Why TSMC is exiting solar

Taiwan Semiconductor Manufacturing Co.—the large, relentless and hugely successful chip manufacturing firm—announced it will shut down TSMC Solar, which produced thin film solar panels made with CIGS (copper indium gallium selenide), by the end of the month. Earlier this year, it sold its LED lighting business.

“Despite six years of hard work, we have not found a way to make a sustainable profit,” said Steve Tso, chairman of TSMC Solar. TSMC’s foray into clean energy began in stealth mode in 2009. The company was actively recruiting executives to head up a new solar group and meeting with VCs. In 2010, it announced it would invest $318 million into solar and lighting manufacturing and lined up investments in solar makers Motech and Stion.

The overall idea was that TSMC would take a promising idea and then apply its volume manufacturing know-how and economies of scale to gain market share and eke out profits where others failed. TSMC also played close to its strengths as well: both solar panels and LEDs are semiconductor-based products. The company wasn’t dabbling in organic chemicals. Others took a similar tack.

We won’t be needing this anymore

So what went wrong? While TSMC correctly predicted that solar would be a growth market, it bet on the wrong material. CIGS is a compelling, but notoriously difficult to material. Not only are the chemicals more difficult to work with than silicon, every CIGS process is different, forcing manufacturers to design their own equipment sets. Investors have bet, and lost, billions trying to mass manufacture CIGS. In fact, most of the major failures in solar—Solyndra, Nanosolar (whose early investors included Sergey and Larry), HelioVolt, Hanergy—revolve around attempts to commercialize CIGS.

Weirdly, many of the former high-ranking executives at some of these companies came from the chip equipment industry and knew how challenging developing new equipment sets for a single customer could be. Only one company—Japan’s Solar Frontier—has ever succeeded in a significant way with CIGS.

TSMC, however, also didn’t fully anticipate the business model that is emerging in renewables. TSMC is a classically horizontal company—it concentrates on a particular segment in the market and works hard not to compete with its upstream or downstream customers. By contrast, solar is becoming vertical. SolarCity, First Solar, SunPower and others are succeeding by controlling as much of the value chain as possible. By participating in both manufacturing and project development, these companies can both lower their costs and earn recurring revenues.

Lighting is also somewhat vertical. LED lights last for decades. There is very little recurring revenue once a bulb or fixture is sold. To win, companies need to capture as much revenue up front as possible or try to sell lighting  as-a-service. TSMC was a classic manufacture in a market moving away from classic manufacturing.

TSMC isn’t the first successful tech company to trip in clean energy. Applied Materials invested millions into trying to commercialize amorphous silicon solar panels, only to pull out of that segment in 2010. SpectraWatt, a company spun out of Intel in 2008, was out for the count by 2010. On the other hand, Panasonic has expanded its reach in batteries through its collaboration with Tesla Motors in batteries while Toshiba is a power in the grid equipment market. It just depends on your angle.

http://www.forbes.com/sites/michaelkanellos/2015/08/27/why-tsmc-is-exiting-solar/2/