Indian Wells, Calif.--Should solar companies vertically integrate operations, combining everything from wafer manufacturing to power plant development under one roof, or should they try to claim a horizontal niche in the value chain?

We might see the answer in the next few years -- and the results could be surprising.

A number of large manufacturers such as SunPower, Yingli Green Energy and Canadian Solar are pursuing vertical strategies to varying degrees. Producing wafers and cells in-house guarantees a base level of components while giving a manufacturer an opportunity to capture more profit from utility-scale projects.

The degree of verticality differs from company to company, but generally it can include all the steps from wafer production to building power plants. Many of the companies seeking vertical nirvana are Chinese or have Chinese operations and already own gigawatts of cell and/or module capacity.

The tricky part is that all of these tasks -- wafer production, module making, project development -- are different disciplines requiring separate project and engineering teams. Suntech Power Holdings ventured into project development with a joint venture called Gemini Solar. About two years later, it sold off its interest in Gemini, explaining that manufacturing, not real estate, was its core competency. (Interestingly, the focus on advanced manufacturing has prompted the company to start designing more of its own factory equipment.)

While the list of the vertical companies might make it look like vertical has won, Brian Micciche, president of equipment maker Komax Solar, told us during the Solar Industry Summit that the second largest market right now after China are the contract manufacturers. It was interesting that he thought of them as a distinct group, as if they had formed their own nation.

Another source (during the cocktail party) told me that Samsung, which has publicly said it wants to be the largest solar company in the world, is interested in wafers and cells, but is more lukewarm when it comes to building modules.

Bruce McPherson, vice president of research and development at Suniva, meanwhile, called during a panel discussion for solar makers to revisit the horizontal business model pioneered by the IT industry.

Suniva has built a 170-megawatt module factory and has laid plans to erect a 500-megawatt facility. The company’s core competence, however, is in developing inexpensive cells with a relatively high efficiency (19 percent and beyond) through techniques like ion implantation. By leveraging contract module manufacturers, Suniva, potentially, can reduce its capital costs and sink more into R&D.

Innovalight, of course, was reborn when it gave up module making to become a producer of silicon inks.

And then there is SunPower, the somewhat vertical company that has already outsourced some module manufacturing and entered into a joint venture with Asia’s AU Optoelectronics.  Note too that all of the companies advocating horizontal organizations are from the U.S.

Who will win? It’s hard to say, but if history is a guide, horizontal might be the ultimate winner. The computer industry began with large, vertically integrated outfits like IBM, Sperry-Rand and Digital. IBM, the only survivor in the group, shifted focus when it began to trust Microsoft and Intel to make the components for its PCs.

Meanwhile, chip makers outsourced factory equipment manufacturing after the process became too expensive and complex. Later, the unthinkable happened: most of them got rid of factories all together, outsourcing to foundries like TSMC. Over the long run, it’s just a more efficient model for research and capital. There has been a return to verticality in hard drives, but mostly out of desperation.

By contrast, the desire for verticality to some degree is rooted in two moods that never last: a fear of shortages (verticality in the solar industry began with the silicon shortages of 2004-2008) and a belief in world domination. Ultimately, however, they become creaky and expend their energies trying to hold the family together.

U.S. automakers in the 1920s and 1930s increasingly sought vertical control over their supply chains as a way to scare off potential competition. The interlocking conglomerates of Japan represented a new world economic order -- 23 years ago.  

“We’re not going to ride the dinosaurs to the tar pits,” the CEO of a major computing conglomerate told me once. The company got swallowed up less than two years later.