Intel felt the hardship of the ARM revolution more than anyone— smartphones and tablets completely skipped over Intel's processor range and opted for the cheaper and more efficient ARM alternatives. This was largely their own fault as until just a few years ago, Intel themselves neglected to focus on that end of the market.
Out of all of the ARM manufacturers who looked to herald in the smart device revolution, Qualcomm was the one who enjoyed the most success. Their sales figures show this too, with Qualcomm closing 2014 with revenue of $26.5 billion— around 7% more revenue than 2013 which brought in $24.87 billion, with both of those figures continuing a pattern that started just a few years earlier. The San Diego-based semiconductor company had some of the most sought-after smartphone and tablet chipsets in the entire industry, setting the benchmarks for excellence and breaking performance records with every new chipset release they brought to the market. It certainly has been an amazing few years.
But that's going to change very, very soon. We've reached the summit. It's over. It's finished. Competition is the backbone of capitalism, and no single manufacturer can ever claim persistent, unchallenged dominance forever.
This extended the expertise of the ever-growing
MediaTek hive mind
Enter MediaTek. A relatively unknown semiconductor company headquartered in Taiwan, founded in 1997, MediaTek had humble beginnings, starting with manufacturing chipsets for optical drives. Eventually moving forward from the CD-ROM, they started producing chipsets for DVD drives as well. Slowly, this grew into the portable electronics industry and the company started swallowing more and more marketshare, taking hold of everything from DVD players to TV sets. They had a high reputation in Taiwan as being one of the top fabless manufacturers, but 2007 is when they really started to shine.
It began with the acquisition of the cellular division of an American firm called Analog Devices, allowing them to get their foot into the Chinese phone market. A few years later came the acquisition of Ralink— if you've ever bought a cheap WiFi dongle, chances are it featured a Ralink chipset. This extended the expertise of the ever-growing MediaTek hive mind and solidified their investment into the mobile world.
Then came the smartphone explosion. RIM's BlackBerry changed the way we think about phones, and the Chinese world naturally started producing BlackBerry-influenced phones of their own. This is where MediaTek initially really got their feet wet. With the production of the MT62XX series chipsets, they began their upwards advance towards glory.
The chipsets were really as Eastern as chipsets could get, often offering things like multi-SIM support and support for television channel playback— if you remember those Chinese phones from a few years ago that had giant antennas on them for watching TV, there was a good chance that it was a MediaTek chipset powering the phone. Coupled with limited band support, these early smartphones and MediaTek themselves never really entered the Western market: it was a different game with different requirements.
Android changed everything.
The Chinese manufacturers found their holy grail
Moving on to the explosion of Android in the Asian world, the smartphone game had taken a different form completely. By now, Android had reached some level of maturity and had entered version 2.0. Apple had already set the standard for the future of smartphones where apps are king and in order for the Asian market to progress, they needed to have a competitive product. This is where Android really shined: it was free, it was heavily supported, and it was flooded with mainstream apps. Oh, did I mention it was free? The Chinese manufacturers found their holy grail.
The manufacturing game was up to that point a complicated task. In order for a manufacturer to produce a phone, they largely had to jump through hoops to get it done. It involved customizing large parts of proprietary operating systems like Nucleus RTOS, it involved writing dozens of drivers for various different keyboards, displays and chipsets, and it involved having to hardcode any special applications that the company wanted to include to make themselves stand out.
It was a chore at best, and it was expensive too. Android took all of this away and produced a single operating system that simplified a large part of the phone manufacturing process. Coupled with the removal of buttons and a touchscreen-only experience, it meant that even the driver-coding part of things became significantly easier. As for the phone's software, as long as the screen was able to register an input at some position [x, y], Android would handle the rest. The game wasn't over just yet, though.
MediaTek's billionaire CEO Tsai Ming-kai reminiscing on his more humble days
when he was just your average Taiwanese multi-millionaire.
