IBM's new Z13 mainframes start at $200,000

When newcomers barge into an industry, they do the easy stuff. That’s logical enough. But as they get experience, they shinny up the value chain and displace the incumbents. “Oh, they’re just handling the bottom of the market,” the incumbents say, as they abandon one market after another. “They can’t touch our jewels.”

I saw this drama play out when I was covering steel in the ‘90s, and I see it now in the computer industry, specifically with IBM.

First steel. If you wanted quality steel in the 1980s, you went to a company that made it the traditional way, by reducing barge-loads of iron ore to liquid metal in a 2,900-degree (f) blast furnace, and then refining it. It was an enormously expensive (and dirty) industrial process. The two biggest producers were U.S. Steel and Bethlehem Steel.

During those years, so-called minimills were spreading across the country. They “made” steel by melting scrap. Viewed as little more than junkyard entrepreneurs, they crawled into the industry at the ground floor. They sold rebar, those ugly gnarled rods that reinforce concrete.

Good riddance, the big steelmakers said. They weren’t making money with rebar, anyway. But by ceding those lowly markets, they were unwittingly feeding a powerful insurgency. Through the years, the minimills, led by Nucor Corp., invested in new technology and refined their processes. Every time they entered a different steel market, prices collapsed, forcing out the traditional steelmakers. By the time I got to Pittsburgh in 1992, the beleaguered big steelmakers would say defiantly: “But do you think Nucor can make the hood of a Cadillac?”

Nucor is now the nation’s biggest steelmaker. Its market capitalization is $14.1 billion. That may not sound like much, but it’s more than four times the value of US Steel. Even if Nucor steel doesn’t go into Cadillac hoods, the company has helped itself to a big chunk of the market. What’s more, the traditional steel makers have had to invest in exotic technology to hold onto their shrinking top-end, from car hoods to shiny stove tops and washing machines.

Something similar has been happening in technology. In the 1950s and ‘60s, IBM grew into the world’s biggest and richest computer company by selling expensive mainframe computers to companies around the world. Mini computers, personal computers and most recently, cloud computing, ate away at that market. But IBM could still sell hardware to companies that demanded top-of-the-line performance, reliability and security.

As commodity computers claimed more markets, IBM shedded much of its hardware business, turning to more lucrative software and services. But Big Blue still clings to mainframes (in large part because they link customers to the company's software and analytics services). The newly unveiled Z13 will sell for $200,000.

As commodity producers invade a market, they always leave opportunities at the top. But those markets shrink. With time, the usurpers dominate, while the displaced champions, if still alive, are left servicing a small and shrinking niche.

Often, the old champions embrace the new technologies. I read that U.S. Steel is considering building an scrap-melting furnace at its Gary Works. This would have been heresy in the ‘90s. And IBM is pushing hard at cloud computing. But it's challenging for old companies to adapt their entrenched processes and culture to the new ways. It requires rapid change, and this hard work is often up to executive teams that have spent 30 years climbing up through the old system--and are only five or 10 years from retirement.