This afternoon, it became official: pending regulatory approvals,
Apple has wanted to build its own modems for a very long time. Appleâ€™s entire chip business is predicated on integration and cost optimization. Apple owns the operating system and it owns the SoC, so why shouldnâ€™t it also own the modem? Furthermore, its biggest silicon competitor is Qualcomm, who already integrates its modems into its Snapdragon SoCs. Apple loves to do things in-house and if it can, it will. Modems are just the next step on Appleâ€™s journey to integration, albeit perhaps one of the most expensive and complex steps to date.
Intelâ€™s biggest push on modems really began last year when it announced it would spin up a 1,000+ person 5G modem division in San Diego. That was increased to 1,200 people earlier this year, announced days before Appleâ€™s now-settled lawsuit with Qualcomm began in San Diego. I suspect that Apple will continue to expand and hire. Apple acquired Intelâ€™s modem business because it would have taken it a minimum of 5-10 years to build a modem business of its own from scratch. With Intelâ€™s teams and IP, it can likely shorten the time to 3-5 years for a top-to-bottom Apple design (while Apple likely inherited some of its own designs from Intel, Iâ€™m sure it has some things it wants to change). We could also see an Apple SoC with a built-in modem within that timeframe as wellâ€”something that Intel struggled to accomplish. Since Apple likely isnâ€™t married to Intelâ€™s fabs, it can probably go with its already TSMC-optimized SoC designs and manufacture a chip with a modem inside with fewer problems.
Apple still has a modem supply and licensing agreement with Qualcomm, which means it is not going to kick Qualcomm to the curb right away. What will more likely happen is that Apple will work diligently to build a 5G modem that caters exactly to its future needs, which can be integrated into all of its custom SoC designs and give it complete control of connectivity. There is also a significant amount of RF design for 5G that needs to be accounted for, especially with mmWave. Many have struggled with this, including Intel. RF gets more complex with 5G since youâ€™re dealing with even more types of bands, bigger bandwidths, and eventually supporting both mmWave and Sub-6 5G concurrently. RF is hard, and thatâ€™s part of the reason Qualcomm has been able to keep its leadership position for so long.
Itâ€™s really unfortunate to see Intel get rid of this division; the company had the potential to be a real supplier to more than just Apple. At the end of the day though, things just werenâ€™t moving fast enough for Intel to compete with Qualcomm and satisfy its only customer. My biggest concern for Intel is how this will impact its future PC business; Intelâ€™s lack of cellular connectivity could be a weak point for Qualcomm to attack. I expect that other ARM vendors will join
Overall, this brings the number of smartphone manufacturers with their own in-house modem teams to three: Apple,
Disclosure:Â Moor Insights & Strategy, like all research and analyst firms, provides or has provided research, analysis, advising, and/or consulting to many high-tech companies in the industry including Intel, Qualcomm, Huawei, and Microsoft.Â The author does not hold any equity positions with any companies cited in this column.