Chinese manufacturers needed a reliable SoC manufacturer that would deliver high volumes of chips at cut-throat Chinese pricing— 'Western' manufacturers just wouldn't do. Many stood up to try to take that section of the market, but it was MediaTek that ended up in the dominant position. They already had a high reputation within the Chinese market thanks to their acquisition of Analog Devices and Ralink, so they seized the newly forming Android market as soon as they had a chance. Starting with the MT651X series, we saw the very early 65nm 400 Mhz and 600 Mhz chipsets based on ARMv6 start to take center stage in the Chinese smartphone world. It wasn't until 2010 that we saw MediaTek producing chipsets with 3G support, and a year later we saw the first chipset based on ARMv7. By this point, Mediatek was gaining plenty of steam in the Chinese Android market.
Part of MediaTek's position in the industry was— and remains — not only in providing SoCs but also providing a service to compliment it as well: MediaTek works very closely with their customers and handles most of the slack in getting Android to run on their chipsets, leaving the Chinese manufacturers with just having to tweak the OS to their liking and designing how the smartphone would physically look.
This was an extremely innovative step in the Chinese smartphone industry which significantly increased MediaTek's adoption levels, especially given that it buys straight into the entrepreneurial Chinese spirit of heavily simplifying the manufacturing process and making things a lot easier for the newcomers breaking into the highly competitive Chinese smartphone market.
Over on the other side of the ocean, we weren't celebrating 3G— we were already expecting HSPA+. Qualcomm was beyond MediaTek to the point where it was almost a joke. The MediaTek cavemen were barely experimenting with their first single-core 1 Ghz chipsets while Qualcomm was releasing the dual-core 1.7 Ghz Snapdragon S3 Scorpion. By the time the dodos at MediaTek had finally released their first dual-core 1 Ghz processor— the MT6577— it was already 2012. Just
two, no, three years ago. What was Qualcomm doing at this point? Aside from running circles around MediaTek , they had released their first quad-core Snapdragon S4 SoC. MediaTek who?
Mediatek must have gotten their hands on an Adderall
prescription or something
Enter 2013. Some group of engineers at MediaTek must have gotten their hands on an Adderall prescription or something because in just one year— one single year— they were suddenly producing 28nm quad-core processors. Although they were celebrating their 1 Ghz single-core chipset and swimming in the fast network speeds of 3G just a year earlier, they were now producing multi-core chips with HSPA+ support as if they were born for it. Suddenly, MediaTek was now starting to be seen as a very legitimate potential threat to other semiconductor manufacturers. The obscure Taiwanese company went from producing years old technology that was long forgotten in the West to producing SoCs that were exceeding all expectations. Over a period of just six months, their GPUs tripled in clock frequency. A month later, their CPUs went up 40% in clock frequency too. But that was just the start.
Eight-core chips? Who the ____ wants or even needs that?
Then, almost out of no where, MediaTek went from chasing old technology to literally breaking records and leading the way for innovation. It was something that was expected for some time, just as a novelty, but no one was delivering on it. All of a sudden the underdog, MediaTek, jumped up and took the lead. Just one month after improving their last round of quad-core chipsets they decided to really make themselves stand out as more than just a follower. Introducing... the MT6592. The 2 Ghz octo-core processor with the Mali-450 GPU clocked at 700 Mhz. The smartphone world threw their hands up in amusement, amazement, disdain or blatant disgust— depending on who was reading the news. MediaTek, however, was celebrating their engineering achievement.
MediaTek grew up from a blip in Qualcomm's radar to becoming a real competitor. If it wasn't public before, when Qualcomm made a video mocking MediaTek's eight-core chipset, it was definitely set in stone now.
For the record, though, a few months later we saw Qualcomm eat their words with the announcement of the Snapdragon 615— their first eight-core processor. It eventually made its way into the HTC Desire 820.
