Apple To Persist With In-House Modem Efforts, Acquires Intel’s Assets for $1 Billion

Apple is the most vertically integrated technology company. However, 60% of Apple’s revenues come from cellular-based devices (e.g.iPhones, cellular iPads, cellular Apple Watch) and this proportion will increase in future with the expansion and penetration of cellular-based offerings. This makes it super-important for Apple to have more control over the technology. Developing in-house reduces its dependence on leaders such as Qualcomm, which has more power in the form of both advanced connectivity solutions and intellectual property (IP) for embedded 3G/4G/5G capabilities.

Cellular Becoming Key, Intel’s Failure to Deliver 5G Solution, Apple’s In-House Ambitions

Apple has been working closely with Intel over the last few years to first build a better 4G LTE modem and RF integration and second, a 5G modem. The 5G modem never saw the light of day due to inherent RF integration complexities and overall sub-par performance. For Intel, which acquired Infineon’s business in 2010 for US$1.4 billion, it has consistently lost money. It had also spent more than US$15 billion between 2011-2018, from its earlier ambition to develop integrated SoCs (which failed), to the standalone discrete 4G modems that reached 100% share in Apple’s newly-launched iPhones, from September 2018, thanks to Apple’s inability to use Qualcomm modems due to the multiple legal disputes the two companies were contesting.

However, Intel’s inability to develop an optimal 5G solution, on time, for Apple combined with its not wanting to throw more money at R&D, meant Intel finally gave up the 5G modem effort in April 2019. This led Apple to sign a 5G modem procurement deal with Qualcomm (after settling its long-running IP disputes).

Apple’s ambition to take the modem and integrated SoC development in-house, dates back more than two years. We were the first to highlight when Apple hired one of Qualcomm’s lead RF VPs, Esin Terzioglu, to kick start and then lead its “Wireless SoC” development ambitions, back in May 2017. Since then, Apple has built a team of more than 1,000 engineers working on the integrated SoC and discrete modem integration project in San Diego. Attracting talent from San Diego and Orange County, a hub for RF engineers from Qualcomm, Broadcom and also from Intel, further north in Silicon Valley. To accelerate the efforts, Apple needed both core IP and additional engineering resources, both of which it acquired in its US$1.0 billion deal to buy Intel’s modem business.

There are already more than 12 5G smartphone models announced/launched, and markets such as South Korea, USA, UK and later China are going to see faster adoption of these devices in 2019 and 2020. Apple will miss this two year cycle and this will affect its market share and mind share – though die-hard Apple fans are likely to hold-out until a 5G iPhone comes to market – probably in late 2020.

Opportunities & Challenges for Apple’s in-House Modem Ambitions

  • According to our Component Tracker Service, Apple spends between $7 billion and $8 billion per year in 4G RF component purchasing, including modems, antennas etc.
  • This spend will increase as Apple bakes-in more cellular capabilities beyond iPhones into other product categories. And even more so as it brings 5G products to market.
  • Consumers are already upbeat  about buying 5G smartphones and are ready to pay higher prices as well.
  • 5G NR technology is a “Spectrum Game” to begin with, shaping how a device is designed. Devices need to be optimized to effectively communicate using spectrum, ranging from as low as 600MHz right up to mmWave bands beyond 26 GHz. This makes it no longer just a “modem game”. The 5G ecosystem is maturing fast and is poised to grow faster than 4G.
  • This means a complex RF front end alongside modems with different power characteristics. These components also range from CMOS based to GaAS/GaN-based. In future, they will also need to support moving from Non-Standalone to Standalone 5G NR environments. This will further increase the effort required and cost of a 5G modem, versus a 4G modem or 4G integrated SoC.
  • The failed effort from Apple-Intel to co-develop a competitive 5G solution shows that developing a modem is extremely difficult, and an integrated 5G SoC is even tougher.
  • Apple has decided that it has to do this in-house – bringing it into line with Samsung and Huawei (both of which own considerable IP and network infrastructure divisions). It will also allow Apple to reduce its dependence on Qualcomm.
  • Acquiring Intel’s IP and resources will give Apple’s R&D push a short-term boost, but it won’t be enough to achieve a superior in-house developed 5G-based device any time soon. Apple will have to play a longer game and go back to basics.
  • The basics include thinking ahead to what might be required for a sixth generation cellular capability, rather than just 5G, and start building intellectual property for these next generation of networks. This will require building a team of researchers that can actively participate and contribute towards the cellular standards within standards bodies like 3GPP.
  • Apple can also study and learn from Huawei, which was almost non-existent in the 2G/3G/4G IP landscapes, but has grown to one of the top three IP holders, and 3GPP contributors, in the 5G era.
  • At some point, Apple will have to start acquiring or partnering with RFFE players to have a more robust in-house integration at the system level.
  • While this will help build a solution, its next big challenge will be to certify this chip with global carriers to ensure it meets operator RF standards and promises a good user experience.
  • Having said that, this entire process will take at least three to four years. So, don’t expect a first appearance much before 2022. Any sooner can be regarded as a massive achievement.
  • This means by the time Apple first brings its integrated 5G SoC to market, the competition will have amassed several years of optimization and commercialization experience, and will likely be rolling out their third or fourth generation 5G solutions.
  • Apple will likely learn a lot from integrating Qualcomm modems and other third party RFFE components, but it won’t be enough if the final goal is an integrated 5G SoC.
  • The secret sauce remains how you tightly build and optimize the modem to the antenna system, with highly efficient power management. Building a significant IP portfolio along the way, with which it can trade, will also be a bonus.

