Greg Raleigh knows just how to attack the Wi-Fi chip set business. The CEO of US company Airgo has a technical edge on his competition—a newly commercialised radio technology that helps expand the range of a wireless network eightfold and its speed as much as sixfold. Airgo has proposed the technology as a standard to succeed the current wireless networking standard. But it could take two years for approval. So Airgo is selling what Raleigh calls “pre-standard” 802.11n chips. Belkin, a Wi-Fi router and access card maker, has begun selling equipment with Airgo’s chip, in hopes of getting early adopters. It’s the same tactic Broadcom used to sweep into the top spot in wireless networking chips in 2003. If it works, Airgo will get a foothold in this emerging market.
“Our aim is to change the technology direction for wireless LANs and then move into other markets,” says Raleigh. “By the time the others get their first generation out, we have to move on to the second. It’s the only way for a startup to gain traction against entrenched competitors.”
Well, maybe not the only way, but one way. When a new technology becomes as widespread and low-cost as Wi-Fi has, chip providers have to navigate the market more nimbly to maintain market share and profits. And that’s what Wi-Fi vendors are doing today. Joe Byrne, an analyst at market research firm Gartner, has identified a range of business strategies that Wi-Fi vendors are using to stay in the game, including remaining ahead of the technology curve, creating technology platforms, forging strategic partnerships and playing in markets where Intel isn’t. Most vendors that have stayed in the Wi-Fi business have employed at least one, and sometimes several, of these strategies. Some companies use more than one strategy at different stages in the market, says Byrne. Those that hit the right balance can coexist with the big players such as Intel or the low-cost makers such as the Taiwanese. With the right business strategy and the stamina to stay in the market, these companies can benefit from the continued explosive growth of Wi-Fi. International Data Corp. (IDC) predicts that the Wi-Fi chip market will grow 42 percent a year, with revenues rising from US$805 million ($1.04 billion) in 2003 to US$2.4 billion ($3.1 billion) in 2008.
When technology ruled
It takes creativity to succeed in the Wi-Fi market today. Intel’s move into the market in March 2003 with the Centrino platform and concurrent US$300 million ($387 million) marketing campaign boosted the size of the Wi-Fi market but also drove down prices (The Centrino platform is a bundle of Intel’s microprocessor, PC chip sets and Wi-Fi into a laptop computer architecture). In addition, Taiwanese companies such as Ralink and Realtek entered the market in 2003, further eroding prices. In 2004 alone, prices for some chip sets fell from US$12 to 7 ($15.5 to 9) a chip, according to IDC. This, in turn, brought system costs down: The price of wireless routers fell from US$1,000 in 2000 ($1290) to just US$59 ($76) in 2004. In that kind of market, only the strong and the crafty survive.
In the first wave of the Wi-Fi market, technology ruled. The early leaders in Wi-Fi chip sets, Agere (distributed by Soanar ) and Intersil (Arrow ) , captured most of the initial market, because they pioneered the technology and its initial standard, 802.11b. But they weren’t able to keep pace with market changes. Agere’s market share fell from 31 percent in 2001 to 8 percent in 2003 and was expected to be even lower for 2004, according to IDC. Intersil sold its Wi-Fi business in August 2003 to GlobespanVirata, which later merged with Conexant (Caelera ).
Startup Atheros came on the scene in 2001 with 802.11a, a new technology that would become a standard for delivering data at 54 Mbits/s. Then many companies flooded the market. By 2003 Gartner counted a total of 50 contenders in the Wi-Fi market.
Broadcom (Soanar) was the next to take the technological lead. It chose to enter the Wi-Fi market in June 2003 with a chip based on the draft of the new 802.11g specification, nine months before it was approved as a standard. Because it arrived early with a part that everyone wanted, Broadcom shot to the top. Even though it sold chips for only half of the year, Broadcom closed 2003 with 22 percent of the US$805 million ($1.04 billion) market, according to IDC.
Even as falling prices hurt rivals that had only commodity chips, Atheros recognised that it could reap higher prices by integrating more functions into one chip. It created “dual mode” chips that combined a, b and g features in the same chip set. Those chip sets still command higher prices, US$10 to 12 ($13 to 15) a chip. Because it stayed ahead of the technology curve, Atheros saw its unit sales grow from 460,000 units in the first quarter of 2003 to 3.6 million units in the third quarter of 2004. Accordingly, Atheros’ revenues grew to US$38.3 million ($49 million) and net income to US$5.5 million ($7.1 million) in the third quarter of 2004, even though its average selling prices fell by half.
Broadcom and Atheros also moved more speedily, in part, they say, because they developed digital radios that could be built in standard CMOS factories. That capability has helped those players stand apart from the pack, says Arnab Chanda, an analyst at Lehman Brothers.
But technology can carry a company only so far when a market is headed toward commodity status. “Technology leadership can work early on, but it becomes harder when everybody gets the technology,” notes Byrne. “At some point, the companies with other strategies catch up with the technology leader.”
Selling a platform
Intel was not the first to the Wi-Fi party and was not the technological leader, but it became a big player by virtue of its platform strategy. It used its dominant position to bundle its microprocessor, PC chip sets and Wi-Fi, thus creating Centrino. Thanks to Intel’s heavy marketing of Centrino, consumers embraced Wi-Fi faster than almost any other new technology. In 2003 only 31 percent of all laptops had built-in Wi-Fi. But now 61 percent of new laptops come with built-in Wi-Fi, most of them using Centrino chip sets. With Centrino, Intel took the leading market share in Wi-Fi chip sets for laptop computers, capturing 35 percent in 2003, according to IDC.
