The smartphone has been a fulcrum of innovation for miniaturization and integration ever since the first iPhone arrived in 2007, but 5G brings it not just to a new level but also greater speed of roll-out. This has put enormous pressure on both device makers themselves and their chip suppliers, with demands of innovation and scale economies driving consolidation down to a handful of major players in each case.
Smartphone makers have to deal with an increasingly fragmented spectral ecosystem, while squaring up performance, aesthetics, battery life, cost and speed to market.
The latter is a particular challenge because lead times in the mobile industry have been declining, and almost demanding a rewrite of Moore’s Law, as rate of increase in expectation is accelerating rather than merely doubling every two years. It may be impossible to meet these expectations, but meanwhile smartphone makers are locked into a race to move as quickly as possible. The only relief is that expectations of device longevity remain low, at around two years, after which most consumers expect to replace their handsets, far sooner than for other consumer electronics devices such as TVs and tablets.
A few of our analyst peers have suggested that the expanding remit of integration means that device makers will no longer be able to rely as much on their chipset suppliers for overall system design. The argument is that because antenna systems, RF components and power management now all need to be integrated tightly to meet performance and footprint demands, device makers will have to become more expert than they already are at integration.
In fact the opposite is true – chipmakers will play an even more intimate role in system design since they are already integrating these additional components. Without such integration at the chip level, it will be impossible to deliver competitive 5G handsets.
Probably the biggest single additional challenge is posed by the expansion and fragmentation of spectrum, making it much harder to integrate the 5G modem and RF front end (RFFE) components. The need to support beamforming, dynamic antenna tuning (DAT) and dynamic spectrum sharing (DSS) add to the complexity. Even the largest smartphone makers will want to outsource integration of those components in order to achieve those tight time to market deadlines and contain costs through scale, while retaining enough flexibility to differentiate on features such as battery life as well as design.
So the big smartphone makers, even Huawei, Samsung and Apple, are seeking to reduce not increase their involvement in RFFE design and integration as the complexity increases. Initial extra complexity results from the 3GPP 5G specifications requiring support for NSA (Non-Standalone) operation where 4G LTE provides anchor signals. 5G signals then at this transition stage provide supplemental bandwidth on top of that LTE base layer.
Existing LTE RFFEs themselves have already become complex enough through integration of 4×4 MIMO and aggregation of multiple carriers, for example. 5G adds additional features such as dual-radio support for 4G and 5G in NSA mode. It is true that RFFE design will be simplified when the end goal of 5G SA (Standalone) is reached because the dual radio will no longer be needed. But this saving will be more than offset by the complexity posed by the growing number of 5G bands that need supporting across a wide range of fragmented frequencies, eventually spanning a 100-fold increase. Even early models that do support millimeter wave frequencies will be replaced by future models expanding into the higher ranges up to 70 GHz.
Qualcomm was quickest to seize this opportunity to expand its smartphone real estate into antenna and RFFE integration, coming out in February 2019 at Mobile World Congress with its Snapdragon Mobile Platform. 5G was integrated into a system-on-chip (SoC) incorporating what its second generation 5G mmWave antenna module and sub-6 GHz RFFE components.
Christian Block, general manager of Qualcomm’s RFFE business, believes this market will see 5G-driven CAGR of at least 12% over the next three years, reaching a market value of $18bn at the end of that period. That would be up from $13bn for the handset RFFE market at the end of 2019. Qualcomm’s target is to secure 20% of the total market by the end of 2022.
Samsung, in turn, was quick to take this up and cease focusing on what it considered generic 5G capabilities embodied in RF and antenna integration, even though previously those had been in the realm of device rather than chip or component makers. Samsung first incorporated Qualcomm’s suite of 5G components in its Galaxy S20 range and then continued in its higher end Galaxy S series.
This seemed to represent a significant shift in strategy by Samsung since it had previously been moving towards design of its own 5G chipsets, being a significant player in semiconductor fabrication as well. But now Samsung has decided that its efforts are best focused on higher level differentiating features that consumers will notice, such as larger immersive displays and high resolution camera zoom.
Qualcomm was well placed in 5G integration, but there is an interesting side plot concerning its relationship with Apple. We recall that Apple was locked in litigation with Qualcomm for several years in an ongoing saga that does not need rehashing too much here. The main points were that Qualcomm was determined to protect the business model responsible for a lot of its revenues, licensing of patents around RF and modem technologies, while Apple wanted to free itself from what it considered taxes on its innovations.
But as this saga dragged on, competitive realities increasingly exerted pressure on both companies to settle. For Qualcomm, the dispute compromised its attempts to dominate in 5G integrated chips, while Apple like Samsung wanted to outsource the RFFE design. It was therefore not as surprising as some reports suggested when on only the second day of the federal trial, the two companies announced they had agreed to settle.
Essentially Apple freed itself of ongoing royalties by making an unspecified payment to Qualcomm, whose chips it would be able to source for six years starting April 1, 2019. Qualcomm in turn retained Apple as a customer while deriving acceptable revenues for its patents via that one-off payment.
Not surprisingly though, that protracted battle over patents, although resolved, left Apple reluctant to depend too much on Qualcomm, not that it would have wanted to anyway. Apple also therefore came back to Broadcom and signed two deals worth up to $15bn depending on future sales covering chips and RF components destined for phones until around 2023. This would appear to cover three generations of iPhones, but probably cheaper models given that the deal appears to cover lower end products, leaving Qualcomm to fill in at the top. This reflects Qualcomm’s leadership in 5G at present.
But the other subplot is that Broadcom has been looking for some time at a possible exit from RF to focus on its other core areas including WiFi. The Apple deal shores up Broadcom’s RF business but at the same time makes it more attractive to potential buyers, worth around $10bn.
Apple itself has long been rumored to be considering a move and that $15bn deal probably makes that more likely to happen, as does the growing importance of integration to smartphone makers in the 5G era. Samsung has inched closer to Qualcomm, recalling that in June 2019 the two companies announced that the former’s foundry would be producing the latter’s next flagship 5G chipset, the Snapdragon 865.
Having been adopted for around 140 existing or prospective devices, this was followed a year later in June 2020 by the Snapdragon 865 Plus 5G Mobile Platform with enhanced performance for gaming.
Qualcomm’s extensive business with other handset makers might inhibit a Samsung takeover, as well as the likelihood of US federal intervention. It was in 2018 that President Trump intervened to block a Broadcom move for Qualcomm, which would have been a merger between near equals in revenue terms. But the US government would be quite happy to endorse an Apple acquisition of Broadcom’s RF business. This would meet Apple’s desire to be in sole charge of its 5G destiny and it can hardly be that while relying on a chip maker for the vital business of RF component integration.