Welcome to 2016! This greeting is a bit belated to everyone that already celebrated New Year and a bit premature to those about to celebrate Chinese New Year, but I hope everyone shares my excitement for the coming year. The beginning of the year is usually the time to make resolutions. One of my resolutions is to work out more. I have to admit that if I really increase my efforts, I’ll reach the “sporadic” level! Sometimes, as I’m working up a sweat, I bump into a guy named Rich. Rich knows a little about the compound semiconductor industry and he’s a very optimistic guy. I haven’t bumped into Rich since the summer, but I bet he is still optimistic about 2016 and so am I!
I think the compound semiconductor market has entered the year looking for clarification. There are a number of cross currents at work in the market. How will oil and the Chinese economic trends affect the global economy? What is the net effect of wireless data traffic growth, slowing smartphone and carrier aggregation heating up? What does the Qualcomm/TDK joint venture and their interest in GaAs PAs mean to the compound semi industry? When and how will 5G affect the industry? I think that 2016 may not provide all the answers, but we will have a lot more clarification exiting 2016 than we do at this point.
One segment that has become clearer to me is the base station market. I just published a forecast update on the overall base station market, along with the RF amplifiers used in base station transceivers (Base Station RF Amplifier Forecast and Outlook: 2014 - 2020). The accompanying spreadsheet provides the actual data for the forecasts and segmentations in the report (Base Station RF Amplifier Forecast and Outlook: 2014 - 2020 (Data Tables)). Some very interesting and perhaps counterintuitive trends seem inescapable.
The fundamental trend driving the base station market is the rapid increase in wireless data traffic. That shouldn’t come as a surprise to anyone who reads my reports or attends my presentations; I attribute trends in every compound semiconductor market segment to increasing data traffic! Having a reliable “go-to” driver makes my job easier, but there is a twist in this case. Wireless data traffic increases will indeed require more sectors and more capacity, but fewer of the new sectors come from macrocells.
This statement has some very interesting implications for the entire compound and silicon semiconductor supply chain. The first implication is that the number of base station sectors becomes less important than the type of sector. The ITU estimates that 95% of the world’s population has access to 2G or better wireless coverage. The classical wireless network deployment model involves establishing wide area coverage with macrocells, then adding lower power cells to augment that coverage. Population saturation is a very bad thing for macrocells, but the loophole in the ITU estimate is data coverage. The ITU estimate drops to 43% of the world’s population when they consider only faster W-CDMA, HSPA+ or LTE coverage. Improving wireless broadband coverage is the remaining macrocell opportunity.
Since 2013/2014, the three major operators in China have been deploying LTE base stations to improve that country’s wireless broadband capability. These deployments pushed the macrocell sector quantity to record levels, but we anticipate that this activity will be complete by the end of this year. In addition to deployment activity slowing appreciably in China, equipment manufacturers are contributing to the slowdown by increasing the capacity of macrocells. This seems a bit counterintuitive, but Ericsson has shown upgrade scenarios where an operator can increase the capacity of their macrocells at a faster rate than wireless data traffic is growing. Techniques like MIMO, higher order modulation schemes, carrier aggregation and active antennas power this evolution. It also appears that the long-anticipated small cell architecture is finally getting traction.
Because of these trends, macrocell sectors will decline in the future, while small cells will grow quickly. The net result is growth in the total base station sectors deployed each year, along with a shift in the technology content of these sectors. The biggest and fastest growing segment will be small cell sectors with transmit powers less than 1W. The highest power sectors will see a double-digit decline during the same time. This has big implications for RF base station amplifier technology. The shift to lower power cells is beneficial for GaAs. The growth in small cells will almost offset the effects of price erosion and fewer driver amplifiers. GaN will be the only technology to show growth, but it will be into a declining application, so the revenue results may not be as high as some manufacturers anticipate. LDMOS is the biggest cautionary tale. This technology has been the workhorse for base stations power applications, but it will lose share to GaN and face with the same declining application.
It is shaping up as a very interesting year, so check back often as I try to add clarity to the events that unfold in 2016.
-Eric