME drives. That's a very rapid change. So let's up level it a bit. We're talking about all this geeky stuff down here but what I'm interested in is why we need this. And obviously the obvious question is there's so much more data now but it's also AI. We talk a lot about the new innovation sandwich of being data plus AI plus cloud combine those things together and that's really what's driving innovation. How real is AI? I presume we need all this stuff to be able to support these data-driven workloads but how real is AI? I mean it feels like it's pretty substantive. I mean we've got a lot of these shows you hear about digital transformation and all these buzzwords and the edge and IoT. Of course AI is one of the big buzzwords but it does really actually feel like a superpower to invoke one of Pat Gelsinger's words. It is and AI could only operate if there was all that data available. So it's the availability of that data because the algorithms in AI go back a long way. There's nothing new in that. But AI has now the availability of processing that data, a large amounts of data which makes it much more powerful. And now you're getting AI and things like a cell phone. The amount of AI that goes into recognizing your face is enormous and it's now practical everyday things are being done in AI and it's going from being a niche to being just everyday use and it's impact long-term is profound. I mean it'll do all the jobs that humans do, many of the jobs that humans do much more efficiently, driving a car. It'll be better at driving a car than human beings. Yeah I mean you see AI everywhere. You're right, I mean ad-serving still stinks but it's getting better. Fraud detections getting much, much better. Email is now finishing my sentences for me. You've noticed that in the last year or so. Basically say I like that choice, boom I'll take it. And so as much as we hate autocorrect. So those are some small examples but what the industry likes to talk about is how it's changing lives, what it's going to do for healthcare, autonomous vehicles, those are some of the big picture items which really haven't kicked in yet just in terms of, or have they in terms of consuming demand for things like DRAM and NAND. It's relatively small at the moment but it has the potential to be very large obviously. Looks good, I'm going to finish your thought. Because in the next 10 years we're going to see automated cars. It's going to be in pieces. You're going to have the trucks going first and then other cars later. So I know you're fairly sanguine and optimistic about autonomous vehicles. I know there are a lot of skeptics out there that talk about, you know, we don't have enough data and we'll see but we'll talk more about that. I want to talk about Micron a little bit. Micron's a company, last year they were a $30 billion company. They had $23 billion in revenue this year so dramatic drop in revenues and that was really due to the change in the supply demand dynamic. Now historically when these things happen the stocks of these companies would just, you could predict it. You say okay time to sell because here comes the oversupply and then when you hit the bottom of that time to buy Micron's done an amazing job of sort of steadying that. Managing its demand and supply balance. Also, I mean obviously doing share buybacks have helped the stock price but the stock price has held up pretty well. So Micron's now a $23 billion company. Last year they threw off $17 billion in free cash flow this year, $13 billion. But still well over 50% of their revenue is going back to free cash flow which is quite large. Their market cap's 51 billion so they're trading at a 2.2 X revenue multiple which is very strong. And they got a 30% gross margin, right? I mean the PC business, you think about that. So the DRAM business is a good business, right? I mean that's a nice business because they don't have a giant sale, direct sales force so they don't have that cost. It's all through OEM. So it's a fairly efficient business and they've managed it pretty well. Your thoughts on Micron as a company? They have, they've managed the timing of every new release very well indeed. If you go too early, you over rotate then you are struggling to get that out. The costs are higher and the people who are selling the previous generation are going to do better. But they've always timed it perfectly. Yeah, now they've efficient some challenges. I mean we talked about the supply-demand imbalance but they're managing that. China, you know the tariffs are at them. Huawei was a big customer, they can't sell to Huawei anymore. China coming after companies like Micron, really going after consumer flash, building fab capacity and then to begin with and then eventually China's going to aim at the higher value enterprise. What are you seeing there? I agree with you. They've had to rotate because of this problem with the tariffs that have been put on China. So what's the reaction they're going to have to invest? And that long term is good news for consumers and good news for everybody else but it's going to be bad news for other people in the business. So a bunch of announcements today, we can't talk about it because they're not public yet but you're going to see some SSD stuff coming out, maybe some acquisitions announced. You might see some other things around 3D cross point which is something that we really haven't talked much about but we will. I know your thoughts on that or it's still kind of niche-y, you remember the HP memraster, right? Which is, nobody talks about that anymore. But now Micron's in a different situation. They'll figure out where that fits but it's still a niche in your view because it doesn't have the volume but we're going to be talking about that stuff. But again, up leveling the conversation to some of those big mega trends, those super power drivers, data, AI, IoT and the edge and some of the things that are really driving change in not only industry but also our lives. David, appreciate the insight. David and I will be here all day today so you're watching theCUBE from Micron Insight from San Francisco. We'll be back with our next guest right after this short break.