Eight-core chips in and of themselves weren't exactly special— any manufacturer could have done it, provided they put the time and effort into it. The pseudo-eight-core chipset was already established in the wild: ARM's big.LITTLE technology— something most of Samsung's Exynos latest processors were being based on at the time— had eight-cores in there. The catch was that they didn't all work at the same time; only four of the cores would ever operate at any specific period. This design revolved around battery efficiency, with the thought that less intensive tasks didn't need the more energy-hungry cores that would be used in gaming, for example. As a result, one set of cores would be used for general OS things, and when more grunt was needed, the four more powerful cores would be used. So, eight cores in total, right?
It was still a feat of engineering that MediaTek took over its rivals
Well, not really. Mediatek was using all eight cores simultaneously. Anyone could have done it, but they were the ones who dedicated the resources and efforts to do it. Granted that to this day there isn't a single app on the market that utilizes it besides the Android OS itself, it was still a feat of engineering that MediaTek took over its rivals.
By this point we had the 20nm Snapdragon 800 chips. Clocking in on AnTuTu around 31,000, MediaTek thought it had finally gotten their arms around Qualcomm as their octo-core chips were also bringing in as much as 31,000 points in the lab.
It didn't take an eight-core processor to get MediaTek buzzing in the Chinese world, though. At that point their position had been completely solidified. Every "off-brand" Chinese smartphone— each and every single one— had a MediaTek chipset. Of course, some of the bigger brands had some of the more sought after Qualcomm chipsets, but they were far and few between when placed next to the sea of MediaTek. It wasn't just staying within China, either.
Because of MediaTek's unbelievably cheap SoC pricing, the no-cost Android operating system and the cookie cutter smartphones being produced in China's endless factories, MediaTek's chipsets are in almost all of the smartphones we see floating around in the rest of the world— including the whole of Africa and Asia (and both of these are really big places with lots of people).
But that was 2013. By now, 2014 had already rolled around. This was the year when Apple changed everything again by making their smartphones bigger, when Samsung took the world by surprise by making their smartphone bigger too, and Android had reached a level of maturity that allowed them to focus on their Metro-inspired flat UI (but don't tell them I said that). All of a sudden it looked like Qualcomm got their second wind. Ignoring their Snapdragon 800 refresh called the Snapdragon 801, with the release of the Snapdragon 805 they had practically made fools out of MediaTek and their octo-core processor. MediaTek thought it had finally managed to catch up to Qualcomm before Qualcomm hit the boost button and took off at full speed. 31,000 points in AnTuTu? That was pretty cute, especially compared to the 46,000 point beast called the Snapdragon 805 which Qualcomm was ready to release.
It took center stage as the Android world prepared to enter the 64-bit era of smartphones
For the first half of 2014 we didn't see MediaTek do anything special. It was almost like the spirit of the company was crushed by Qualcomm's massive second wind.
Wait, who are we kidding— companies don't have souls. MediaTek continued on truckin' and didn't slow down at all, mostly turning their attention on their quad-core range. Early 2014, however, saw the announcement of the MT6595: a quad-core 4G LTE beast that was eventually benchmarked at some 47,000 points on AnTuTu. This was the equalizer. No fake benchmarking magic, no gimmicks— or at least equal amounts on both sides. They employed the same big.LITTLE technology that Samsung had been using and delivered a product that was on par with Qualcomm's latest offering.
But things didn't stop here— they had a little surprise waiting to come out. Around July 2014— around the same time that the Snapdragon 805 was starting to really make waves— MediaTek announced their first 64-bit chipset: the MT6795, again beating their rivals to the chase. Although previous chipsets weren't too special, this one was. Employing eight cores, LTE, 64-bit support, support for 2K display, clocked at 2.2 Ghz with enhanced dual-channel RAM support, it took center stage as the Android world prepared to enter the 64-bit era of smartphones.
MediaTek's booth at CES 2013.
