Ranked: The worst phone processor fails of all time

While the industry is eyeballing the upcoming Exynos 2500 SoC, which may or may not power Samsung’s Galaxy S25 series, with a sense of uncertainty, this sense of foreboding is nothing new to mobile chipset aficionados. This is particularly so with Samsung’s Exynos versus Qualcomm’s Snapdragon arrangement for flagship Galaxy smartphones, where design wins ebb and flow between the two chipset manufacturers every few years.

Here at Android Authority, we’ve seen our fair share of hits and misses in the tricky game of chipset development. So here’s a look back at some of the worst phone processor fails — and the devices they ruined.

Exynos 2300 no-show

Let’s start with the most recent failure in the Exynos vs. Snapdragon saga. The Samsung Galaxy S23 series shipped exclusively with a Snapdragon 8 Gen 2 for Galaxy processor, ending the brand’s long-running dual-chip strategy. While this was a win for Qualcomm, it was a bit of an embarrassment for Samsung.

The Exynos 2300 never made it to market, leaving Samsung unable to power its own flagship smartphones. While Samsung ducked questions about the chip’s absence, lingering manufacturing issues were believed to be the leading cause.

Samsung reorganized its system-on-chip (SoC) business throughout 2022, following a rather poor showing for the Galaxy S22’s Exynos 2200. The chip, along with the Snapdragon 8 Gen 1 produced on the same process, suffered some performance throttling owing to Samsung Foundry’s underperforming 4nm manufacturing node. Exynos’ absence was short-lived; the 4nm Exynos 2400 returned the following year for the Galaxy S24 series.

Qualcomm’s overheating Snapdragon 810

Perhaps the most troubled chipset of recent times is 2015’s hotter-than-hot Qualcomm Snapdragon 810. Its smaller Snapdragon 808 sibling also paid the price for Qualcomm’s dented reputation.

Before the chip landed, rumors abounded that the processor ran into overheating issues. Sure enough, the LG G Flex 2, the first phone sporting the chipset, suffered performance throttling due to heat, handing in benchmarks below previous-gen chipsets. As more and more handsets were launched, widespread fears of a chip-level problem were confirmed. The HC M9 and Xiaomi Mi Note Pro were two other high-profile launches that ran more than a little hot that year.

It’s noteworthy that Samsung didn’t use Qualcomm for its Galaxy S6 release that year — sticking exclusively with the in-house Exynos 7420. Overheating was reported as the reason but neither party ever confirmed this. Samsung returned to using Qualcomm’s 800 series chips the following year. Make of that what you will.

Interestingly, Qualcomm also offered a lower performance, hexa-core Snapdragon 808 chip in 2015. This was a new approach for the company’s 800 series we haven’t seen since. The chip offered two fewer big CPU cores and a smaller GPU than the 810. Following its debacle with the G Flex 2, LG picked up the 808 for that year’s G4 flagship. Although the 808 ran cooler than the 810, enthusiasts and pundits weren’t happy with the lower performance point and still complained about temperatures. Talk about lose-lose for LG.

Intel didn’t even make it to 4G

Intel makes this list not for the damage its chips did to products but for its negligible impact in the smartphone space. However, it wasn’t for lack of trying. Intel and Google entered a joint partnership to provide Android support for Intel processors in September 2011, followed by Intel Atom processors designed for phones under the Medfield, Clover Trail, and Moorefield architectures.

Two of the first Android smartphones to use an Intel processor were the Lenovo K800 and Motorola RAZR i in 2012. But the early Atom lineup was perhaps most popular in the tablet space. None of these became knock-out devices. ASUS turned out to be the most avid adopter of Intel’s mobile chipsets. ASUS’ 2014 Zenfone series launched with dual-core Intel Atom processors. The company moved on to a quad-core Atom chipset for 2015’s Zenfone 2 and Zenfone Zoom handsets. But that was as far as the company went — ASUS eventually moved on to MediaTek and Qualcomm chipsets like the rest of the industry.

To be fair to ASUS and Intel, this partnership produced phones that offered decent performance for their price. However, they fell well short of the CPU and GPU capabilities of leading flagship chipsets at the time.

Intel had further ambitions for Atom with its Silvermont CPU- and Arm Mali GPU-based SoFIA lineup. However, Intel was far behind in modem development, and Wi-Fi-only variants weren’t attracting customers. Despite announcing 3G and 4G chipsets, the SoFIA project continued to miss its launch dates. Intel eventually canned the whole portfolio in 2016 and bowed out of the smartphone space, barely making an impact. However, some of Intel’s smartphone CPU work lived on in budget processors developed by Unisoc.

MediaTek’s dabble with deca-core

While we’re on the subject of chipsets that didn’t end up in consumer hands, does anyone remember the MediaTek Helio X20 and X30? 2017’s Helio X20 was well ahead of its time — it was the first chipset to sport a tri-cluster CPU arrangement which you’ll now find in all high-end Android mobile chipsets.

Despite boasting a novel tri-cluster design and 10 CPU cores, the Helio X20 and its successors were all underpowered. The use of just two big cores and eight low-power cores, four of which had low clock speeds, left the chipset lacking the grunt of rival flagship processors. Although not a good look for a supposedly flagship-tier chipset, the X20 found a home in affordable phones from Doogie, Elephone, LeEco, Sharp, Xiaomi, and more.

