Mediatek: Telecom's invisible force

First impression: AN outsize map Dominates the sparsely furnished reception area. Countries where MediaTek HAS oFFICES glow in red; LIT uP to the right is a large cluster—China, Taiwan, India, Indonesia, and Malaysia, among others. To the left are sporadically glowing landmasses in Europe, West Asia, and America. In another room is a gigantic wall Display of cellphones, Television sets, DVD players, CD-ROMs, XBoxes, Blu-Ray players—Pretty much every device that MediaTek’s chips have ever powered. Between the giant, glowing map and the wall of gizmos, you get a fair idea of MediaTek’s evolution (from HIGH-end Televisions to handsets) and its growing sphere of business—Taiwan, China, and then, the rest of the world. Located inside a Science Park (a bit like a Special Economic zone) in Hsinchu, 70 kilometres south of Taipei, this is the global headquarters of the $3.8 billion(Rs 18,639 crore) MediaTek, Taiwan’s largest and the world’s second-largest fabless semiconductor company after Qualcomm. Essentially, both these companies work on designing the chip but outsource their manufacturing.

The fourth and 11th floors of the 11-storey building house the labs. Outsiders are rarely allowed here. It’s peopled with folks in white coats and transparent goggles, hunching over desks, eerily silent. This is where the company creates its weapons for world domination. MediaTek’s chips run cellphones, most of which retail for $150 or less, and have been a powerful force behind the mobile telephony revolution, particularly among the masses. These chips ran 500 million phones across 152 countries last year: 60 million of those in India alone. In comparison, Qualcomm shipped 400 million globally.

Paul Su, manager, application engineering department, points to some of MediaTek’s work in progress. In the middle of one room full of rectangles, squares, and triangles painted blue, orange, purple, and yellow stands a mannequin with a phone strapped to its ear. It’s fitted with electronic sensors which mimic human hearing capabilities. On either side of it are two hi-fi speakers. Here, Su explains, they test the effectiveness of chips destined for “high-decibel markets” (read: noisy places) such as India.

In another room is a miniature model of a town—complete with a lake, houses, people, cars, and even pets. Around it are phones fitted with MediaTek chips, clicking pictures. Su says that once a picture is taken, the software in the chip will allow the user to view it from any angle, like a 3D picture. This software will be loaded on phones in the next few months. “Our R&D is pretty evolved,” says Su in halting English.

Indeed, nearly 90% of MediaTek’s 5,500 employees are engaged in R&D; they are also young (average age 31.7 years). It’s this single-minded focus on research that allowed MediaTek to emerge from nowhere and disrupt the global handset business, especially in India and China. It’s now training its sights on the smartphone (and 3G) space, hoping to silence critics who allege that MediaTek’s innovation worked for feature phones and is, therefore, very yesterday. Chairman and CEO Ming-Kai Tsai, who holds a master’s degree in electrical engineering from the University of Cincinnati, says India will play a critical role in this effort, given its technology skills. “China and Taiwan cannot offer talent in every aspect of our business. India has talent and could be used as a strategic location in the future.” Already there is talk within the company of opening an R&D centre in Bangalore or Noida. Grant Kuo, MediaTek India’s managing director, says he has begun looking for engineers, but refuses to divulge more.

IT'S NO EXAGGERATION to say that MediaTek was perhaps largely responsible for the mobile telephony explosion here. Rewind half a decade. The rural market was yet to open up and the cheapest colour handset cost around Rs 7,000. The grey market accounted for 15% of sales. Nokia was indomitable, with more than 70% of the 67 million handset market. Nokia’s 1100 model, with a monochrome interface and priced around Rs 5,000, was a bestseller, having sold about 250 million units globally. Behind Nokia were Motorola, Sony Ericsson, Samsung, and LG.

Micromax’s Vikas Jain and Rahul Sharma, or Hari Om Rai, who later started Lava, or Karbonn’s Shashin Devsare, played on the periphery of the handset industry. Jain used to sell fixed wireless terminals for Airtel. Devsare’s Jaina Marketing, which owns the brand Karbonn, was a distributor for Motorola and HTC handsets.

MediaTek, meanwhile, was shifting tack. It was born in 1997 as a spin-off from Taiwan’s electronic giant United Microelectronics and for the first half of the 2000s, made media processor chips for DVD players, TVs, and optical storage devices such as CD-ROMs. In 2001, it began working on wireless chips and by the fourth quarter of 2003, as cellphones were beginning to take off in neighbouring China, it entered the market. In 2004, it shipped around 3.2 million units.

