Apple’s self-developed baseband is not to make the iPhone signal better

There is no denying that Apple is now a semiconductor giant. On closer inspection, Apple began to develop self-developed chips in 2007 at the earliest, and now the A-series and M-series chips are blooming everywhere, but only in 15 years.

CCS Insight analyst Wayne Lam also predicts that Apple's chip division will become the world's top 12 chip company by revenue.

Apple is not a chip design company in the traditional sense. They only build chips to better serve the product experience, and they are not sold outside. They only exist in Apple products.

▲ Craig Federighi, vice president of Apple software, sighed twice at the press conference about the M chip: How cool is that?

At the annual press conference, we always like that Apple uses its own chips to "hang" the products of the same industry, and also gives its own products a unique advantage in energy efficiency ratio.

It can be said that the Apple Core creates a unique experience for Apple products.

But under the bright, there are exceptions. Apple frequently encountered Waterloo on the network baseband. At that time, it also had a lot of trouble with Qualcomm, and even delayed the launch of the 5G iPhone for a year.

In order to check and balance Qualcomm, Apple mixed the Intel baseband with it, making the problem of poor iPhone signal a major pain point in the experience.

Until now, after the full turn to Qualcomm, this problem has also become a lingering "shadow" of the new iPhone. Although the self-developed 5G baseband was temporarily blocked and reconciled with Qualcomm, Apple did not delay the self-developed process and determination.

Next year's iPhone will not see Apple's self-developed baseband

From the beginning of core-making to small achievements, in terms of A-series chips, it was probably the A7 chip in 2013 that Apple's core-making ability was gradually recognized by the outside world, and it changed the pattern of mobile phone chips.

▲ iPhone 5s with A7 chip

The court case with Qualcomm, and betting on Intel, and then shaking hands with Qualcomm, and acquiring the Intel baseband department, this farce is only two or three years, and Apple is remodeling the team while plugging in the Qualcomm 5G baseband. The process of self-developed baseband with funds is not smooth.

In a Qualcomm earnings report, they said they would continue to provide basebands for the iPhone until the end of next year (that is, the iPhone 15 series).

Earlier, foreign media said through sources that Apple's self-developed baseband may be mass-produced on the iPhone 15 series and used on some models.

As a result, Qualcomm's stock price has also been subject to certain fluctuations, and the market believes that Apple's move will affect some of Qualcomm's performance.

After Qualcomm's financial report stated that it would continue to provide basebands for the iPhone, some sources also said that Apple's self-developed basebands were blocked and encountered overheating problems, which delayed the mass production of self-developed basebands, perhaps until 2024.

▲ iPhone SE 2020

In the iPhone in 2023, self-developed basebands may all be used in entry-level products such as iPhone SE, while Qualcomm basebands may only account for 20%, and flagship products may use mixed basebands.

In addition, Qualcomm has also lowered its financial report expectations in 2025. After all, it is very likely to lose a major customer. Of course, it does not rule out that Apple's self-developed baseband is still blocked.

Difficult to replicate the success of A-series chips

Apple's self-developed SoC has been smooth, and it can be said that the right time and place are right. In the Jobs era, a core-making team was formed, and it spent money to buy high-level architecture licenses from ARM. From the A4, Apple has embarked on chip-making. Highways rarely fail.

▲ Johny Srouji, Senior Vice President of Apple Hardware

In 2008, Jobs acquired two legendary chip designers, Sribalan Santhanam, Jim Keller, and Johny Srouji, who worked at Intel and IBM, through the acquisition of PA Semi and Intrinsty. They also became the souls of Apple's chipmaking team.

The follow-up M chips actually stand on the "shoulders of giants" of the success of the A-series chips. The core-making team has completed the definition and development of the M chips in combination with the requirements of the design team, the software team and the Pro Workflow team. The result is an SoC that is power efficient for the Mac.

If Apple has decades of R&D experience in SoC and the top chip design team in the industry, then in the self-developed baseband, all this is actually blank.

Until 2019, Apple spent $1 billion to acquire Intel's baseband chip patents and related team employees, which can be regarded as a starting point for Apple's self-developed baseband. Unsurprisingly, this team is still led by Johny Srouji.

Apple's self-developed baseband is actually very similar to its self-developed SoC chip. It has poached people everywhere to form an initial team, and quickly gained market recognition through several generations of product iterations.

However, the self-developed baseband is much more complicated than the self-developed Arm chip, and the Intel baseband department acquired by Apple is not the head team in the industry.

