Intel’s Peripheral Vision
Intel has been spending its money on some, let’s call it, peripheral stuff – McAfee, contra-revenues, share buy-backs, high divis, wearables companies, Altera etc. and is this at the expense of investment in process technology?
It seems that, while Intel’s 16nm CPU processes are ahead of the industry, Samsung, TSMC and GloFo have caught up with Intel at 14/16nm on SoC processes and look set to overtake it at 10nm.
Earlier this month it was reported that Intel has slipped to third place in the capex spenders behind Samsung and TSMC with Samsung spending $15 billion in capex this year, TSMC $10.8 billion and Intel $8.7 billion.
This is odd for Intel which implemented the two biggest recent advances in process technology – high K metal gate and finfet – three years before the rest of the industry.
Some would say Intel has kept to its Moore’s Law cadence of a shrink every two years while the foundries have run faster with the technology to catch up.
But it is also true that Intel has slipped behind its original roadmap on both 14nm and 10nm though – on 14nm – Intel has shrunk both transistors and interconnect while the foundries have shrunk only the transistors while preserving the 20nm interconnect.
Last week IBM upped the ante with its 7nm technology which will, no doubt, be licensed to Samsung and GloFo, while TSMC is talking about getting 7nm into production in 2017 – the same year as it moves 10nm into production.
So has Intel been spending its money on ‘peripherals’ while neglecting its core competency of process? It looks that way. And, for the first time in decades, Intel us beginning to look a bit strapped for cash with which to repair the damage – it has $14.6 billion cash, $12 billion debt and has to find $16.7 billion to buy Altera.
Last week Intel’s President resigned. We don’t know why, but peripheral vision might have had something to do with it.