Ultimately, MediaTek would do much more than that and alter the way handsets were made. In those days, companies such as Qualcomm, Infineon, and Texas Instruments made the chips, which were fitted in phones made by the likes of Nokia or Motorola, and ran on carriers such as SK Telecom or Airtel. While chips from companies such as Texas Instruments ran cheaper handsets in the emerging markets, they rarely innovated on designs specifically for such markets. Their focus, much like that of their customers, was on the developed markets, where margins were fatter.

“The Western chipmakers weren’t willing to adapt to the local needs of our markets; it was too much effort,” says William Lu, vice president, Gionee Group, one of the biggest original device manufacturers (ODMs) of phones using MediaTek chips in Shenzhen, China.

On its part, Texas Instruments did make a go for the emerging economies. In 2006, to back the GSM Association’s global mandate to create handsets below Rs 2,000 for Asia, Africa, and Latin America, it built a chip which was used on handsets such as Motorola’s F3. Unfortunately, the phones were not evolved enough
to support aspirations like a QWERTY keypad, expandable memory, or a camera. The next few years, they disappeared.

MediaTek had no privileged relationships with established phone manufacturers and needed to think differently. It began talking to small device makers who were ready to make cellphones, but had no chipmaker backing them. (Other than the chips and software that run it, making a phone doesn’t take much; it’s mostly an assembly job.)

There’s even a word for this in that part of the world. Shan zhai in Cantonese means building a business around rip-offs. While shan zhai firms produced only cheap imitations initially, there have been instances where they’ve built successful businesses.

Estimates put the number of shan zhai firms in China between 100 and 200 towards the late 2000s, located mostly in Shenzhen, producing approximately 150 million handsets a year. The majority of them ran on MediaTek chips. Between 2004 and 2008, its cellphone chips division grew by 262% year on year. By 2010, this accounted for 70% of overall revenues. The rest comes from TVs, Blu-rays, gaming consoles, etc.

MediaTek did more than just sell chips. It sold itself as a partner. It didn’t just sell hardware but also the operating system and software applications that infused life into it. It gave the reference designs on which a phone’s components are placed. It offered its own engineers to check signal quality and device compatibility over different networks. So MediaTek took away the ODMs’ need to have a field engineering and testing team, or software R&D team. This turnkey approach was unique. It cut down the R&D cost of ODMs and they could focus on industrial design or the form (look) of the device.

“Earlier, 80% of the effort went into stabilising the phone after we got chipsets from the vendors. Now we could focus on additional features [like embedding Swarovski crystals on designer phones] for different markets,” says the MD of a Shenzhen-based ODM, who did not wish to be identified.

Soon MediaTek started integrating chips for different features, such as power, radio frequency, processing audio and video files. All the ODM had to do was to put in the camera, the SIM slots, etc., according to the look of the phone. MediaTek also helped out by facilitating dialogue between component suppliers and ODMs. And though a chip accounts for just 15% to 20% of a handset’s total cost, the volumes generated by the ODMs, thanks to such chips, allowed them to negotiate better with other component suppliers, thereby reducing overall costs.

“I think MediaTek’s approach was that everyone in the value chain should make money, not just the chipmaker,” says Alvin, vice president, overseas business, Shenzhen Konka Telecommunications Technology, another ODM. Konka, once the second-largest buyer of Germany’s Infineon chips, now sources only from MediaTek.

Micromax’s Sharma remembers meeting MediaTek executives in Taiwan in early 2007. He was bowled over by their willingness to go the extra mile. By then MediaTek had captured a significant chunk of the Chinese business and was eyeing India, the other gargantuan mobile phone market. MediaTek introduced Sharma to Chinese ODMs. By late 2008, Micromax launched X1, which had a battery life of 30 days. Its phones were made by Shenzhen’s ODMs and powered by MediaTek.

The features and prices of Micromax handsets took India by storm. Phones with colour interfaces, cameras, video recorders, QWERTY keypads, etc., were priced at around Rs 4,000, nearly 60% cheaper than any other multinational handset brand. By 2010, MediaTek-powered phones had captured more than 40% of the market here. Some of these players have even gone global. Their route: India, followed by the neighbouring Saarc countries, then Southeast Asia and Africa, and now Latin America. Micromax is now the world’s 12th-largest handset company (12 million handsets sold in 2010). “One of the major contributors to our success is MediaTek,” says Micromax’s Jain.

Lava’s Rai adds that MediaTek figured out the needs of the Indian brands smartly and suggested the right ODM partners in China. “Their culture of support and cooperation goes beyond giving presentations and conferences to get business.” Barely three years after launch, Lava sells nearly 300,000 phones a month.

One of the advantages of the MediaTek model is that it allows ODMs to churn out models far quicker. MediaTek officials say that nearly 2,000 models are launched every year. Compare this with Nokia, Samsung, or LG, where this number, per company, is less than 40. For BlackBerry, Sony Ericsson, or HTC, the numbers are even lower—between 10 and 20.

