Bill Watkins’s $200,000 Bentley (VOW:GR) convertible is littered with leftovers from Jack in the Box (JACK). He’s collected $160 million in investor money for his latest company, a high-tech lighting manufacturer called Bridgelux, but rails against venture capitalists. He calls himself a liberal and is against corporate taxation.
None of which is to say he’s difficult to pin down. If you want to know what Watkins thinks about something—anything—just ask. For example, the details of his 2009 departure from disk-drive maker Seagate Technology, where he served as chief executive officer for more than five years. Watkins wanted to place a bet on flash drives through a merger or acquisition, but when he presented the idea to board members, they refused, claiming the stock would take too big a hit. “I got in a big fight, and to be honest, I personalized it by insulting people and losing my cool. They thought I was going to punch out one of the board members. I didn’t, but I did call him, ‘You f---ing motherf---er,’ ” Watkins says with a guffaw that nearly makes him choke.
We’re having lunch in a restaurant across town from Bridgelux’s Livermore (Calif.) headquarters. Watkins, 59, who has short graying hair and a pronounced hobble (bum knees), orders a glass of chardonnay and a pork chop the size of a shoe. “Someone said after,” he continues after chomping through a big portion, “ ‘I lost count after the 30th “f--- you” in your rant.’ We still have this debate. Did they fire me or did I fire myself? I could have handled it better.”
Seagate officials declined to comment, and the company appears to be doing fine. Its stock has more than doubled since Watkins’s departure during the doldrums of early 2009, but he attributes that to last year’s flooding in Thailand, which wiped out competing facilities while leaving Seagate largely unscathed. Now the industry is recovering, and Watkins contends that Seagate is still behind in flash. Either way, he’s back to building a company again in a market that seems to have an especially bright future: LED lighting.
Graphic by 731
A privately held company founded in 2002, Bridgelux generates about $90 million in revenue. It makes LED lighting components, bulbs, and arrays that are used primarily in streetlights, office buildings, and factories. When Watkins arrived in 2010, the company was a mess, and he quickly identified a situation like the one he’d just left. Bridgelux was building the right product but in the wrong way.
LEDs are more energy-efficient than other technologies. Unlike incandescent bulbs, which generate light with glowing filaments, or fluorescent lights, which pass electrical currents through mercury vapor, LEDs are basically semiconductor chips that glow. They comprise a substrate material fused with gallium nitride (GaN), which lights up when stimulated by voltage. They waste less energy and emit a more focused light.
These chips are produced much the same way Intel (INTC) makes microprocessors. LED lights, which can last more than a decade and don’t contain mercury, are increasingly used in smartphone displays, TVs, cars, billboards, and notebook computers. By 2015, manufacturing giant Philips (PHG) expects LEDs to account for roughly 50 percent of a $100 billion global market for lighting. Watkins sees even more impressive growth. He says the energy savings that LEDs provide will help drive the overall lighting market toward half a trillion dollars, with LEDs taking the lion’s share.
As of early 2010, Bridgelux was manufacturing its own chips using solely synthetic sapphire—a costly material—as the substrate. Having lived in and around Silicon Valley for decades, Watkins was familiar with a much cheaper and more abundant substrate: silicon. Working with silicon would let the company use existing semiconductor factories, known as fabs (short for fabrication plants). Watkins saw an opportunity to reduce his manufacturing costs by as much as 60 percent. But he envisioned more than just manufacturing efficiencies: He wanted to base a new era of smart, inexpensive lighting on silicon chips, to make every light socket a node in a network, accessible by mobile phones and computers. In his imagination, each cheap silicon-based LED would become part of a system, providing wireless connectivity and digital video security while dimming and changing colors at the tap of a smartphone screen.
The only obstacle to these would-be wonder bulbs: No one, least of all Watkins, knew how to make them. He had no experience in lighting. He’s not even a technologist; he has a liberal arts degree from the University of Texas at Austin. But he and his engineers figured out how to make LEDs on silicon work. “The reason silicon wasn’t happening was not because of physics,” he says. “It’s because it’s hard.”
