Is Flash FPGA A Good Idea?

Actel, for long known as an anti-fuse house, went for flash-based FPGA some while back and has built up a substantial business based on that currently running  at about $60 million a year representing about a third of the company's total revenues.

John East, Actel's CEO, targets flash FPGAs at: "What I call the 'Remote' market."

By 'Remote' he means handheld products. "The handheld market needs flash", says East, "flash does that well."

A particular beauty of flash-based FPGA is that: "Xilinx and Altera don't go after that". Says east, "the market wouldn't be big enough to help Xilinx grow its $2 billion  market."

"We're going for handhelds", says East, "in 2007/2008 it had huge growth. In Q2, 20% of our sales were into consumer products, up from 8% five years ago. Flash is 30% of total sales including sales into the industrial and medical markets."

Why medical? "Well you'd think medical was huge heavy pieces of machinery but at China's largest medical equipment company most of the products are battery operated, because hospitals like mobile equipment which is fast and easy to deploy," replies East, "flash is good for that, flash FPGA allows you to add functions as you go along, and ours are the smallest and lowest power FPGAs. The biggest difficulty is persuading people they need a programmable when they're used to ASICs."

Of course the apogee of the handheld device is the cellphone. "High volume cell phones require flash FPGA to do the interfaces", says East, "the baseband is redesigned every three to four years but the peripherals change like crazy. Like displays, USBs, audio, cameras, accelerometers  etc. Baseband chips don't have native interfaces to those peripherals because all those things keep changing. So they need to be able to support that."

So far flash FPGA has created a business for Actel running at $60 million a year. And that sounds like a good idea.