 Hi everybody, this is Dave Vellante of Wikibon.org, and there's big news today. Fusion I.O. is being acquired by Sandisk, and we're going to break this down. Joining me is David Floyer for his breaking analysis. David, thanks for coming on. Hi there, Dave. Yes, exciting news, isn't it? Well, so Sandisk is about a $6 billion company with a $23 billion market cap. Fusion I.O. is about a $430 million company with today a $1.23 billion, $1.3 billion market cap, because of the announcement that stocks up about 25% on the day. At one point, Fusion I.O. was trading an evaluation of over $2 billion, and the company was very narrowly focused, a lot of its sales were focused on big, hyperscale giants like Facebook and Apple and the like. So the business was kind of lumpy, wasn't it, David? Very lumpy indeed. And there was a downturn in sales, which hit them hard. Their long-term ambition is always to be the memory company, to regard flash as an extension of memory in the server itself. But they never really got there. It was always treated as disk drives and very lumpy sales, as you said, at the high end. They never got the volume to get it into the general marketplace. Yeah, so as I said, there are stocks trading up today. It's been hovering pretty consistently up over 11, 11.4 roughly at a nice premium, about a 23% premium to where it was at the Friday's close. But not an enormously disgusting premium, David, I mean does this signify that potentially somebody else could come in and who else might be interested in Fusion IO? Yeah, this is, as you say, it's not a high premium. You can either look at it from two points of view. Who else as a flash provider would be in the market for this? So people like Samsung, people like obviously Toshiba is a partner of Sandisk. People like Micron or people like Intel, those could look at it from that point of view and look at it as a way of getting the software stack to make, to add value to the very up and down market price that is flashed. And if you look at what happens when a new version of the, when a new smaller size comes out of the flash chips, the prices crash for a bit and then they stabilize and even go up a bit and then the next one comes and they crash again. So it's a very difficult market to be just the flash supplier. And obviously, Sandisk wants to be in that market and other people that could buy it would be similar people. Samsung, I think, has more aspirations. They've tried already to buy Sandisk. Itself, so that's a possibility later on. And the other guys, for me, Intel would be the natural one because they're closest to the server and could use that to create a very good product. So you don't see any of the server guys or do you see any of the server guys interested? I was just going to say that the other group of people could be the server guys, IBM would have been a natural, but they've sold off everything to Lenovo and I'm sure Lenovo doesn't have the capability of buying them at this stage, so they're out. Well, let's be clear on that. So IBM sold its x86 business to Lenovo and I'm inferring from what you said that that's really where FusionIO would fit. So when you said everything, that's what you meant. I mean, in theory, they could try and apply to other things, but it's the 86 business that matters. No volume, yeah, right, OK. So who else? So HP, I can't buy anything and they are fixated on Memoristers. What's very interesting to me is Sandisk announced a month ago, the 15 nanometer flash coming out later this year. There's 128 gigabit flash chips. Isn't that amazing? So Memoristers are nowhere near that level of density or cost or anything else like that. So HP should be thinking about it, but they can't. And that leaves people like Dell and they're not really in that business. And who else? So you say HP can't because you're saying that Meg is still paying down the debt and she has said we're not going to do any big acquisitions until we've paid down that debt, right? That's your point there. That's right, yeah. And of course that HP Discover, Martin Fink put forth a vision of the machine, which was a large shared memory system, but you know, distributed cluster of nodes based on Memorister. And I had tweeted out and David, this is your point all along, is that they got to compete with the price of flash. And that's going to be very challenging them for them to do that. Longer term, it may have some potential. We'll see. There are people like Jean-Luc Chatelaine, a CTO of Data Domain, who's very high on Memorister, former HP employee, by the way, just for full disclosure. But David, you've been saying that really flash is going to be here to stay even in the enterprise for quite some time. Flash has two things. It has the consumer market underlying it. So all of the 15 nanometer is driven to a large extent by the consumer market and the flash for the enterprise and the flash for, you know, big data and things like that is a small proportion of that total marketplace, whereas Memoristers are only at the top. So, yeah, it's flash is coming down. The idea was that Memorister was going to come in or other things like IBM's PCM technology was going to come in and take over from flash. But flashes continue to come down. The roadblocks that people foresaw in it have been overcome by smart people. All right, David, let's let's break this down in a little bit more detail. Let's look at Fusion IO. Fusion IO, as we said, specializing in the hyperscale business was the high flyer at one point in time, some lumpy business. It came back very strong after the first downturn when Facebook and Apple and other big hyperscale guys put the brakes on. Never came back quite the same way. A second time around they had to make management changes. Shane Robison came in ostensibly to beef up the sales. Well, this is one way to beep up the sales. It's to sell the company to somebody with a larger channel. We were very high on Fusion IO, in particular because of their technology, their software technology that allowed primitives to be written directly to flash as an extension of memory, eliminating not only the