This was best visible with dual SIM phones in India. In 2008, after six new telecom service operators launched services, users were faced with multiple tariff options. Entrepreneurs such as Rai figured two SIMs would allow people to switch between carriers. The same year, MediaTek-powered dual SIM phones were launched. They became a big hit overnight and caught market leader Nokia off guard. By 2009, dual SIM phones were outselling the single SIM variety. Today, they account for over one-third of the market. It took Nokia nearly three years to launch its dual SIM phones. That delay was one of the main reasons that its market share drop from 51% in 2009 to 28% in the first quarter of 2011 (according to figures from Strategy Analytics, a consultancy).

At a gathering of journalists in Delhi in September, Nokia CEO Stephen Elop took a veiled swipe at shan zhai when he said that his dual SIM phones were “moving out the Chinese devices” and spoke glowingly about how Nokia had “innovated”. He refused to disclose any numbers.

Given that bundling of services and handsets isn’t a practice here, telecom experts often describe India as a ‘velocity’ market, where the buzz created by handset makers drive cellphone choice, rather than what carriers offer. So it isn’t difficult to see why Micromax, Karbonn, and Lava do so well. They continue to launch nearly three to four new models a month each.

“We realised that the technology adoption curve differs from country to country, but the aspirations are same. We were closer to our market that was asking for good quality, feature-rich phones, but at affordable rates. We had the advantage of low R&D and employee costs and that made the difference,” says Tsai.

Meanwhile, other chipmakers, particularly Shanghai-based Spreadtrum Communications, have slowly begun increasing their share. Between 2008 and the first quarter of 2011, its market share has increased from 2.2% to 15.7%. While much of this is at the cost of Texas Instruments’ falling share, MediaTek now has a worthy rival.

One reason for the increasing market share of companies such as Spreadtrum is that they are beginning to follow MediaTek’s strategy. Diana Jovin, Spreadtrum’s vice president, strategy and investor relations, says its location in mainland China gives the company better proximity to ODMs. It may also help that Spreadtrum’s chips at $2.4 are a dollar cheaper than MediaTek’s. Of course, MediaTek continues to be cheaper than its Western rivals such as Qualcomm’s $12.7 chip or Texas Instruments’ $6 chip.

Spreadtrum’s progress doesn’t bother MediaTek. Its market share continues to be over 30%. But can the company make the cut when it comes to the next wave of mobile telephony—smartphones and 3G and beyond?

MediaTek may be highly focussed on chip design, software, and business process innovation, but it’s perceived to be lagging in technology research compared to leaders such as Qualcomm and Intel. It also trails in processor speeds. Its best to date, commercially, is a 650 MHz processor, while the world has moved beyond 1 GHz and dual core. A lower processor speed limits a device’s load-handling capacity and is one of the reasons for MediaTek’s delayed entry in smartphones. Add to this the patent-related issues that accompany designing smartphones: hardware, software applications, and user interface.

A bigger bogey for MediaTek is Qualcomm. The U.S. company owns the intellectual property rights for 3G and, unlike in the past, is now actively collaborating with ODMs for emerging markets. While imitation may be a form of flattery, for MediaTek, it’s a bad signal. At a Qualcomm event (IndiaOn) in Delhi this August, its executives held closed-door meetings with vendors of Chinese phones, the same shan zhai it ignored earlier. After the event, representatives of a few handset brands who work with MediaTek mentioned that Qualcomm executives were particularly curious about MediaTek’s ways.

No surprises there: The San Diego-headquartered Qualcomm has announced how it’s trying to develop a $100 smartphone for India.

If that happens, many of the Taiwanese company’s partners here would happily jump ship. Handset labels are chipset agnostic. “The end customer is not bothered about what chip runs his device. He is looking for better features and quality. We, in turn, look for the best solution to meet such demands,” says Shashin Devsare, owner of Karbonn Mobiles.

Given that wide-scale adoption of smartphones in emerging markets such as India and China looks 18 months to 24 months away, MediaTek has that much time to get its act together. Or else, the company may potentially be at risk.

IN THE FAST-CHANGING TECHNOLOGY WORLD, it may not be wise for the company to start R&D on 3G and 4G devices now. To get around that, MediaTek is in some sense trying to ally with the enemy. It has licensed Qualcomm’s patents for 3G chips but its ODMs will still pay a 3% to 5% royalty on each handset sold. It has licensed technology for long-term evolution (LTE) from NTT DoCoMo; its chip is expected to hit the markets by 2013. It has joined the Symbian Foundation, Open Handset Alliance, and partnered with Microsoft, with an eye on markets in Europe and America.

