On Jan. 28, FPGA maker Xilinx (XLNX) issued a press release indicating that it had begun shipment of its first high-end FPGAs built on TSMC's (TSM 1.60%) 16-nanometer FinFET+ manufacturing process. Xilinx, it said, is now shipping these products (presumably for revenue).
It would appear that Xilinx has now beaten its archrival, Altera (recently acquired by Intel (INTC 0.61%)), to commercial shipments of 14/16-nanometer class FPGAs. Here's why this is a significant positive for Xilinx and likely a major negative for Intel/Altera.
First to node wins the market share
On Intel's second-quarter earnings call last year, CEO Brian Krzanich told investors that "history tells us that the FPGA vendor who is first to a manufacturing process node enjoys a market segment share advantage over the life of that node."
In this case, we have Xilinx telling its investors that it is now shipping (presumably for revenue) product built on 16-nanometer FinFET+. This comes not too long after Xilinx made sure to let investors know that it began sampling these chips back in September.
On the other hand, Altera, by way of Intel, hasn't said much to investors about its upcoming 14-nanometer FPGA. Altera/Intel put out a press release back in June detailing some of the architectural innovations that the product will bring and said in that release that "engineering samples" of the FPGA in the fall, but there was no follow-up release indicating that such sampling had begun.
It's starting to look as though Xilinx will "beat" Altera to the 14/16-nanometer manufacturing technology node, which could mean market segment share loss for Altera/Intel and corresponding gain for Xilinx.
A potential exception to the "rule"
Although it would appear that Xilinx is first to a 14/16-nanometer FPGA, it's worth noting that Intel maintains that its 14-nanometer technology is significantly denser, as well as more sophisticated, than competing foundry 14/16-nanometer offerings.
That was before Intel acquired Altera, but once the latter went public that it would be manufacturing its next-generation high-end FPGA on Intel's 14-nanometer technology, this was a big selling point that Altera used to try to convince investors that it would indeed gain a sizable edge over Xilinx.
I also wouldn't be surprised if Intel were, shall we say, eager to have Altera go public with this announcement as a way to promote Intel Custom Foundry to potential customers and, more importantly, investors.
When will investors know more?
I doubt that Intel will host a special event to go into the nitty-gritty of what it expects from the Altera business. However, it typically hosts yearly events when it launches next-generation server processors. This would be an opportune time for the company to also provide an update on its stand-alone FPGAs, particularly as these should become an increasingly important part of its overall data center portfolio over time.
After this event, the next major opportunity for Intel to talk about the Altera business will be at its 2016 investor meeting, which will likely be held in November of this year. Management wasn't able to go into too much detail about the Altera business or any plans to develop integrated FPGA/server chips as the acquisition had not yet closed.
With the deal now done, I fully expect the company to go into quite a bit of detail on both the plans for the stand-alone FPGA business as well as more comprehensive plans for integrated products later this year. After dropping more than $16 billion on Altera, Intel stockholders should expect nothing less.
