The real problem with space is there’s not enough of it—at least not when it comes to places to put people. With billionaires funding their own space programs and investors pouring cash into new start-ups aimed at building a true space economy, millions of humans may end up working and living away from Earth in a century or two. If that happens, all of those people will need somewhere to live. But as of yet, no would-be captain of space industry has proposed a viable housing plan.

Mars, which gets a lot of attention as ground zero for humanity’s future, is really just a good place to die. The Red Planet is a frozen desert with a thin atmosphere and no protecting magnetic field. The threat from high-energy solar radiation is so extreme that cities would have to be built underground or covered by thick domes. And the surface of Mars (or any planet or moon) sits at the bottom of a deep gravitational well (that is, a high-gravity region) that rockets must climb out of or carefully drop down into. This makes visits between the surface and space an expensive proposition: It takes a lot of rocket fuel to climb up and down those wells, and rocket fuel doesn’t grow on trees (and trees don’t grow on Mars).

Any serious efforts to live outside Earth will have to go beyond the immediately obvious destinations such as Mars. There are options that don’t involve the great expense of climbing out of gravity wells and the great danger of falling back down them again. Based on recent work, my colleagues and I may have found a candidate with none of Mars’s tricky problems. If we’re right, our real future in space may be not on the surface of planets, but inside asteroids.

NASA has been considering non-planet, non-moon options for our future space housing as far back as the early 1970s, when it began exploring designs for “space habitats.” Then, in 1976, the Princeton physicist Gerard K. O’Neill published The High Frontier: Human Colonies in Space. The book, which became an instant classic, was full of plans for permanent space cities contained in shining metal cylinders, the largest of which was 20 miles long and four miles across.

[From the January/February 2015 issue: 5,200 days in space]

These “O’Neill cylinders” addressed the biggest and most fundamental problem of not living on a planet: no gravity. Research indicates that when people live in zero gravity for more than a few months, their eyeballs bulge out, their retinas can detach, their muscles atrophy, and their bones become brittle. While there aren’t many data on the problem, evidence suggests that humans may require somewhere around one-third of Earth’s gravity in order to function properly. That’s one reason people such as Elon Musk are so bullish on Mars: Its surface gravity is just above that one-third limit. An O’Neill cylinder wouldn’t be massive enough to generate significant gravity, but O’Neill designed them to rotate around their long axis, producing gravitylike centrifugal force. Residents would live happily on the cylinder’s inner surface, pulled “downward” away from the center.

The stunning artwork that accompanied O’Neill’s designs showed futuristic communities set amid beautiful, landscaped parklands where the horizon curved upward. The pictures fired the imagination of a generation of space nerds, including me and, more notably, Jeff Bezos, who owns the rocket company Blue Origin and is a big fan of space habitats. But O’Neill cylinders have a crucial problem: Bringing millions of tons of raw materials up from a gravity well (even a shallow one such as the moon’s) and then fabricating them into the necessary steel beams, trusses, and construction essentials would be prohibitively expensive. I’ve seen an estimate of $100 trillion a cylinder.

One could conceivably get the material from asteroids, which don’t have significant gravity. But even after grinding the asteroids down, you’d still need giant space factories for transforming the extracted raw materials into O’Neill cylinders. That wouldn’t be cheap either. The one thing asteroids have going for them is their sheer quantity: Thousands of the flying space mountains zoom around on orbits that pass near Earth. Actually, if you think about it, that’s a lot of potential real estate.

[Read: NASA is practicing asteroid deflection. You know, just in case.]

To be clear, we can’t live on asteroids (too little gravity, too much radiation). And we can’t live in metal tubes made from asteroids (too expensive). But there may be a third option: living inside asteroids.

