Profits at Stake in IC Foundries’ Push for Leading-Edge Technology

Figure 1 shows that although GlobalFoundries is expected to have a greater percentage of its sales dedicated to ≤45nm technology than TSMC this year, TSMC is forecast to have more than twice the sales volume at ≤45nm as compared to GlobalFoundries in 2012 ($6.23 billion for TSMC and $2.79 billion for GlobalFoundries). In contrast, China-based SMIC only entered initial production of 45nm technology in early 2012, more than three years after TSMC first put its 45nm process into production. In fact, less than 1% of SMIC’s 2012 sales are expected to come from devices having ≤45nm feature sizes, which is why its revenue per wafer ($759) is expected to be so low this year as compared to TSMC ($1,190) and GlobalFoundries ($1,157).

In 2012, the ≤45nm foundry segment is expected to represent 30% of the total pure-play IC foundry revenue, up from 22% in 2011. Older technologies (i.e., >0.18µ segments) are forecast to account for only 13% of the pure-play foundry market in 2012, down one point from 2011 and two points from 2010.

As shown in Figure 1, only 11% of UMC’s sales are forecast to be dedicated to ≤45nm technology this year (UMC is offering a 40nm process after bypassing 45nm technology). In contrast, GlobalFoundries is expected to have almost two-thirds of its sales in 2012 come from ≤45nm technology. Moreover, TSMC is forecast to have about $1.8 billion in sales of 28nm devices alone this year, up almost 10x from the $185 million worth of 28nm product the company sold in 2011.

IC Insights continues to believe that the more profitable (i.e., successful) foundries will be those that keep at the leading edge of the process technology roadmap (e.g., GlobalFoundries’ recent announcement that it plans to begin foundry production using 14nm FinFET IC technology in 2014). This can be accomplished through joint ventures and licensing agreements such as the partnership between IBM and GlobalFoundries and/or through significantly increasing R&D spending to develop advanced technology, as TSMC has done.

Figure 2 illustrates how important it is for the major IC foundries to be able to offer leading edge of IC process technology. As shown, there is an obvious correlation between the percentage of sales a major foundry has had over the past 18 months for advanced IC devices and its net income percentage.

In 2012, TSMC and GlobalFoundries were in high-volume production of devices using 28nm feature sizes. Although many of the pure-play foundries other than the Big 4 focus on specialized processes and technology, the process technology gap between the other significant pure-play IC foundries and the leading-edge producers is enormous. Of the 14 pure-play foundries ranked 5th-18th, only four (TowerJazz, Grace/HHNEC, Dongbu, and Xinxin) are expected to be able to produce ICs using ≤90nm feature sizes in 2012, and this production is likely to be relatively limited. Collectively, these 14 “non-major” IC foundries are forecast to account for $4.6 billion in sales, or about 15% of the total pure-play IC foundry market in 2012.