So, in summary, the road to 5G for Apple is going to be tricky. But Apple has to think about the longer game and not rush an noncompetitive solution to market. It would be better to start building cellular IP rather than just acquiring it. It is behind in 5G, but Apple is capable of thinking very long term about a future which could be even brighter.

Apple is the most vertical integrated technology company as you can get. However, 60% of Apple’s revenues come from cellular-based devices (e.g.iPhones, cellular iPads, cellular Apple Watch) and this proportion will continue to increase in future with expansion and penetration of cellular-based offerings. This makes it super-important for Apple to have more control over this technology and also develop it in-house and reduce dependent on leaders such as Qualcomm which has a higher selling power in the form of an advanced solution and Intellectual Property (IP) revenues for embedded 3G/4G/5G capabilities.

Cellular Becoming Key, Intel’s Failure to Deliver 5G Solution, Apple’s In-Hose Ambitions

Apple has been working closely with Intel over the last few years to first build a better 4G LTE modem and RF integration and also a 5G modem which did not see light in the day due to inherent RF integration complexities as well as overall sub-par system-level performance. For Intel, which acquired the Infineon’s business in 2020 for US$1.4 Billion, since then its been losing money and has spent more than US$15 billion between 2011-2018 right from earlier ambition to develop integrated SoCs (which failed) to the standalone discrete 4G modems which had a decent run reaching to 100% share of the modems of the Apple’s newly launched iPhone lineup since Sep 2018, thanks to the Apple-Qualcomm legal disputes.

However, the delays with respect to building an optimal 5G solution on time for Apple and wit fear of losing more money on developing the same, Intel finally gave up the 5G modem effort for iPhones in April 2019 which led Apple to sign a 5G modem procurement deal with Qualcomm. Though, Apple’s ambitions to take the modem and integrated SoC development in-house dates back more than two years. We were the first one to highlight when Apple hired one of the lead Qualcomm RF VP Esin Terzioglu to kickstart and lead its “Wireless SoC” development ambitions back in May 2017. Since then Apple has built a team of more than 1000 engineers working on the integrated SoC and discrete modem integration project in San Diego (in Qualcomm’s backyard) attracting talent from San Diego and Orange County, the hub for RF engineers from Qualcomm, Broadcom and obviously from Intel up north in the Valley. To accelerate the efforts, one thing missing was IP and additional engineering resources which it acquired yesterday from Intel in a US$1.0 billion deal.

Opportunities & Challenges for Apple’s in-House Modem Ambitions

  • According to our Component Tracker Service, Apple spends close to $7-$8 billion from the modem to antenna per year in 4G era.
  • This spend will increase as Apple bakes in more cellular capabilities beyond iPhones into other product categories and more so in 5G era.
  • 5G NR technology is a “Spectrum Game” to begin with.  This is shaping how a device is designed, optimized to effectively communicate utilizing the multitude of spectrum ranging from 600MHz to 60GHz.
  • This warrants a highly complex RF Front end along-side modem with different power characteristics as these components which also range from CMOS based to GaAS/GaN-based and moving from Non-Standalone to Standalone 5G NR in future will drum up the efforts required and cost of a 5G modem discrete or integrated vs a 4G modem or 4G integrated SoC
  •  The failed effort from Apple-Intel to co-develop an industry or rather competitor-grade 5G solution signals the road is challenging for