It’s an idea that science-fiction writers, at least, have taken notice of for more than a decade. Hollowed-out space rocks, set into rotation to generate centrifugal force, play a central role in Amazon’s TV series The Expanse (based on the books series of the same name). In that fictional future, about 100 million “Belters” live in asteroid cities, where the thick rock walls offer free, natural protection from high-energy solar radiation. (The Expanse is on Amazon for a reason: Bezos loved the show so much that he rescued it from cancellation from the Syfy channel).

But science fiction is one thing; the laws of physics are another. I and my colleagues at the University of Rochester wanted to find out if asteroid habitats are really a possibility. Within the next several centuries, humans may very well have the technology to dig out big living spaces in an asteroid and set it spinning. But when my colleague Alice Quillen, an asteroid expert, and I ran the calculations on that plan, we found that spinning a large asteroid fast enough to create artificial gravity would crack and fracture the asteroid rock matrix, leading to the whole asteroid flinging apart. As a huge fan of The Expanse, I was bummed out by this result. But Alice came up with another approach.

[Read: A handful of asteroid could help decipher our entire existence]

Most asteroids, it turns out, aren’t solid rocks. The ones smaller than about 10 kilometers across are a mixture of sand, pebbles, rocks, and boulders held together by the weak force of their own gravity. Spinning such a body would send its component parts flying into space immediately—unless, as Alice pointed out, you could hold them together. All we needed was a very big bag.

Imagine a swarm of robots easing up to one of these rubble-pile asteroids and covering it with an ultra-strong, ultralight elastic webbing. (Carbon nanowires, which can be as thin as a few atoms, are a good option. They’re currently produced in only small amounts, but we could manufacture them en masse in the future.) Once the cylinder-shaped “bag” is set, the asteroid could then be slowly spun via rocket motors anchored deep in the rubble. As the rubble pile spins faster, it would begin flinging out those pebbles, rocks, and boulders, pushing the carbon-nanowire webbing outward with them.

Rendering of a space habitat made by spinning an asteroid (Michael Osadciw/University of Rochester)

At some point, the bag would hit its maximum extension, and the expanding rubble would slam into now-rigid webbing. Debris flung into the taut bag would be compacted, forming a giant, hollow, concrete cylinder. Once the dust clears, towns, cities, parks, and farmland could all be built on the inner surface of the cylinder, just like in O’Neill’s designs. That surface could be enclosed with a transparent roof to hold in an atmosphere; imagine a bicycle tire with clear plastic running around the inner circumference. Outside the living area, the thick concrete walls would protect against radiation.

[Read: Just like that, we’re making oxygen on Mars]

Led by a very smart engineering Ph.D. student named Peter Miklavcic, our team ran a series of simulations to study the plan. It seemed to work: Our models showed that a small rubble-pile asteroid, just a few football fields across, could be expanded into a cylindrical space habitat with about 22 square miles of living area. That’s roughly the size of Manhattan, where more than 1.5 million people currently live. Multiply that by the solar system’s tens of thousands of asteroids, and The Expanse’s 100 million space dwellers could be easily accommodated.

Humans do not yet have the technology to build asteroid cities. It’s also possible that we will never have the kind of industrial space infrastructure necessary to make this idea a reality. Maybe it won’t be possible to manufacture enough carbon nanowire to make the asteroid bags. Maybe social systems are unstable in permanently enclosed habitats. These are all possibilities. But the paper summarizing our results, which we published last year, suggests that the overall plan is viable. The technologies we proposed using don’t break any laws of physics; they are reasonable extrapolations of capacities we have now. Such extrapolations can seem fantastical, but they prove true more often than you might think. Remember that in 1900, no one had ever flown an airplane. Now, as you read this, hundreds of thousands of people are safely hurtling five miles above the ground at hundreds of miles an hour.

[Read: When a Mars simulation goes wrong]

Perhaps future generations will blaze ahead with Mars settlements anyway, deciding to drop down treacherous gravity wells and take their chances on an unforgiving desert. But I’d rather imagine the solar system hung with thousands of asteroid cities, each a spinning jewel of human creativity and promise.