There is one flaw in MediaTek's armor however: the GPU. While MediaTek's chipsets are doing great in terms of raw CPU performance, the GPUs they carry are lagging years behind. Literally, in that the Snapdragon 800's GPU outperforms the latest and greatest from MediaTek, and the Snapdragon 805 almost doubles that. In this aspect, MediaTek is still stuck in 2012 and working on technology that has already arrived and departed in the Western world. Gaming is a large selling point on smartphones, and MediaTek's chipsets fall behind significantly in that respect.
All hope is not lost, however, as improvements are being made in this area as well. Comparing the first generation eight-core chip, the MT6592, to the second generation eight-core chip, the MT6595, we see that it has doubled in performance. If MediaTek's next generation chips can double again, or at least grow enough to rival that of Qualcomm's, then the Taiwanese company's chips would have truly changed the game for mobile computing, phablets, smartphones, tablets, and everything in between.
This isn't exactly difficult, it just appears to be something MediaTek doesn't care about. MediaTek has often utilized ARM's Mali range of GPUs, and some of those GPUs have as much as 378 GFlops (for context, the Xbox 360 has 240 GFlops and the Wii U has a little over 350). By bumping up to the six-core Mali-T760 variant, the Mali-T760 MP6, they can truly equalize with Qualcomm in that respect.
A single difference in company culture between Qualcomm and MediaTek is the bane for their chipset adoption
That was the story of the underdog— the absolute underdog— entering a market and challenging a company ten times its size and bringing in over a hundred times its revenue. But the story doesn't end here, not at all. Back to the actual point of this article: the fall of Qualcomm. Barring the graphics performance, there are only two major things that stop MediaTek from challenging Qualcomm in the Western world, and one of these are the smartphone hobbyists.
A single difference in company culture between Qualcomm and MediaTek is the bane for their chipset adoption: Qualcomm releases the source code for the Snapdragon's kernel, whereas MediaTek keeps it closed. This means that third party ROMs, updates and upgrades are far and few between, with no official builds of things like Cyanogenmod. MediaTek only releases the source code to "trusted partners" and does so at a fairly hefty fee under a NDA, or so the rumor goes anyway. Tech journalists and reviewers are typically those hobbyists who'd want to modify their device and MediaTek simply stonewalls their potential contributions. The second issue which is largely fading away is the lack of band support— initially MediaTek's processors were only supporting certain bands or certain implementations of things like LTE, elements of which weren't well supported in the West. This was before, anyway, as that has largely become a moot point.
So where do we go from here? It all depends on Qualcomm and MediaTek. Qualcomm's latest Snapdragon 810 must really, really change the game. And I don't mean supporting something like 4K video streaming or recording— MediaTek's MT6795 will have that too. What I mean is that it really needs to turn the smartphone and tablet industry on its head and live up to their unofficial "the future of mobile is here" motto— big names like Samsung are reportedly already losing interest.
MediaTek has chased everything Qualcomm has done so far and has finally equalized with Qualcomm on everything except for GPU performance— something that can be fixed in a single generation. MediaTek's chips are... as cheap as chips (no pun intended, I think). They are efficient, they support the bleeding edge of mobile computing, and they are powerful enough to handle most of what can be thrown at them in 2015— so far, anyway. It's not as if MediaTek doesn't intend to chase after Qualcomm either. Although initially they were focused on the low-end market, feeling that abandoned technologies like 2G and EDGE still had life, they now really do have Qualcomm in their sights as a target— MediaTek has already said as much.
Unless Qualcomm shifts its position back into where they were in 2012— that position of dominance where their rivals were barely considered as competition— then the company's position as the market leader is all but already over. It will just take a single act of goodwill by MediaTek: make the kernel open-source; give in to the demands of the hobbyists; open yourself up to the third party world. Add a little bit of time on top of that, improve the GPU, make a couple of deals with companies like Lenovo and Motorola, make some sort of deal with Cyanogen... and it's a wrap. That is of course if Samsung, Huawei or Intel don't destroy Qualcomm first.