Economist Mariana Mazzucato likes to describe the investment world as a landscape; a “bumpy and complex one,” as she puts it. Government’s job—when government is visionary and well-run and not dogmatically following Keynesian or libertarian ideology—is to get out ahead and smooth the way. NASA and other agencies invest in new mission-oriented technology “until fear-inducing uncertainty” gradually levels into “mere risk.” Economists argue about the difference between uncertainty and risk; but as an engineer and physicist, I like to use numbers to distinguish between the two. In short, risk has numbers. Uncertainty lacks them.

Take the weather as an example. You turn on the radio or check your smartphone before you go to work, and the forecast says there’s a 30% chance of rain. That means there is a 70% chance of a nice day. If you leave without an umbrella or raincoat, you risk getting wet. The risk, according to the forecast, is about one in three. You know the odds, they seem to be in your favor, and you take the risk. It has a number. But imagine it’s pitch dark outside, you’ve lost your phone, and your wifi and radio are on the fritz. You can’t get the weather forecast and have no idea what it’s like outside. Worse, you don’t know what the weather will be this evening when you return home from work. There could be a blizzard for all you know. Or a hurricane. The clouds looked pretty weird the night before. But you can’t put a number on last night’s clouds. That’s uncertainty. And what do most people do in the face of uncertainty? They behave more cautiously than they usually would. You probably would be more inclined to bring a raincoat. Or, if you feel a little freaked out about the uncertainty, you might even stay home. Better safe than sorry.

Similarly, imagine you’re a venture capitalist investing in an industry where an average three out of every four startups fail. Among the successful companies, just one out of eight doubles its value every year for the first five years. What are your odds of making a good investment if you backed only one company? How about backing four companies? Ten? While the math is complicated, a good computer algorithm can parse the odds precisely. Will it eliminate your risk altogether? No. But it can tell you exactly how much risk you would be taking.

Now suppose you are offered an investment in a market that does not even exist yet—extraterrestrial mining, say. Two entrepreneurs come to your venture capital firm and make a pitch. They want to start a company that captures a small asteroid, parks it at one of the Lagrangian points, and then extracts valuable minerals. Some will get sold to other companies for use in space construction, and some will be brought back to Earth. “We’re talking minerals worth as much as billions!” says one of the entrepreneurs, showing a beautiful slide with that beautiful number on it.

“How much of an investment do you need?” you ask.

The entrepreneurs look a little embarrassed. “Three billion,” one says.

“Three billion!”

“Actually, three billion a year,” says the second entrepreneur.

“For ten years,” says the first.

All right. In this scenario, suppose you actually have the money. Your firm has made a killing in Silicon Valley. But what are the odds that the venture will succeed? Thirty percent? Sixty? You have no idea. There are no odds, because there is no record of companies’ success or failure. You have no investment version of a weather forecast. You’re completely in the dark. While the idea of vast profit has a certain appeal, your firm would not see that profit for at least ten years. What venture capital company in its right mind would make that kind of investment?

While I made up the eager pair of entrepreneurs, I did not make up the extraterrestrial-mining part. A great many highly intelligent, well-informed people are working up plans for the industry right now. The only problem is, the industry won’t exist for at least another two decades. Even five years requires a vast amount of patience for a venture capitalist. And VCs tend to invest at a relatively safe moment in the life cycle of a company, after a market exists and industry is defined. Many startups face a dangerous time, when they have burned through most of their capital and have not yet begun to receive significant income. Economists and investors use a term for this phase, borrowed from the tech professions: the Valley of Death. Most of the big money waits for companies to get through the valley, proving their business model. If markets and industries make up the terrain that includes valleys of death, then government programs allow new terrain to be discovered. Venture capitalists like to help fast-growing young companies with the means to expand, tap a larger market, and then reap the profits when those companies get sold or go public.

New industries have not even reached the Valley of Death. When the brilliant physicist Richard Feynman first conceived of the field of nanotechnology in the 1950s, for example, the term “nanotechnology” had not even been invented yet. (A scientist at a public university in Japan coined the term in 1974, and an American engineer at MIT popularized the concept with a book in 1986.) The nanotech industry is just getting off the ground, with nanomaterials a billionth of a meter thick being employed in sunscreens and carbon-fiber bicycle components, among other uses. But the industry remains in its infancy, and the far-out, science-fiction-worthy products—nanorobots repairing patients’ organs, clothing that can change color and shape—require patient, and uncertain, investment over decades. That is why the U.S. government created the National Nanotechnology Initiative in 2001, with an investment totaling $21 billion through 2015 and continuing today. An effort combining universities, private companies, and government agencies (including NASA), the initiative works to create a trove of knowledge ore and pass it to the private sector as a whole, all while helping train a workforce skilled in the new industry.

Extraterrestrial mining and nanotech are two examples that speak to Mazzucato’s assertion that the State “often takes the lead, not only to fix markets but to create them.” She notes that in biotechnology and the Internet, venture capitalists began investing 15 to 20 years after public funds had been committed to the industries.