This article is first published by Forbes.

Size is an asset until it’s not. Within the space of one week, J&J, GE, and Toshiba were all disassembling themselves. Companies over 100 years old suddenly found their current form unfashionable. General Electric will separate into three listed groups, each focusing respectively on aviation, healthcare, and energy. Johnson & Johnson will spin off its consumer products. Even the Japanese conglomerate Toshiba can’t withstand the relentless pressure and will split into three firms.

Simply put, investors view these companies as less valuable should they continue with their previous business models. They want them to work as independent entities.

It’s easy to declare that giant conglomerates are falling out of fashion. But they’re not. The world has come to embrace a new type of conglomerate. We must ask ourselves, “What is Amazon?” Its e-commerce retailing bears little resemblance to its cloud computing services. Microsoft no longer focuses only on office productivity – Xbox is a major player in the gaming industry. And Tesla is not just a car company; its buildup in solar panels and superchargers belongs not to the automotive industry, but to the energy sector.

Tellingly, Amazon, Microsoft, and Tesla are among the world’s most valuable companies. They don’t suffer from being conglomerates. There is no discount on their share prices. Compared to their industry peer groups, they enjoy a premium.

The right question to ask is this: Why do some companies suffer a conglomerate discount, while others enjoy a premium? Conglomerates are here to stay. But what will they look like in the second half of the 21st century?

Don’t be big for the wrong reason

The industrial revolution gave rise to big companies. In the 19th century, steam engines and assembly lines transformed the industrial landscape. Heavy machines required heavier capital investment. Small was not beautiful; it was simply vulnerable. Vanderbilt, Carnegie, and Rockefeller understood that big companies had access to bigger capital and worked to gain access to that capital. Flush with money, they could buy off competitors and merge operations. Those captains of industry became robber barons.

Exploiting the combined advantage of being an industrialist and a financier is not uniquely American. The Japanese conglomerates, keiretsu, are the true masters. A typical keiretsu is Mitsubishi. The Bank of Tokyo–Mitsubishi sits at the heart of this nexus of operations. It supports other Mitsubishi businesses by trading shares of parent company Mitsubishi Shoji. It finances the industrial manufacturing of Mitsubishi Electric. Meiji Mutual Life insures all the other members of the conglomerate. As many as 30 companies are spread out across a huge range of industries.

Outside of Japan, you have chaebols in South Korea, and in Taiwan and elsewhere in Asia are seemingly unrelated businesses linked by family ties. What they have in common is access to capital.

However, General Electric figured something out.

How management practices are sold to investors

What GE understood is the democratization of corporate finance. In the United States, access to capital alone doesn’t translate into any substantial advantage. During the mid-20th century, big banks like JP Morgan Chase and Bank of America learned how to lend to companies of all kinds. Private companies could list themselves on the stock exchanges. Publicly listed companies could issue bonds to raise debt. A fledgling venture capital sector emerged to fund startups.

GE, however, found magic in its management system, and sold that story to Wall Street. Smart analysts at a sprawling company’s headquarters can give it a real advantage. The headquarters can provide real expertise in accounting control, it can force local operations to slash costs, it can fund moonshot projects in research and development. Those projects require the patience that a local manager often lacks. Moreover, headquarters can pioneer a new manufacturing process like six sigma, roll out its standard and reap the benefits worldwide. Headquarters should build a corporate university that brings in top talent from around the globe.

GE’s pursuit of management practices put it in the top spot. People, culture, and organization were said to be the common glue, and investors bought the story. Business schools love writing about this. Sir George Simpson, a legendary British executive known for turning around struggling companies, once teased GE’s CEO Jack Welch. He asked, “Jack, how do you do it?” Simpson replied: “How do you get 50 P/E with that bag of shit you’ve got?” Simpson had noticed that several of GE’s businesses were pretty average, but in terms of the price to earnings ratio (P/E) they were stellar. These mediocre businesses were riding on GE’s halo, but that halo would soon come undone.

Digital glue is the new conduit

What Amazon, Tesla, and Microsoft have come to represent is the third wave of conglomerates. What makes them unique? It’s not capital; everyone has that. It’s not management practices; everyone is now using them. Their uniqueness lies in their prowess in managing data flow. Amazon has built an infrastructure known as AWS; it serves external customers, but is also the glue for its sprawling empire.

Here’s the thing. Data advantage always compounds exponentially. When Tesla knows people’s collective driving habits, it then knows where to put the charging stations. When Tesla knows how long people are willing to wait at the supercharger, it then knows what technologies to invest in. The more business that you engage, the more insights you can glean; as long as you have the data infrastructure that glues all these businesses together.

Amazon likes to talk about “mechanisms”. It understands that to innovate at scale, you must encode the desired behaviors into the way everyone works. You do this in part by training. But in today’s world, you develop digital tools to facilitate innovative ideas. Data is the new oil.

From this angle, the story of GE, J&J, and Toshiba is simple to understand. Their disparate businesses have no relevance to each other because their data remain siloed and they have failed to imagine how to unleash their knowledge across different divisions. And investors don’t think they could either. What’s the point of companies having all these under one umbrella if they don’t glean any insights? So, let’s split them to cut overhead; and make some savings by getting rid of the fancy headquarters.