Sixty years ago this October, while Elvis Presley’s “All Shook Up” was topping charts, a simpler tune picked up by a BBC monitoring station shook the world. A short, repeating string of boops and beeps turned out to be the broadcast of a distant metal orb the size of a beach ball with four long antennae. The Soviet Union’s Sputnik 1 had become the first artificial satellite launched into orbit. A global space race followed.

Today billionaire technology titans are pouring money into space, creating the fastest pace of change in decades. SpaceX, a side project of Tesla(ticker: TSLA) founder Elon Musk, has grabbed a leading market share in commercial satellite launches by undercutting its main rival, Europe’s Arianespace, by up to 40%.

Musk’s long-term goal is to colonize Mars. Jeffrey Bezos, founder of Amazon.com (AMZN) and privately owned Blue Horizon, plans to compete on launches, too, and has a vision of moving manufacturing to space. Richard Branson, of Virgin Group, hopes to sell space flights to vacationers. Other players are planning to mine meteors for precious metals.
 
 
SpaceX Iridium-1 Launch, Jan. 14, 2017 SpaceX


Launch costs have fallen more in the past decade than in the prior whole of the space age.

Meanwhile, CubeSat, a university-developed standard for small, cheap satellites, has opened the market to new players. Some of the better-financed ones aim to launch communications satellites into low orbit by the hundreds or even thousands, covering the globe with data services and upending incumbent players that rely on a handful of costly, higher-orbit satellites.

Public investors are best off staying clear of the upheaval in commercial satellites, and focusing instead on exposure to the U.S. defense and intelligence sectors. Goldman Sachs predicts that spending there will grow 6% a year, compounded, over the next five years. That will benefit aerospace giants like Boeing (BA) and Lockheed Martin (LMT). Orbital ATK (OA), a much smaller defense contractor, looks like another winner.

On the commercial side, there is already a cyclical downturn, because new satellites offer so much more bandwidth than those they replace that demand has yet to catch up. ViaSat (VSAT), a provider of broadband internet service from space, could be vulnerable to future swarms of small satellites.

Nascent space booms have fizzled before. In the late 1990s and early 2000s, insurgent satellite players included Globalstar, Iridium Communications, and the Bill Gates–backed Teledesic.

All went bust; the first two have since relaunched. But it doesn’t take a rocket scientist to understand why past investors have turned to the space business.
 
 
 
 
A traditional satellite takes two to three years to build and costs $300 million—half for the hardware, a third for the launch, and the rest for insurance and other costs, according to estimates from Goldman. Once operational, it can generate $40 million to $60 million a year in cash, for 15 years. Today, there are about 1,500 satellites in orbit, including a quarter for military use, although these account for three-quarters of spending. The U.S. is the biggest player, controlling 40% of satellites, followed by China at 13% and Russia at 10%.

Most launches use multistage, expendable rockets to escape Earth’s atmosphere. The Space Shuttle program, built around crewed, reusable spacecraft, ended six years ago after more than 130 successful flights and two deadly failures. But what’s old is becoming new again: This past March, SpaceX launched a Falcon 9 rocket from Cape Canaveral, Fla., and delivered a communications satellite into orbit, while returning the rocket’s valuable, 14-story main body to a drone ship floating in the Atlantic Ocean. What made the flight unique is that last year, this particular rocket helped deliver supplies to the International Space Station, making its latest payload mission the world’s first with a used rocket, and probably signaling more launch price cuts to come.
 
MOST SATELLITES OPERATE IN what’s called low-Earth orbit, typically 100 to 400 miles away, where they whip around the planet about every hour and a half to two hours. Others operate much farther out, about 22,000 miles away, in geosynchronous orbit, where they travel with the Earth’s rotation to stay above the same point. Geosynchronous satellites provide constant coverage over wide areas, making them ideal for broadcasting television and radio.

They’re useful for two-way communication, too, but with a slight lag—fine for airline Wi-Fi, but not so much for demanding tasks like gaming.

Low-orbit satellites can provide broadband service with minimal lag, but with their limited and constantly changing coverage area, it would take a vast fleet of them to make such service viable. They are used largely for imaging.
 
 
Clockwise from top left: Kevork Djansezian/Getty Images; David M. Benett/Getty Images; Mark Mann/AUGUST; David Ryder/Getty Images
 
   
Enter the disrupters. Privately held OneWeb, backed by funding from Virgin Group, Qualcomm (QCOM), and SoftBank Group (9984.Japan), received regulatory approval in June for a constellation of 720 low-orbit communications satellites. It plans to start launches next year and begin service as soon as 2019. Its goal is to bring affordable, high-speed internet access to the entire planet by 2027.
 