Intel did this even without making the Wi-Fi chips itself. Because it was late in developing its own Wi-Fi technology, the company initially relied on buying various components of Wi-Fi from others, including Texas Instruments , Philips Electronics and Symbol Technologies. That might have seemed unusual on one level. But the company achieved its larger mission with Centrino, which was to broaden the usage of laptop computers using its processors and increase consumer preference for the Intel brand, according to Jim Johnson, general manager of Intel’s wireless group. Intel does not make money on its wireless chip sets, according to Charles Glavin, an analyst at Needham & Co. (Intel declines to comment.) But the Centrino platform’s US$5 billion ($6.45 billion) in sales over the past two years has generated considerable profits for Intel, he notes.
Other chip companies that have broad product portfolios also plan to use the platform strategy. Texas Instruments, the largest maker of chips for mobiles, intends to create a mobile platform that combines multiple chips, including Wi-Fi, that can be sold as a package to customers. TI lost market share in Wi-Fi when Intel started making its own Wi-Fi chips for Centrino in 2004, but TI intends to make a comeback by integrating its Wi-Fi with digital signal processors and other mobile chips, according to Avner Goren, marketing manager at TI. Already TI has a design win at Motorola for a Wi-Fi mobile, he notes. “When you think of who else has all of those capabilities, the list is pretty short,” says Rich Templeton, TI’s CEO.
Indeed, the mobile market, representing 700 million units per year, is perhaps the biggest frontier for Wi-Fi chip makers. By 2008 as much as 15 percent of the 360 million Wi-Fi chip set units will be in mobiles, IDC estimates.
Partnering and other dance steps
Although Centrino is an example of the classic platform strategy, it also incorporates elements of what Gartner’s Byrne calls the buddy strategy. Because its own chip sets weren’t ready yet, Intel had to partner with other vendors of wireless chip set components to launch Centrino.
Broadcom also has used partnerships to help it attain and keep a lead in the Wi-Fi market. Just as the Centrino marketing campaign was popularising Wi-Fi, Broadcom was readying its pre-standard 802.11g chip, which used less power than the former variants. At the same time, Dell, which was seeking a way to regain momentum in the laptop market, asked Broadcom for a chip set using the upcoming standard, according to Jeff Thermond, vice president and general manager of Broadcom’s wireless networking group. Thermond, who joined Broadcom when it bought his home-networking chip company, Epigram, in 1999, had been upset that Broadcom was late to the wireless networking party. So when Dell asked Broadcom for help with Wi-Fi, it was a perfect opportunity to get Broadcom’s pre-standard chip designed into a major computer vendor’s system.
“We had the technology,” says Thermond. “But we also focused on tailoring our products to the market demand. We targeted a few large customers, and we innovated on top of the standard.”
When Intel made a big splash with Centrino, Broadcom was able to turn Intel’s success to its own advantage by successfully appealing to Intel’s customers in the OEM market with a second-source pitch.
“The principle of judo is to use your opponent’s strength against that person,” says Broadcom’s Thermond. “No one wanted to be entirely dependent on Intel, and we were happy to be a second source.”
There was yet another strategy that Broadcom and others exploited: play where Intel isn’t. Broadcom and Atheros have moved into retail consumer markets, selling their chip sets to OEMs that make Wi-Fi access points or routers for the home as well as cards that can be inserted into laptops to enable Wi-Fi reception. Whereas Intel confined itself to selling chips to laptop makers, Broadcom and Atheros sold their chips to equipment makers such as Netgear, Linksys and D-Link.
Broadcom went one step further and provided Wi-Fi chips that were integrated with its cable modem and digital subscriber line (DSL) modem chips. Intel says it has no plans to move into these niches. The strategy of combining Wi-Fi with cable modems or DSL modems helped Broadcom reach higher volumes than it otherwise might have if it had focused only on stand-alone Wi-Fi chip sets. Broadcom and others can combine Wi-Fi chip sets with other functions for home networking gear companies such as Linksys and Netgear, says Byrne.
Keep on moving
The rest of the players aren’t just sitting around waiting for Intel to move into their markets. Atheros has achieved eight design wins with TV makers that are designing Wi-Fi into their products to enable them to beam high-definition video and recorded TV shows from one TV to other TVs in the same house. TI is selling chip sets to printer makers such as Hewlett-Packard, because Wi-Fi-based printing is starting to catch on. Wireless handhelds, game consoles and set-top boxes are all expected to adopt Wi-Fi capability as well.
And the next step for Wi-Fi may represent another opportunity for a company to take the technological lead in the market. Intel is now pouring resources into WiMAX, a new radio technology that Intel believes will widen the range of Wi-Fi hot spots as much as 10 miles (16 km), according to Intel’s Johnson. Intel has launched its first standard-based chip sets to implement WiMAX networks and believes that WiMAX networks can serve as backbones for Wi-Fi networks, carrying traffic for many hot spots into the communications infrastructure. It also believes that WiMAX will extend wireless Internet access to rural communities.
But Raleigh is putting Airgo’s efforts into 802.11n, whose range is less than WiMAX’s but which is more appropriate for moving video streams around within a house or an office, because it can reliably penetrate walls. It also promises faster throughput, especially as consumers start sending video from one machine to another. But the standard is still in its early stages, and IEEE approval isn’t due until 2006. Airgo has an alliance of chip makers, including TI, behind it, but Intel is leading a rival standard group that is proposing its own version of 802.11n. If Airgo’s approach isn’t approved, then its pre-standard chips won’t be compatible with the final standard. That would limit their adoption and possibly lead to consumer backlash. Still, Raleigh thinks the risk is worth taking.
“We’re just following the strategy that others have proven,” he says. Yet, as the Agere and Intersil examples show, technology leadership doesn’t last forever.
Further information: This article was first published in Electronic Business (US) in February 2005.