As economics changes the face of the semiconductor industry, which way forward should it take?

But events of the last few years have upset the apple cart. From being an industry which operated on the basis of certainty – there was always new technology, new markets and, above all, money – semiconductor manufacture is now a different world. At the recent Global Semiconductor Alliance Forum, companies looked at what the future might hold, with hot topics including collaboration, innovation and profit. There can be little argument that innovation drives the electronics industry forward. Traditionally, venture capital has driven innovation. Despite the 1 in 10 success rate of start ups, that one success kept the investors happy and there were always generous management fees to be charged. Then the global economy tanked; with no exit strategy, VCs closed their funds and the money available to start ups all but dried up as other markets – such as biotech and green – promised better returns. You can understand the move in part when you look at the economics: designing and building an SoC on a 28nm process requires an investment of more than $100million. Future processes will cost even more. And then, can you get a place in the foundry's production queue? Yet, almost by definition, that leading edge chip is application specific, unless you're an Altera or a Xilinx. What application will consume the millions of devices needed to recoup the investment? And bear in mind, when you think of mass market products, that only two mobile phone manufacturers are profitable today. Are we – in some weird parallel of Moore's Law – en route to a world with a couple of foundries, a couple of big chip companies and a handful of product developers? Or could it be that collaboration around legacy technologies will provide the way forward?