Speaking at The ConFab 2012, in a session focused on prolonging the longevity of existing equipment, Sanjay Rajguru, director of ISMI, pointed out that more than half the current fab capacity today comes from facilities that are more than ten years old. “That’s a pretty significant volume,” he said. Also, supply from 200mm and 300mm fabs continue to dominate the industry’s total output. “This is a real issue for our industry,” Rajguru said. “Can we support legacy fabs and legacy tools given that the demand is significant?”
Rajguru said the problem was really brought to a head to as a result of the earthquake in Japan. “When the Japanese companies started to rebuild their factories, and restock their parts and equipment, they realized the supply chain for many of the older 200mm fabs was really broken. They brought us in to escalate the visibility of this issue for the industry and take on a series of projects and activities to rebuild the supply chain,” he said.
The graph below shows the history and a forecast of where silicon is going to be sourced. “What this particular graph shows is what wafer sizes are going to be produced and demanded. We continue to see 6-7% compound annual growth rate in our industry and you can see that drives some of the growth in the amount of silicon being procured. We do assume that 450mm silicon supply will start and accelerate at a fairly significant rate starting in 2017,” Rajguru explained.
The graph shows a couple of important trends. 150mm fabs are going to start declining in their output, but, surprisingly, 200mm fab output is steady over the next 10 years. The OEMs anticipated 200mm fabs declining at a fairly rapid rate as 300mm reached its peak. That hasn’t happened for two reasons: 1) it’s quite lucrative to run a fully depreciated fab. The ASPs on some of these products had declined to the point where that’s the only place that you can make these devices. 2) Now 200mm fabs are being repurposed to produce LED, MEMS and power devices.
The green line shows that 200mm fab output is fairly steady over the next years. In addition, 300mm silicon demand continues to be fairly robust. “Because of the age of some of these 300mm fabs, they are also going to start experiencing difficulty in getting parts, tools and skills (i.e., those with the knowledge of those tools are retiring) for their fabs,”
Rajguru said the challenge for the legacy fabs is twofold: 1) equipment is getting old and it’s getting difficult if not impossible to get parts, skills and manuals for these tools. That’s what we consider obsolescence management, 2) Equipment capability extension and tool re-use. “Can we extend the lifetime of these tools by repurposing our fab? We are currently making silicon devices. Is it more profitable and effective for us to convert to a MEMS fab?,” he asked.
ISMI has taken some initial steps to help the industry deal with this issue. In the short term ISMI is:
· Developing a risk assessment application specific to our industry that will enable it to develop contingency plans and financial strategies, and allow the industry to assess risk for new process nodes or acquisitions
· Developing a public website that will be a “single source” for obsolete parts and services
· Commissioned a white paper to quantify the financial impact of obsolescence on the industry.
ISMI also has two longer term projects in process: it is developing a strategic roadmap towards an alternate supply chain for obsolete printed circuit boards (the main cause of equipment obsolescence), and it has introduced the “Equipment Obsolescence Forum”. This forum will be the working group that guides all of ISMI’s work on equipment obsolescence.