Shortly after his arrival, Watkins converted Bridgelux’s Livermore headquarters into an R&D operation, transferred manufacturing operations to Asia, and began a tour to sell venture capitalists on his vision. The VCs were skeptical at first. “People understand when you’re asking to buy a piece of equipment with their money, but when you’re saying, ‘I’m buying a bunch of engineers and putting them in a big room to develop world-changing stuff,’ they get really, really nervous,” he says of those investor meetings. “That’s a hard nut to swallow.”
Watkins was able to make his case, and last year Bridgelux spent $42 million on R&D. While engineers are generally happy to accept money and a challenge, the Bridgelux crew wasn’t convinced that Watkins’s mandate was even feasible. “Bill made a bold move,” says Steve Lester, the company’s chief technology officer. “But it wasn’t clear that it was going to work. I did a survey of the team early on, and some of them said we had a 10 percent chance of success in two years.”
The team had two basic technological hurdles. First was that the key component of LEDs, gallium nitride, has a different atomic arrangement than silicon, so placing one material on top of the other causes defects and degrades the chip’s performance. Second, silicon and gallium nitride are even less compatible at extreme temperature ranges, and the LED manufacturing requires heat as high as 1,100C. “Because of a difference in thermal expansion rates, the gallium nitride films have historically cracked when you cool down to room temperature,” says Lester, “sometimes so badly that there’s no usable material whatsoever.”
To solve those problems, the team came up with a new manufacturing process. It’s a variation of an existing method known as metal-organic chemical vapor deposition, or MOCVD, which involves depositing inert chemicals between the gallium nitride and the silicon. Think of it as laying roughly five dozen layers of padding between your living room rug and the plywood beneath—except here, the layers can be as little as one nanometer thick. “The control and design of those layers has to be extremely precise. A one-nanometer layer has a very countable number of atoms,” says Lester, holding up an LED chip at Bridgelux headquarters. “I won’t say what those layers are—a company’s MOCVD is its secret sauce—but it’s a combination of materials that controls the cracking, facilitates the elimination of defects, and manages the stress that develops in the film.”
Lester’s engineers then devised a way to keep the silicon from absorbing the light generated by the gallium nitride. Once they had it working on a single chip, they set to stamping out 30,000 at a time on discs of silicon. “A lot of people in the world still don’t believe it works,” says Lester.
Sapphire-based LED manufacturing typically uses two- or four-inch wafers. By shifting to silicon, Bridgelux could use eight-inch silicon wafers. This is important because by now, the mainstream semiconductor industry has largely abandoned eight-inch wafers in favor of 12-inch and even 16-inch wafers, leaving many eight-inch fabs in the U.S. and Asia looking for work. Japanese semiconductor manufacturer Toshiba was the first to bite. It signed a joint development agreement with Bridgelux earlier this year, and Bridgelux officials expect mass production of silicon LEDs in the first half of 2013. “The Toshiba deal is a validation for a small company like us,” Watkins says over lunch. And with all the fallow fabs around, he’ll build his silicon LEDs wherever he can get the best deal. “The key is to own the technology and the patent,” he says. “If you have the technology, you’ll always find someone to manufacture it for you.”
The new process should bring greater automation and allow 20 times as many LEDs to be placed on a single wafer. The net effect would be a reduction in the size of LED light sources and a significant drop in price. A 17-watt LED bulb, which emits roughly the same amount of light as a 75-watt incandescent, sells today at Home Depot (HD) for $40. Watkins expects silicon-based bulbs of similar wattage to sell for $5 within two years. Not long after that, he sees prices dropping to as low as $1 or even 50¢ per bulb. The bigger disruption may come when these low-cost, long-lasting LEDs incorporate all sorts of functionality and make the very notion of a bulb that produces only light anachronistic.