spinning disk, which all flash does, but eliminating the disk protocol, for example, SCSI protocol, which is very overhead intensive and chatty. We saw that as an opportunity to completely change application development application design. It never happened. Why not? Were we wrong about that? It's taking longer. I mean, it is happening. You'll see you're seeing now a whole set of announcements coming out from the database vendors. MySQL now supports the atomic rights. That's eliminating the SCSI overhead, both all three, Pocona, Intel, sorry, Pocona, Oracle and MiriaDB. Very interesting announcement from Microsoft. SQL 2014 has got some very nice flash-aware technologies built into it. So the database vendors are coming out with that. And the ISVs, the people who are designing the systems born in the cloud, are starting that type of design. But the whole cycle is a longer cycle when it goes to when it comes to the database design. That's a 10-year cycle, as opposed to the two, three-year cycle for technology or a five-year cycle for technology. That's going to be a longer. But I still believe that's where the future is. And SanDisk, if they get into that marketplace and helps support the applications in a different way, support the operating systems, for example, putting all of the operating system into flash and being able to get that at lightning speed, support the databases, being able to do the same thing, very large memory databases supported by flash and then support a completely different sort of design where all way you can do transaction processing, whether you do analytic processing and get rid of that 40-year dinosaur, which is the disk drive itself. OK. So I tweeted out, actually, this morning in reference to the sort of alternative bidders, doesn't everybody want, you know, atomic rights to go compete with DSSD from the EMC just acquired? Howard Marks actually had an interesting tweet. He said, at DeValente, when T10 approves atomic rights, it becomes standard operating procedure. And then the other piece of information he shared was PCIE flash is an OEM market and flash foundry vendors like SanDisk are in the catbird seat. We'll talk about SanDisk in a moment, but he's talking about a standards committee, T10 approving atomic rights and then NVME becomes a standard. What are your thoughts on that? Well, there's a lot more required to get it done than getting the T10 standard. You've got to have a stack itself. You've got to have solid software which interacts with the controller itself, very tightly coupled. That's what SanDisk are getting here. The other very interesting thing that SanDisk has come out with, they bought a smart storage and together they have the capability of putting the flash onto the DRAM on straight onto the memory bus. They've come out with a product which does that. The smart storage allow them to make it look like a SATA disk, which is interestingly a quick way of doing it, but it is not the most effective way of doing it at all. So bringing together the FusionIO software, putting that as an alternative onto the memory bus directly and providing a soup to nuts ability. The T10 is the standard from which a programming standard, but underneath that to provide the controller software, the Linux software, FusionIO have a lot of patents in that area, not a patent, a lot of know-how and actually patents as well in that area. So there's a lot that SanDisk can do to be the leader in that market and that's a much higher margin market than the traditional flash market, the NAND chips. So I think it's a good move for them. And if you look at storage as a whole, we've said consistently it's going away from the sand towards the server and within the server it's going towards the memory itself. This is where a lot of storage, a lot of change is going to happen and the master copy of data is going to be in the server itself. Okay, so let's talk about SanDisk and from their perspective, you touched on that a little bit. SanDisk is known for thumb drives. As I say, it's growing very nicely, growing 22% a year growth rate, $6 billion company from a revenue standpoint, nice valuation at over $20 billion. To me, this is about SanDisk expanding its total available market. I mean, we've seen SanDisk at the various shows. We saw them at Oracle Open World. They were just at HP Discover last week. They had a booth there. They had a presence there. They're hiring people that understand the enterprise. Isn't this about a TAM expansion move for SanDisk? Absolutely. And I think if you're looking for the next EMC in storage, it's going to come from the Samsung, the SanDisk, the Toshibos of the world coming down into that marketplace with most of the added value being in the controllers, in the software at the heart of it. So that's going to be a very different model to the current model where the expertise was in how to manage slow disks, how to manage the very slow transfer of data from very fast processes down to those disks. So I think it's the start of a profound change in the storage industry. Very interesting. You're saying a new breed of storage vendors is emerging. You're seeing all these tectonic shifts from the SanDisk to the server. And that's going to cause some disruptions to the existing storage companies. Will they all make it through the knothole? No, clearly. These things are slow to start with, as I've said in my forecast on server sand, for example, they're very slow to start with, but they are going to accelerate. And over the next 10, 15 years, we're going to see a profound change as the disk drives become just very standard, low-cost, low-access products. And all of the action, all of the invention goes towards the flash end of the marketplace and is owned by essentially the flash providers. All right, David, thank you very much. Great breaking analysis. I appreciate you coming on. And thanks for watching, everybody. Big news today. We'll be tracking this. Check out siliconangle.tv for this and other videos. Check out siliconangle.com and wikibon.org for more analysis. Thanks for watching. We'll see you next time.