Give a workshop in Shenzhen, one of MediaTek’s complete phone kits, the chip already sitting on a reference board with the firmware and test tools bundled in, and about four months later, that workshop was shipping finished handsets. Not prototypes, actual phones under a hundred dollars with two SIM slots because the big brands refused to build them, plus a TV tuner and sometimes a compass that pointed to Mecca. That was the offer MediaTek made to the entire bottom of the market, and I still think it is the most underrated move in the history of mobile silicon.
The industry had a name for the buyers: shanzhai, the bandit makers, the army of ants that a Harvard case study would later credit MediaTek with arming. The polite word was white-box. The mechanism was the same either way. MediaTek did the genuinely hard parts, the baseband, the radio, the reference design, and the software stack, then handed the whole thing over as a near-finished platform, so anyone with a little electronics knowledge and a case supplier could assemble a phone at a kitchen table. Time to market fell from the year or more a real OEM needed down to a single quarter. By the tail end of the 2000s, MediaTek held something close to half the feature-phone chipset market in China, and roughly 40 percent of those China-built phones were re-exported into the Middle East, India, and Latin America. Chairman Ming-Kai Tsai was called the father of China’s shanzhai handsets, a label he clearly hated, ducking interviews rather than wear it.
Here is the part that still gets me. This was never a low-margin race to the bottom. Handset chips were delivering gross margins north of 60 percent, while everyone treated MediaTek as the cheap option. Tsai had worked out that if you strip every ounce of R&D friction out of your customer’s life, you get to own that customer, and thousands of tiny makers with no leverage is a far stronger position than a handful of giants who can lean on you at contract time. Qualcomm sold the crown jewels to a few kings. MediaTek sold complete, competent kits to a whole peasant economy, and that economy turned out to be gigantic.
The wall it hit was 3G. Qualcomm invented CDMA and sat on the patents, so the jump from 2G feature phones to 3G smartphones was the one transition where MediaTek was genuinely behind, and it stayed behind for a few uncomfortable years. Android is what rescued it. Once Google’s OS turned the smartphone into a commodity, MediaTek could run the exact same playbook one tier up. The MT6575 was introduced in Android phones in 2012. A year later, the MT6589 made quad-core cheap and landed in devices like the Oppo Find 5. Then November 2013 brought the MT6592, marketed as the True Octa-Core, MediaTek’s jab at a petty but real industry fudge: rivals were selling eight-core chips where only four cores could fire at once, so MediaTek built one where all eight actually ran and trademarked the phrase to make the point. It already held something like 45 percent of the Chinese smartphone market by 2012. The budget king had made the jump.
Buying the pieces it did not have
MediaTek’s other reflex in these years was shopping, and it shopped with unusual discipline. It had picked up Analog Devices’ cellular radio and baseband business back in 2007, closing in early 2008 for $ 350 million, which is how a company famous for optical storage chips suddenly owned a serious mobile modem team. In 2011, it absorbed Ralink in a share swap and inherited an entire Wi-Fi and networking portfolio overnight, the reason MediaTek’s routers and Wi-Fi silicon exist at all. A Swedish DSP house, Coresonic, followed in 2012. The heavyweight was MStar: the tender offer opened in mid-2012 at around 3.8 billion dollars, though antitrust regulators in China and South Korea delayed the full merger until February 2014. MStar-owned television and display chips, and that deal is the quiet reason MediaTek became a giant in smart-TV silicon, a business almost nobody files under the MediaTek name. The machine kept running past this window, Richtek for power management closing in 2016, and the Airoha buyout for Bluetooth and GPS in 2017, but the habit was set right here. MediaTek bought capabilities and whole verticals, never headlines.
By 2015, the company finally bolted a real brand onto the ambition. Helio arrived that year with the X10, then the tri-cluster X20 and the slimmer P-series, and that is where this chapter stops, because Helio is its own story and mostly its own set of headaches. What I keep circling back to is how completely MediaTek won a game no prestigious company wanted to play. It armed the counterfeiters and filled the phone your carrier handed your parents, pulling 60-plus-point margins the entire time, while the tech press kept it filed under “the cheap one.” I’m leaving the Qualcomm modem-royalty wars out of this; that is a separate and much angrier post. The turnkey empire was never glamorous. It was just quietly, enormously right, and MediaTek then spent the next ten years trying to convince anyone that it could also build something you would actually brag about owning.
Two decades of MediaTek, oldest to newest: Turnkey Empire (2006-2015) · Helio Years (2015-2019) · Dimensity 5G (2019-2021) · Flagship Assault (2022-2023) · The AI Turn (2024-2026)