“We want a light that you think about all the time,” Watkins says. “Why can’t I have a light that’s sometimes blue, and with the push of a button on my phone it turns green or white? Why can’t I have a light that allows me to cut the energy, that has motion sensors, or that adjusts for reading or TV viewing? We’re going to give you a reason to change your light bulb. It’s the same reason you threw away your last phone—because there’s something better out there.”
The LED subsidiary of Philips, Philips Lumileds Lighting, has been exploring GaN-on-Si, but Ted Mihopoulos, one of the company’s senior technology executives, says Philips is sticking with sapphire for now. The company invested heavily two years ago to convert to six-inch wafers, which realizes some of the advantages that Watkins imagines. “That transition gives us big advantages in uniformity, yields, productivity, and costs,” Mihopoulos says. “If a company is still in two-inch wafers, the move to silicon makes sense. But the cost of sapphire is aggressively, continuously coming down.”
Watkins says he isn’t surprised by Philips’s decision. It has a status quo to protect, just as Seagate did. “It’s easy to sell something additive,” he says. “It’s hard to sell an idea where every dollar you invest comes off the bottom line.” Bridgelux, on the other hand, had little choice. It couldn’t compete with the distributor relationships, the big customers, or the economies of scale that Philips enjoys.
Bridgelux customers seem convinced that Watkins made the right bet. Intense Lighting, an 11-year-old Anaheim (Calif.) company that sells lighting systems to Target (TGT), Nike (NKE) stores, and other businesses, says LEDs have grown to about 50 percent of its sales largely because of the energy savings they provide. Allan Gray, Intense Lighting’s president, thinks lower-priced silicon LEDs will expand the market. “It means we’re going to have a better price point on the product, and we can offer it in places that couldn’t afford the technology today,” he says. “We work with other suppliers—Philips, Osram (SI). But we consider Bridgelux one of our better partners. When we do our presentations, we talk about their innovations. It’s a positive for our company to be involved with them.”
Indian IT services company Infosys (INFY), another Bridgelux customer, annually adds up to 15,000 employees and 2 million square feet of office space. A year ago it began working with Bridgelux and other energy-efficient lighting suppliers to help reduce the company’s carbon footprint. “Not only did we reduce lighting energy. My air conditioning bills came down, which means I need smaller A/C units. That’s a huge savings in capital costs,” says Rohan Parikh, an Infosys associate vice president and head of green initiatives.
Parikh talks about silicon bulbs like a fanboy. He imagines buying lights that have all the features Watkins promises—motion sensors, Wi-Fi, and so forth—and controlling them all from his phone. “We’ve been very excited about the goal of having the light maintain all the intelligence on your floor,” he says. “That could be disruptive. It’s reducing costs but also building intelligence. From my perspective as a buyer, I’m very excited.”
Success is hardly guaranteed for Bridgelux. Samsung and two European manufacturers—Germany’s Azzurro Semiconductors and Plessey Semiconductors in the U.K.—all claim to be readying GaN-on-Si products. Given the expected growth of the LED market, there’s potentially room for several players, but a precipitous drop in prices could quickly strip away profit. Or China could just swallow the whole market, the way it did with solar panels. Watkins says he’s aware of these dangers but figures that by continually innovating, he can drive demand and stay ahead of everyone else, China included. “I have a fundamental belief: China steals. They just steal less from their friends,” he says. “But it doesn’t matter. I like the idea that they don’t invest in R&D. If you continue to invest, you’ll stay ahead.”
Almost four years after his departure from Seagate, the dust-up is still fresh in Watkins’s mind. He professes to still love the company—“Those guys made me rich!” he says. But his bigger concern is carrying out his plan for Bridgelux, and now he’s got a board of directors on his side. “I told my current board, in two or three years we’re going to all look back on this moment and say, ‘This is what we did that fundamentally changed Bridgelux, and that’s why we’re successful,’” he says. “Or you’ll all say, ‘This is why we fired that dumb a------.’ ”