MediaTek continued this idea with 2017’s Helio X30, which featured a new PowerVR GPU and Tensilica DSP designed to rival the best in the business. However, the lackluster performance failed to entice customers — Meizu was MediaTek’s only client for the X30. In fact, the 10-core Helio X lineup was seemingly so bad for business that MediaTek dropped out of the flagship chip space for years, only recently returning with the powerhouse Dimensity 9000.

Security troubles for Exynos

Now for a different kind of spectacular fail. Samsung’s Exynos 4210 and 4412 fell victim to a root permissions exploit that was seemingly so “simple” that it was packaged into a one-click application. Talk about lackluster security.

If you’re unfamiliar with the concept of root access, it grants a user or malicious application access to all the low-level files of the Android OS, allowing them to install apps and access sensitive files at will. Deliberately rooting your phone was all the rage back in 2012, as it allowed for apps with advanced capabilities and switching ROMs, so the one-click exploit was a boon to some users. However, it posed a significant security risk to the rest, especially if malware leveraged the exploit.

Alarmingly, the Exynos 4210 and 4412 powered some hugely popular handsets, including Samsung’s Galaxy S2, Galaxy S3, and Galaxy Note 2 flagships. At its peak, millions of customers were at risk. Samsung acknowledged the issue and shipped patches to affected devices, but the usual carrier update delays dragged out this process.

Fortunately, rooting is less popular than it used to be, and chipsets are increasingly secure and harder to exploit. However, Samsung’s faux pas was far from the last major chipset security vulnerability — 2018’s Meltdown and Spectre exploits affected Arm-based mobile chipsets as well as PC components from AMD and Intel.

Apple’s planned obsolescence

Apple’s Bionic processors have been industry-leading for years now, but even the iPhone hasn’t dodged processor controversy. Back in 2017, Apple acknowledged that it was throttling back the performance of older iPhone models to combat the effect of aging batteries. Newer, worn-out lithium-ion batteries can become unstable and suffer from a lower voltage, which could cause a phone to reboot or shut down. By lowering processor performance, Apple also reduces power consumption and avoids the shutdown problem. So in some sense, Apple is doing its customers a favor.

In 2016, Apple pushed an update to all its phones to flip the throttling switch under the circumstances mentioned above. Even though some acknowledged Apple’s good intentions, an uproar followed because the throttling happened secretly. It took third-party research to uncover that iPhones eventually stopped performing at levels consumers expected and paid for. Either way, it was a pretty bad look for a company that sells phones based, in part, on industry-leading performance and one that gave the “planned obsolescence” conspiracy theorist plenty of ammunition.

Apple instituted a lower-cost battery replacement program to address the controversy, even for out-of-warranty customers. A subsequent iOS 11.3 update also includes an option to turn off this generously dubbed “peak performance capability.” Even so, Apple is still throttling the performance of older iPhones once they reach a certain age.

Dishonorable mentions

That’s it for our top five, but quite a few other chipset fails spring to mind. Here’s a roundup of a few of the more noticeable ones:

• Qualcomm choosing the Snapdragon 835 to launch its Windows on Arm initiative was a low-level disaster. The mobile SoC couldn’t provide the performance required for a desktop environment — see the Lenovo Mixx 630 — which detracted from the chip’s battery and networking advantages. In fact, even Qualcomm’s recent Snapdragon 8xc Gen 3 still looks to be far off the industry leaders.
• Although it offered some impressive gaming chops, NVIDIA’s Tegra 4 was a disappointment after several years of decent smartphone chipsets. It wasn’t a bad chip as such — the Tegra 4 ended up powering a range of tablets. Still, without a competitive modem setup, NVIDIA’s proposition couldn’t keep up with its competitors’ increasingly advanced integration solutions. The Tegra 4i revision was too low-power to captivate the mobile crowd, leaving the Tegra 3 as the high point for NVIDIA’s smartphone chipsets. Still, Tegra has been a boon for the NVIDIA Shield TV.
• Did you know that Xiaomi also dabbled in phone SoC development? Its Surge S1 was built from eight low-power Cortex-A53 CPU cores, middling Mali-T860 MP4 GPU, and an outdated 28nm HPC process. It only appeared in the Chinese-exclusive Xiaomi Mi 5c, so hardly made a splash. The Surge S1 was an OK affordable processor, but we haven’t seen anything better from Xiamoi since that limited 2017 debut.
• Speaking of octa-core Cortex-A53 processors, remember when they swamped the budget market? Thank goodness they’re gone, but MediaTek stubbornly held onto this arrangement for a couple of years after Qualcomm and Samsung (mostly) moved on. See the entire Helio P series up to 2019’s P35 and 2020’s Helio G25 and G35 as the last examples. Although these chips may have been cost-effective, we’ve seen significant performance jumps in more affordable smartphone segments since vendors started implementing a couple of big cores.

Of course, there have been plenty of other chipset ups and downs over the years. Did we miss any of your biggest smartphone processor fails? Let us know in the comments below.