In May, SpaceX told a U.S. Senate committee that it wants to launch more than 4,000 small satellites into low orbit by 2024. SpaceX’s Musk could be looking for a cash cow to better compete with Bezos, who briefly dethroned Bill Gates in recent weeks as the world’s richest man. Bezos’ Blue Origin announced in March that it has its first paying customer for its heavy-lift rocket, New Glenn, which will enter service in 2020. Another Blue Origin program called New Shepard aims to provide suborbital space tourism, allowing passengers to experience cheek-stretching takeoffs and several minutes of weightlessness via a reusable rocket and capsule, with crewed launches planned for next year.

Virgin Galactic, which is also pursuing space tourism, is testing a satellite launch service that will get a boost from repurposed jumbo jets from its airline sibling, Virgin Atlantic. Not to be outdone, Microsoft(MSFT) alum Paul Allen unveiled the world’s largest plane in May, called the Stratolaunch aircraft, with triple the wingspan of a Boeing 737. It is meant to carry three launch vehicles, each of which can put a 1,000-pound satellite into low orbit.
 
THINGS ARE STARTING to look bubbly. “My business model right now…is I sell about $1 billion of Amazon stock a year and I use it to invest in Blue Origin,” Bezos told reporters in April.

A report last year by the Tauri Group noted that 50 venture-capital firms invested in space in 2015, with spending eclipsing the prior 15 years combined. SpaceX’s latest money-raise of $350 million values it at $21 billion. By comparison, L3 Technologies (LLL), a longstanding supplier to government space programs, which generates $900 million a year in excess cash, has a stock market value of $14 billion.

SpaceX investors, including Google parent Alphabet (GOOGL) and Fidelity Investments, aren’t in it for a quick cash-out in an initial public offering. Musk has said he won’t take SpaceX public until he is on the way to realizing his long-term goal: to build self-sustaining colonies on Mars, complete with, as he described it at a conference last year, iron foundries and pizza joints. Bezos’ long-term vision is closer to home but no less ambitious: heavy industry will one day be moved into space, where access to solar power is abundant, leaving Earth for homes and light industry. 
 
ViaSat isn’t sitting still. It has plans to launch three satellites in the next four years to increase its data capacity and expand coverage. But free cash flow isn’t expected to turn positive until 2021. The problem is that consumer broadband services are squarely in the crosshairs of the new low-orbit players, so ViaSat could face more competition by the time it reaches its satellite payday. Shares have fallen 17% over the past year.

Among military providers, Lockheed has more exposure than Boeing. Space systems make up 18% of its revenue. But Orbital ATK has more riding on space: Its satellite division brings in 26% of revenue, and a flight unit that makes rocket systems for both weapons and launch vehicles contributes 35% more. The shares are up 43% in a year but still look attractive.

Orbital is expected to grow overall revenue 4%, to $4.6 billion this year, and earnings per share, 11%, to $6.17, leaving shares at 17 times earnings. For investors looking for gee-whiz appeal, consider that Orbital is the first mover in a business that it expects will become lucrative in a world awash in space gear: launching machines that can coax new life out of expired satellites.

BUBBLES AREN’T ENTIRELY BAD. A Florida land boom in the 1920s ended with a painful crash but left behind new roads, rail stations, and entire cities. The dot-com stock craze of the late 1990s got the timing wrong but the emphasis right. America’s five largest companies today sell software, smartphones, and services that revolve around the internet.

If space investing is indeed a bubble, it might prove to be a benign one, because most of the players who will take losses can afford them, and some of the stuff left behind could prove world-changing. Half of humanity lacks reliable internet access. New imaging technology is advancing the study of Earth’s economic activity, conflicts, and natural resources. Other developments will rival science fiction for wonder.

A company called Deep Space Industries sees potential to send mining vehicles to meteors, grind their metals into powder, and use that to make space infrastructure through a manufacturing process called 3-D printing.

If that idea sounds a little out-there, consider that the biggest obstacle it faces is launch costs. After launch, much of what happens in space is relatively cheap. Let Musk, Bezos, Branson, and Allen clobber one another with their wallets to bring those costs down, and the innovation that follows might astonish.