Its been about two months since the flash memory company SanDisk agreed to pay about $1.2 billion for Fusion-io, the company that was among the first to really push flash into the enterprise data center.
Now, the question is: What is SanDisk going to do with Fusion, now that it owns it? I checked in earlier this week with Sumit Sadana, executive VP and chief strategy officer, to ask exactly that question.
Basically, it comes down to this: SanDisk has been successful in selling flash memory — SD cards for cameras, USB thumb drives for computers and the like — to consumers. According to the research firm IHS, it holds 50 percent of the flash card market and 30 percent of the USB flash drive market.
But it’s also a relatively big manufacturer of flash chips, controlling, in IHS’s reckoning, about 20 percent of the world’s flash-manufacturing capacity. And it clearly wants more.
One way to do that is to go after the enterprise, where flash chips are increasingly overtaking hard drives as the storage medium of choice. SanDisk has, in fact, made four other enterprise-oriented acquisitions since 2011, spending a combined $1.8 billion for SMART Storage Systems, Schooner Information Technology, FlashSoft and Pliant Technology. Fusion was the biggest and best-known of that lot.
“Four years ago, we realized the world was changing around us,” Sadana says. “We saw that cloud computing was going to be an important business driver for us in the data center. We decided to reposition ourselves not just as the top brand for flash storage in the consumer world, but for storage, period.”
Fusion’s biggest business has been supplying memory technology used to speed up servers that run in huge data centers. It turns out that processors in computers waste a lot of time waiting around for something to do. Fusion’s insert cards eliminate that lag time and help keep the processor constantly busy with a steady stream of data. A busier server is a more cost-efficient server, and aside from marquee names like Apple and Facebook, Fusion’s customers have included a who’s who of Wall Street investment banks.
It had been a good business, amounting to about $433 million annually by 2013, but during the years it was publicly held it was profitable only once — in 2011. Its business model brought with it some problems: Sales and demand cycles at Fusion have tended to be “lumpy,” he says, meaning that when its customers were building new data centers or upgrading existing ones, business was good. When they weren’t, it wasn’t.
“Our goal here has been to look past some of the lumpiness with a few large customers,” he says. “The portfolio we have now is something we’re going to take to the market in broad way … Fusion had been really reliant on just a few big accounts through much of its history, but it had recently started to diversify.”
For on thing, more mainstream servers were being sold to large companies via mainstream manufacturers like Hewlett-Packard, Dell, IBM and others. “We already have a terrific relationship with those companies,” he says, supplying them the flash technology like solid-state drives for PCs and servers.
So what’s the goal? Before the Fusion deal, the combined value of the existing enterprise flash business plus solid state drives sold to PC makers amounted to about 29 percent of revenue, Sadana says. (SanDisk had $6.2 billion in revenue last year.) The plan had been to push enterprise sales to $1 billion by
2017 2016. “After the Fusion acquisition, we think we can get there by 2015,” he says.
Look also for SanDisk to expand into software capabilities related to flash in the data center. Fusion had been working on that too, and last year bought a small software firm called ID7. One thing Fusion had been really good at, Sadana says, was making software run faster when running on systems with its flash cards on board. SanDisk he says, it going to make that technology work in tandem with flash from other companies, too. “It created a sense of vendor lock-in that doesn’t make sense,” he says. “We want to make the use of flash in the data center a lot more democratic.”
Clarification: I mangled Sadana’s quote about the revenue goal in my notes, and thus in the story. The original goal for $1 billion in enterprise sales by 2016, not 2017. I fixed his quote to reflect the new goal of 2015, not 2016.