The iPhone 5 overheating episode – for which a solution seems to have already been found – is not a trivial anecdote, as it could damage the sales and reputation of the famous smartphone in the most relevant period of a year that is not going well for Apple. In such a case, the income statement would suffer and perhaps the stock’s run on the stock market would be broken. What has alarmed the industry is something more structural: Apple has acknowledged that for another year, and until at least 2026, its fetish product will continue to carry a modem designed by Qualcomm, thus certifying the failure of its years of efforts to have another designed and developed by its own engineers.
Naturally, every smartphone carries a modem and virtually all brands buy it from Qualcomm, without giving much thought to the issue. In the case of the iPhone, with each generation change – usually in September – the brand’s fans look closely at the evolution of the processors (and the new version of the operating system), obviously vital for the operation of a smartphone. With these ingredients, plus some other variable, they periodically judge the Californian company’s capacity for innovation and will come with greater or lesser haste to acquire the latest model in the saga.
What many fans – except for the most fervent techies – do not usually notice is the modem, that tiny but complex piece of silicon that makes it possible for the iPhone, like its competitors, to communicate with the outside world, a question that is taken for granted. The truth is that it is not so much: it has been trying for years to develop its own modem to replace the one supplied by Qualcomm that successive iphones have carried in their guts all their lives. It is certainly important, not only because self-sufficiency raises the brand image, but also because, in principle, the company would save good money and could provide users with what in industry jargon is called a unique experience. As it does, by the way, with its own processors, designed in-house and manufactured by TSMC. In such a case, the profit margin would be kept by Apple, instead of paying Qualcomm a bill that can’t be light.
This is precisely why the renewal of the contract with Qualcomm to continue supplying the modem for the next three generations of the iPhone is more important than any anecdote. Qualcomm supplies Apple with other indispensable components for sending and receiving signals between the device and the cellular infrastructure. The manufacturer delivers the tiny 5G antennas and radio frequency frontends, which are responsible for filtering interference and amplifying the signals.
In addition, Qualcomm licenses to Apple a collection of patents, which are inherent to the connectivity of its cell phones. According to an estimate by UBS, Apple would pay Qualcomm about 1.9 billion dollars a year in royalties. This was already the cause of a legal battle between the two companies that was closed by a forceps agreement in 2019, effective for the next six years (until 2025) with an optional two-year extension. But patents are, with apologies, another matter.
There is in all this an exemplary story. Starting in 2010, Apple became convinced that it could design the microelectronics of its iPhones in-house: it assembled a team of specialists and they set about the task more or less parallel to the task, at about the same time as the development of the processors now in Macs. Nothing seemed impossible for a company that has inherited the omnipotence of its founder. But a modem poses specific problems that the company was not able to solve. In fact, Apple has had its own processors for years, but has not freed itself from Qualcomm as a modem supplier.
The project to design and develop its own modem began to take shape in 2019 and at its head was the Israeli Johny Srouji, vice president of hardware with fifteen years of seniority in Cupertino. Srouji was commissioned by Tim Cook, Apple CEO, to find the formula that would end that uncomfortable dependence.
As the Wall Street Journal reported a few days ago, Ruben Caballero, then vice president of engineering and apparently the company’s top radio communications specialist, disagreed with Srouji on a crucial point. He was in favor of reaching an agreement with Intel, which at the time had the same intention, as the best shortcut until it had secured the design of its own modem.
At first, Cook decided that the two approaches could be compatible: he gave wings to Srouji’s internal project while agreeing to support Intel’s modem design. Things went awry because Intel’s prototypes were not convincing, and Caballero eventually left [he now works for Microsoft]. Not that Srouji solved the riddle, but at least he has kept the job.
Around the time that Apple and Qualcomm were signing a peace of convenience, Intel threw in the towel and accepted $1 billion to transfer the assets it had earmarked for modem development to Apple. As a result, Qualcomm planned that 2023 would be the last year its modems would equip an iPhone.
In fact, relations between Apple and Qualcomm have been going through bad times. They have accused each other of appropriating each other’s patents, lying to the market and even monopolistic practices. The hottest point was when Apple announced the installation of an engineering center in San Diego, in the vicinity of Qualcomm and promised to create 1,200 jobs with the notorious intention of seducing Qualcomm employees to switch companies.
Apparently, Apple went overboard. Wireless communications is a unique world, in which the company had no experience, while Qualcomm was born in the 80s with that specialization, developing CDMA protocols. In short, it is one thing to develop a processor, with all the complexity you want, but which manages communications within the device, and quite another to develop radio communications between the device and the nearest antenna.
Apart from the above observation, the signals must be compatible, now with 5G but also with previous generations and also take into account that there are different suppliers of cell phone equipment and operators that potentially multiply the protocols to be respected. Intel, with all its expertise in a proprietary architecture it invented, tried to satisfy Apple to win it back as a customer, but in the end it capitulated. It had other problems to deal with.
According to the Wall Street Journal’s tidy account, Apple decided that proprietary modems would start equipping 2024 iPhones, since it would not make it in time for 2023. But without much conviction, by the looks of it: to avoid further unpleasantness, in 2019 Apple and Qualcomm had negotiated a conditional extension of supply until 2025, with a possible two-year extension. In short, Qualcomm had the upper hand: if this extension were to be applied, the iPhone 16 and 17 would carry Qualcomm modems. This will be the case and, provisionally, in the third year the supply would be limited to 20% of the required units.
The internal procedures of the companies give much food for thought. Perhaps the purchase of Intel’s assets led Tim Cook – confident in Srouji’s prestige – to believe that he could do without Qualcomm. Which, on the other hand, has never been an accommodating supplier. In 2021, Cristiano Amon, its CEO, told analysts the following: “in our planning we do not count supplying modems to Apple in 2024”. A dance of expectations that portrays the current news: both parties again accept an inevitable divorce, but give themselves additional time to live together.
Nor is this an isolated case. It is well known that Apple has set itself the goal of the greatest possible autonomy in the electronics of its products. Having exceeded that goal with flying colors in the microprocessor, it is still dependent on another silicon component that it buys from a supplier it has not been able to get rid of: a Broadcom module for its iPhones to manage WiFi and Bluetooth. It has been implied that the contract, valid until 2025, would not be renewed because Apple would have developed its own module by then.