The initial team is far less than the Arm chip team that Jobs gradually formed through a series of operations around 2008.

Since the iPhone 7 era and Qualcomm's mixed use of 4G baseband, Intel's baseband has suffered from poor performance, power consumption, heat and other conditions. In the follow-up Apple and Qualcomm had a bad relationship, Intel also set up a separate team to develop 5G baseband, but it was not until it was acquired by Apple. No substantial progress has been made.

As for the complexity of the baseband chip, Ai Faner had a detailed analysis in the article "Apple Builds Cores to Save the iPhone Signal".

To put it simply, Apple's A-series chips only serve its own devices, but the baseband chips are not only for their own products, but also for more than 100 mobile network service providers around the world, which requires individual testing and tuning.

In addition, the communication standards and frequency band information in various places also increase the difficulty and complexity of the development of baseband chips.

Baseband chips not only test the process or later mass production, but also focus on the accumulation of long-term experience.

▲ Picture from: whistleout

MediaTek, Samsung and other manufacturers with self-developed baseband chips have gradually spent 8 to 10 years to gradually catch up with the first echelon, while Apple has only a few years since the formation of the team.

In addition, in addition to the complexity of baseband chips, self-developed baseband chips also need to bypass Qualcomm's patent licensing, or purchase Qualcomm's patent licensing, and for Apple, it is obvious that they want to get rid of Qualcomm's patent licensing fees through self-development. , do not want to be checked and balanced by Qualcomm.

▲ The iPhone 12 that finally supports 5G networks

But from the current situation, the process of Apple's self-developed baseband is obviously not as smooth as that of self-developed SoC, and it can even be said that it is not clear enough.

After the self-developed baseband is launched, it takes more time and manpower to participate in the testing and tuning of different mobile networks. The first commercial self-developed baseband is started on the iPhone SE model, which can be regarded as a low-cost test. Wrong, at least better than the inconsistent user experience of the iPhone 7 era due to mixed basebands.

Self-developed chips, open source and reduce expenditure

The research and development costs of the chip industry are quite high. For those manufacturers who are mainly engaged in the chip business, they mostly obtain maximum profits through product classification.

For a company like Apple that develops chips in order to improve the hardware experience, the R&D costs can be smoothed out through huge hardware sales, and the profits will be invested in the R&D of the next generation of chips, forming a virtuous circle.

That is to say, for Apple, which has been in the self-developed chip business for more than ten years, the R&D cost of switching to hardware is much lower than that of purchasing from traditional chip manufacturers, and it has also virtually raised the barriers to products.

Just like the current iPhone, iPad, and Mac, although they run different systems, they are essentially chips based on the Arm architecture. It is only a matter of a few lines of code to make cross-system calls and linkages, and there is no need to consider them. Break through the barriers of different chips.

For the Apple software team, the development of a new feature is no longer constrained by chips of different architectures.

In fact, before reconciling with Qualcomm, Apple also contacted MediaTek, Samsung and other manufacturers with 5G self-developed basebands, but at that time, their baseband products still had a certain gap in performance with Qualcomm.

▲ iPhone plug-in Qualcomm 5G chip picture from: wccftech

Even if there is no performance difference, in Apple, the cost of plug-in Qualcomm, MediaTek, and Samsung basebands is actually not much different. Choosing Qualcomm is just a basis for cooperation, and no additional adjustments are required.

Apple's use of self-developed basebands is nothing more than a desire to save the cost of basebands and maintain the profit margin of the iPhone.

In recent years, with the increase in the cost of iPhone functions, design, and manufacturing, compared with the selling price, the profit ratio of the iPhone has actually been decreasing year by year, which means "small profits but quick turnover".

At the same time, in Apple's financial report this year, the hardware business has reached an unprecedented height, and Apple's market value has exceeded the sum of Google, Amazon and Meta.

In this context, many analysts believe that Apple's hardware business, especially iPhone sales, will decline in the next quarter, which will affect Apple's financial report, and saving costs in all aspects and increasing profits is to cope with low sales expectations. a countermeasure.

Increasing the research and development steps of self-developed chips and self-developed basebands will also be the only way to increase profits under the general trend, but the difficulty of self-developed basebands is much more difficult than that of self-developed Arm chips. For Apple, the cost time, energy and manpower will be more.

#Welcome to pay attention to the official WeChat account of Aifaner: Aifaner (WeChat: ifanr), more exciting content will be brought to you as soon as possible.

Love Faner | Original link · View comments · Sina Weibo