Reply to post: Re: Hardly surprising

Re: Hardly surprising

"So what it it is the previous generation of chips, as long as they work that is what matters."

To make my point a little clearer (and why I think China will struggle to compete with bleeding edge processes in the next 5 years) is the difficulty of creating that level.

If you look at volume chip manufacturing, you see a bleeding edge where costs are significant but profits are enormous. Step back one generation and you significantly increase volumes but the profits are much less. Take another step back and generally equipment and plant has paid for itself and it is getting to the point where you can't justify on-going maintenance because the running costs/efficiency barely covers the revenue you receive. Beyond that it tends to be either custom foundries or development.

Getting equipment for bleeding edge is basically impossible (long lead times and huge competition for equipment from the big 3), previous generation equipment used to be available as refurbished/second-hand but is now largely used by the big manufacturers looking to expand their product ranges (i.e. flash/memory) or the second tier manufacturers.

Can China make their own equipment? Eventually, given time, but they need to develop A LOT of supporting industries around very high purity chemical processing (generally only Japan and US doing this at present), very high quality optics (generally only Germany and Japan providing this), high precision steppers (US/Germany/South Korea/Japan) etc.

If we compare semiconductors to other manufacturing processes that have moved to China over time, China tends to take the volume market before taking over the more specialist areas. When China start driving American/South Korean/Japanese memory manufacturers out of business then I will start to believe China can compete. The reality is American/South Korean/Japanese memory manufacturers are struggling/leaving the industry because the big 3 manufacturers are increasing their market share.

Will China want to spend US$20bn a fab for 10+ fabs to possibly be competitive to Intel/Samsung/TSMC in the next 5 years? Assuming they can get a leading edge manufacturing process from that investment - many others have tried and failed.