Apple are credited with many firsts when it comes to the mobile industry. Regardless of if you like the iPhone, when Apple launched their new smartphone, they were something of an industry upstart. The original iPhone was fabulously expensive and used new technology for the screen, one that did not rely on pressure. Apple did more than this: they invited carriers to pay money to sell the iPhone as an exclusive for a number of years. The established smartphone manufacturers – BlackBerry and Nokia – did not pay so much attention to the new upstart. Surely, customers wouldn’t pay such an outrageous price for the new iPhone? As we now know, Nokia and BlackBerry were wrong; many, many people will pay a lot of money for the Apple iPhone.
Since the original iPhone, Apple have released a string of follow-up devices, each one purported to be the best yet. The iPhone has disrupted the market, shifting the balance of power away from the carrier and towards the device manufacturer. Under the device’s skin, we are also seeing a similar pattern: after the original iPhone, Apple have customized the System-on-Chip that runs the device and the operating system. Modern iPhones use an Apple-designed chipset, which affords the business the ability to finely control how the device performs and ensures the operating system is tightly integrated with the hardware. In effect, Apple has taken the established semiconductor and device manufacturer business model and broken it. We are seeing other companies mirror this behaviour – witness the Samsung Exynos and Huawei’s HiSilicon Kirin chipsets, plus other companies such as LG working on similar internal chipset projects.
One of the side effects of Apple’s high iPhone pricing is that the business generates a lot of profit. This fuels Apple’s ability to invest into future projects and the business has poured significant amounts of cash into their System-on-Chip technology. This has a direct two-pronged attack on the industry. One prong is that it pushes development of mobile chipsets exclusively for the iPhone and similar products. Because the mobile industry has something of a technological arms race, the rest of the industry is now scrambling to catch up. Before Apple started developing their own chips, the industry norm was a new chipset every other year. The second prong is that it is putting immense pressure on other businesses to follow suit, which in turn is hurting those dedicated semiconductor businesses such as MediaTek and, especially, Qualcomm. These two businesses are in a slightly different place of the cycle and it is Qualcomm that I am going to concentrate on, because they have endured a difficult year and there are strong signs that things are not going to get any easier.
For Qualcomm, Apple’s influence has been particularly disruptive. Apple released 64-bit architecture at least a year before it was ready in the Qualcomm stable. In 2013, Qualcomm used a custom application core, called Krait, which it was able to fine tune to deliver a blend of solid performance and power consumption. However, when Apple unleashed 64-bit application processors into the market – regardless of the immediate benefit to customers – Qualcomm found itself in desperate need of its own competitor chipset. The business was developing a 64-bit custom core replacement for the Krait family, called Kyro, which was not ready at the time. Qualcomm had to find a solution and had to find it fast, so they used the ARM reference cores, the Cortex-A53 and Cortex-A57, arranged in a big.LITTLE architecture. The result has been a family of System-on-Chips that offers broadly similar performance characteristics as the competition, which by and large had not developed their own custom application cores. Instead, Qualcomm traded the Snapdragon chipset on the integrated communications, location and Internet radios. Unfortunately, for Qualcomm, MediaTek has also developed its own comparable integrated technologies, but offers the technology at a cheaper price. Worse, the flagship Snapdragon chip, the 810, quickly earned a reputation for overheating thanks to smartphone designers putting the processor into hardware not designed to cope with the heat output of the processor, and software engineers not restricting the chip’s performance enough to keep device skin temperatures cool. In any event, Qualcomm’s public relations did a terrible job at dealing with the overheating Snapdragon 810 processor and for all intents and purposes, it looked like Qualcomm rushed the chip out before it was ready.
We need to consider the impact of Samsung not using the Snapdragon 810 in the Galaxy S6 range of devices, because Samsung as the market leader for Android handsets are one of Qualcomm’s main customers. Samsung used their own internal chipset for the S6 family, the Exynos 7420, which is a superior processor to the Snapdragon 810 because it is constructed using smaller components. This means it is cooler and uses less power, which is important for any smartphone or tablet device, especially one given the challenge of turning around Samsung’s struggling sales and technological change to include a sealed battery. The business relationship between Qualcomm and Samsung has not broken down; Samsung remains an important customer for Qualcomm. Despite Qualcomm winning some important orders, the business has not sold as many Snapdragon 810s as it needed and had to cut back on investment, research and development during the year.
Going forwards, we still don’t know if Samsung are going to use Qualcomm’s next flagship processor, the Snapdragon 820, in the next Galaxy S smartphone. We have seen conflicting reports from around the Internet. In the recent earnings call, Qualcomm’s Chief Executive Officer sounded less optimistic about Samsung using the Snapdragon 820. Qualcomm also stated in the earnings call that they did not expect to receive substantial revenue from the 820 until the second half of 2015 and because Samsung schedule their Galaxy S flagship device to launch towards the end of the first quarter, it is looking as though the Galaxy S7 is not going to receive the Snapdragon 820. Following this earnings call, Qualcomm’s stock price dropped by around 14% as investors digested the news that the Snapdragon 820 may not be inside the Galaxy S7. Qualcomm are also planning a further reduction in research and development for 2016 to the tune of $1.1 billion. This won’t have been an easy decision for the business because reducing inward investment at a time when Apple is able to increase research funding is not good news.
Where now for Qualcomm? We have seen that the business appears to be gradually moving into the wearable and Internet of Things markets and in the earnings call, Qualcomm spoke about opportunities in the low-end and mid-range chipsets. Chasing mid-range customers has worked out well for smaller device manufacturers such as OnePlus and WileyFox, to take two examples, but this discounts Qualcomm’s considerable investment into their new high end Kyro custom core as these less expensive chips typically use reference ARM processor cores. It also puts them into direct competition with MediaTek. Given that the business was unwilling to provide earnings and revenue guidance for this time next year, it appears that Qualcomm cannot foresee with any confidence what might happen to their business over the next year. Perhaps Qualcomm will repurpose the Kyro core and use it in a mid-range chipset or two? Perhaps the Samsung / Snapdragon 820 deal is still up in the air?
With so many questions, Qualcomm are in a difficult position. The business needs the Snapdragon 820 to be a commercial success and this would be easier if it powered the next Samsung Galaxy S flagship devices. We will have to wait and see if Qualcomm’s mid-range and low-end mobile System-on-Chips can compete in the market and generate money for the business. Mobile application processors are moving into becoming a commodity item, but we do not yet know the timescale. Independent semiconductor designers and manufacturers are being left on the sidelines as the larger device manufacturers take their designs in-house, and there are fewer and fewer chipmakers left to satisfy the smaller device manufacturers. Qualcomm may be a very different business this time next year. For a business that has celebrated one billion devices with a Snapdragon inside, that it may not reach two billion would be a shame for the wider industry. Apple’s disruption in the System-on-Chip market may ultimately result in Qualcomm seeking their profits elsewhere, and this would be bad news for the market in need of constant innovation.