Such alliances appeal to Tsai, who argues that there’s nothing stopping the company from “investing heavily” in future technologies—provided there’s a market for it. Hence, the alliances. “Even if we start as tech followers, we will ensure that we’re market leaders in shortest possible time,” he says.

Local brands echo his confidence. They say that although the company might not be investing heavily into new technologies, its market insight, frugal engineering capabilities, and multimedia skills for mass-market phones, would enable the company to win again. “There is something within the company that will make a dent in the market. The technology roadmap, which they have shared with us, looks robust and good enough to meet our demands,” says Rai of Lava.

In 2009 Qualcomm and MediaTek signed an agreement to steer clear of legal issues over 3G technology patents. This gave the Taiwanese company headroom to develop 3G chips and in June it unveiled the MT6573, which goes into Android smartphones. Tsai says the early feedback has been promising and they are on track to reduce smartphone prices as well. The buzz in Shenzhen is that some million such chips have already been sold. However, the lower processing power of MediaTek’s chips and the fact that ODM royalties to Qualcomm will push prices up, may be a dampener.

Tsai’s solution: even more innovation from MediaTek to add more features and reduce prices. He points to the recently launched MT6620, a ‘4-in-1’ chip that combines Wi-Fi, GPS, FM, and Bluetooth technologies. “Innovations like this will help us make a difference,” he says.

It is also widely expected that MediaTek will buy its way into 3G and smartphone expertise that can be added on to Qualcomm’s patents. It has done it in the past. In 2004, it lacked a multimedia interface (MMI) without which its chips couldn’t support audio-video applications on handsets. MediaTek outsourced this to a Noida-based firm, Pixtel Communications, and later acquired the company.

“Even before India became a big market for our devices, an Indian company gave direction to our future handsets and our devices became successful riding on Pixtel’s software,” says Kuo. Also in 2007, MediaTek shelled out $350 million to acquire Analog Devices for its TD-SCDMA, a China-specific mobile technology.

“There are companies such as Texas Instruments which have specific technology that would fill the gap in MediaTek’s portfolio and give it access to customers such as Samsung, Nokia, and LG. Even partnering with Marvell (Intel’s former mobile chip business) for its processor designs and combining it with MediaTek’s expertise would commoditise the smartphone business, like it did in the 2G space,” says Sravan Kumar Kundojjala, senior analyst at Strategy Analytics.

Meanwhile, MediaTek is planning a new front. It has created a new team to evaluate tablets and machine-to-machine communication. According to Tsai, the company has a tablet ready and is in the process of customising for low-cost markets such as India. An analyst says if MediaTek indeed does this, it’ll be a “game-changer”.

A lot of MediaTek’s recent efforts to generate apps for its phones should also come handy in the tablets. In June this year, it announced that it would deliver Yahoo mobile Internet services such as messenger, news, finance, weather, and mail on its devices. It is also negotiating with Google, Facebook, and others to load their services on its chips. That way, any MediaTek gadget will come Google- and Facebook-ready, saving individual brands the trouble of negotiating separately. Then there are developers such as MoreMagic in India and mBone in China being supported by MediaTek to develop applications that can be preloaded or downloaded through an app store such as GMobi, SnapTu, or MRP Store available on the handset. The company also invested $20 million in Spice Digital for value added services that would exclusively go with Spice mobile phones running on MediaTek solutions.

“We are working closely with application developers who can use our platform. We will facilitate their interaction with ODMs, local brands, and operators as well as help them put their applications on app stores,” says Kuo. This will make application usage richer and perhaps also open an additional revenue stream for the company.

While MediaTek may be putting its pieces together to fight the smartphone challenge, software companies such as Google and Microsoft will pose the real threat in the long run. At the heart of MediaTek’s strategy has been software which sets its chips apart. But if a Google launches software and applications that can be loaded on any chip to bring a handset alive, MediaTek will lose its edge as its software will become irrelevant.

FOR ALL ITS success so far, and the fact that some Motorola, Alcatel, and LG phones run on MediaTek chips, the Taiwanese company is still not considered among the elite of the mobile semiconductor industry. Among the more frequent charges levelled at it is that its chips are spectrally inefficient—this means MediaTek-powered phones increase congestion and bring down network quality.

At one of the labs on the fourth floor of MediaTek’s Hsinchu headquarters, there are several phones mounted on tripods. All of them save one are powered by chips made by MediaTek’s competitors. And all are simultaneously ‘talking’ to an internal mobile network. Outside, there’s a huge monitor, somewhat like an ECG machine, that’s plotting the performance of each phone. Su points to the machine and says that MediaTek’s performance is actually superior to others. And that’s the way he and the rest of MediaTek would like the world to see it. After all, being a challenger is never easy.