I attended the annual IMAPS local chapter lunch here in Arizona last week and got the low-down on the semiconductor market and forecast from Bill McClean at IC Insights. Bill does a great job at showing the whole picture – comparing the Worldwide GDP and showing the correlation between it and the worldwide semiconductor market. He’s also good at relating theory to statistics, which helps someone like me, who always needs to know WHY, understand the charts and graphs. So rather than recite the percentages and charts here (hey, to get that you can just read the presentation here.) I’ll relay the major takeaways.
Overall, Bill says the next couple of years look good from a growth perspective. Last year, the total semiconductor market —including ICs, and optoelectronics/sensors/discrete (OSDs) — grew 2%, with most of that growth in the OSD side (could that be thanks to CMOS image sensors?), and is expected to grow 7% this year. Capex increased by 17% last year, so he’s expecting a -5% decline this year, while on the materials side, expect a 7% increase. He said the materials market is approaching $50B, and it usually follows the IC market.
“The semi market usually shows a 2 year rebound after a major global recession,” noted McClean.
“This time, we got about half that (or a one year rebound). The decline was so severe in 2009; we’re still feeling the effects of it 3 years later.” He added that because the IC was invented in 1958, we’ve never experienced such a massive recession as we did in 2009. So we had a rebound, replenished inventories, and then had another drop.
McClean credited the worldwide stimulus spending for most of the 2010 rebound. After that, he said we went from stimulus to cuts, which is what we’re feeling in 2011, 2012. Without the stimulus money, GDP in 2009 would have been -3.2 rather than -1.9, and growth in 2010 would have been 2.1% instead of 4.3%. “Without that money, we wouldn’t be where we are today,” said McClean. “It’s a pretty scary world; who knows where we would be. It was like putting training wheels on the economy to get through the rough patch, come out and continue growing.” He added that although the bike is still wobbly, growth without stimulus in 2011 was almost the same as with stimulus.
Calling 2011 “one of the strangest years ever,” McClean cited all the unforeseen events that affected the market, such as Arab unrest resulting in higher oil prices; the earthquake in Japan, which took out the Japanese economy and shook up the supply chain, particularly for substrate materials; flooding in Thailand that caused a shortage of HDD for PC applications; various US natural disasters (11 of which reported damages equalled $1B or more) that caused disruption to local economies and hit US GDP growth; the Euro debt crisis and US debt ceiling deadline; and an overzealous IC inventory build due to PC and tablet sales forecasts; not to mention the impact a 27% drop in DRAM had on the IC market. “This is how projections go from 10% to 2% growth in a year. Uncertainty is the worst thing for the market,” noted McClean. “People can take bad news better than uncertainty.”
He also talked a lot about China’s growth and its social stability factor because the country subsidized itself through the recession. Employment remained high and high growth is expected. At this point, China is the #1 market for consumption for PCs, cell phones, automobiles and digital TV, said McClean. They haven’t fared so well on semiconductor manufacturing, however, largely due to the lack of intellectual property protection. Indigenous IC fabs aren’t doing so well, noted McClean, and therefore China is pushing a fabless model going forward, focusing on developing design companies.
Looking ahead to 2012 and beyond, McClean predicted moderate growth this year – around 3.5% – with double digit growth returning in 2013, 14, 15, and 16. He noted as a couple of positive indicators, the European Central Bank talking about lowering interest rates, and the weakening Euro, which could help Europe’s exports down the road and make the second half of 2012 turn out to be better than the first half.
Focusing on the semiconductor industry itself, McClean noted “I think we’ll remember 2012 as the year where this industry changed for good.” He was referring mostly to Intel and Samsung dominating manufacturing. Both increased capital spending by 24% this year, everyone else is at -19%. “Intel and Samsung are separating themselves from the pack,” he said, citing developments in 450mm, 14nm technology, ad 3D transistor structures. “The other guys are going to be left in the dust. You’re going to have those two, and the foundries. There’s not going to be room for the others. Over next 5-10 years, they’re going to try and take the market with technology and stay so far ahead that others can’t keep up. We’re going to look back at 2012 and see a change.”
At the same time, McClean predicts delays in the transition to 450mm will delay the next cost reduction phase for IC manufacturing. Why the delay? Beyond the few manufacturers who can afford to add the capacity and will benefit from the transition, there’s not a whole lot of enthusiasm around it from the equipment manufacturers who need to develop the tools that will be sold to only a handful of customers. He did say that the difficulties in 450mm could help in the push for development of 3D ICs and advanced packaging solutions.
On this last point, here’s my own personal analysis. It seems that by itself, the only benefit to the transition to 450mm is increased throughput. Increasing wafer size does nothing to improve functionality, performance, and power consumption of the die itself. In reality, many of the processes that have been optimized on 300mm will need to be re-qualified on 450. It seems therefore, that investment in technology advancements such as 14nm, 3D transistor structures and 3D ICs are more worthwhile. And in that area, there is room for some healthy competition. Look out, Intel and Samsung. The game’s not quite over yet. ~ F.v.T.