The push into 2.5D and fan-outs will launch a new round of fireworks in the semiconductor industry.
There has been much talk about semiconductor industry consolidation, but the shift into advanced packaging could have more far-reaching effects than all the mega-deals so far.
Packaging is big business. Yole Développement has pinned the market at $30 billion, but that’s only a thin slice of the pie that’s in play. Companies that win the packaging deals also have a good shot of winning the manufacturing…and vice versa. And while it’s easy enough to add together numbers for Samsung and Apple smart phones, think about tens of billions of IoT devices on limited runs that could involve tens of wafers or just one multi-project wafer. The value of the individual contract is important, but the value of winning market share across wide swaths of new and existing markets is enormous and potentially long-lasting.
The participants in this battle come from several camps, and they’re likely to include others as the games commence. Initially, they include foundries, OSATs and PCB manufacturers. The foundries are obvious contenders here, but how far into this market they venture and what’s considered a profitable run for them will likely be different than for OSATs and electronic manufacturing services (EMS) companies. While the big foundries have been adding capabilities for 2.5D, 3D (monolithic as well as non-monolithic) and fan-outs, they’re going to favor the bigger customers. Companies looking to get to market more quickly, and with smaller jobs, will likely choose other options.
The obvious choices there are OSATs—outsourced semiconductor assembly and test vendors—such as ASE, Amkor, JCET, Nantong Fujitsu Microelectronics, ChipMOS, SPIL, and many others. The largest of these have been investing heavily in advanced packaging and manufacturing capabilities for years, and they’re likely to push further into the foundry space just as the foundries will likely push further into their world. The smaller ones will either need to come up with a niche or an exit strategy.
Where the EMS providers play is anyone’s guess, but with advanced packaging the PCB increasingly is being absorbed into the package rather than a separate SoC that sits on a board. As a result, they either have to move up, or fight for what could well be a decreasing TAM. (The one notable exception here is ASE Group, which has both an OSAT and an EMS company under the same roof, allowing them to each focus on complementary offerings).
That’s the initial battlefield arrangement. Differentiation will come in many forms, which includes everything from cross-industry joint ventures, such as ASE’s deal with TDK, to outright acquisitions of tools vendors to go along with manufacturing. While’s it’s not clear that an EDA company would help a packaging company, another packaging company certainly could. And there is plenty of money on the table in China for those kinds of acquisitions.
Rather than the biggest players buying other companies, the pattern that’s emerging in China is for a No. 2, No. 3 or even a No. 6 company to buy another vendor in order to increase in market rank and reach. That’s what happened with JCET and STATS ChipPAC, and it’s what Tsinghua attempted to do with Micron—a move that appears to have been stunted by the U.S. Treasury Department.
There are many wild cards in this market, and it’s all just beginning to go into active play as the number of advanced packaging designs increases and 5nm looks increasingly dubious. A rise in interest rates and the end of cheap money could slow it down, but the landscape will change nonetheless over the next few years. The only unknown is how and where.