New York CNN Business  — 

SpaceX wants to be your next Wi-Fi provider. It’s hoping to build a constellation of satellites to deliver cheap, lightning-quick broadband from space — and the toughest obstacle could be on the ground.

The company took major steps this year by launching the first 100-plus satellites for its network, called Starlink, with an eye toward growing the constellation to include more than 10,000 devices that will blanket the planet in internet connectivity. SpaceX executives hope initial service could roll out in parts of the United States next year.

The path ahead is filled with risk and uncertainty. Starlink is essentially a multibillion-dollar bet that Elon Musk’s company can deploy a technology that others have tried and failed to build in the past. And the service’s success depends only partly on whether SpaceX can deploy thousands of satellites without running out of money.

Perhaps a bigger and more difficult question, according to industry experts, is how people will access the network. Starlink customers would need user terminals, or high-tech antennas they can stick on their roofs or in their yards to establish a useable internet connection.

The necessary technology exists. But if SpaceX wants to attract consumers, it needs to build user terminals that are affordable, reliable and low maintenance — a massive undertaking.

SpaceX’s chief operating officer, Gwynne Shotwell, said during a recent meeting with reporters that the company is putting “a lot of engineering” resources into developing user terminals.

“Because if we don’t get it right,” she said, “we’re in deep doo-doo.”

Why are ground terminals important?

Even as traditional internet infrastructure has expanded rapidly in recent years, about half the global population still lacks consistent internet access, including millions of people in the United States.

Traditional Wi-Fi and cell services rely on an enormous web of underground cables and cell towers. It would be extremely expensive and time-consuming to connect billions of people using only ground-based technologies.

That’s why there’s big interest in using satellites to blanket the globe with an affordable network. And it’s not a new idea.

In the 1990s, several well-funded ventures attempted to build satellite constellations similar to Starlink. All of them either changed plans, went bankrupt or liquidated after realizing it would be impractical or too expensive.

SpaceX is among a new crop of companies — which include Amazon and Softbank-backed OneWeb — that are trying again. They expect to be successful this time because satellites and rockets are cheaper than ever. While deploying a satellite internet constellation will be far from easy, that effort would be less likely to end in bankruptcy.

Ground equipment may pose one of the biggest obstacle to success.

Antenna technology has come a long way since the 1990s. But even some industry leaders say it’ll still be extremely difficult to build user terminals at the volume and price points that SpaceX will need to turn a profit.

One source involved in Teledesic, a now-defunct internet constellation business backed by Microsoft billionaire Bill Gates, said that the company learned the hard way not to underestimate the difficulty of building user terminals.

It was a key reason why the venture folded in the early 2000s, the source said. In fact, it was “probably” the main reason, the source said.

Why are the antennas so hard?

Space-based data services already exist. They’re powered primarily by massive satellites in geosynchronous orbit more than 20,000 miles from Earth. At that distance, objects orbit at the same speed as the Earth turns, meaning satellites can stay positioned over a specific area of land and provide uninterrupted service.

That could be ideal for telecom services like Dish Network, which provides customers with simple antennas that sit atop their roofs. They can point at the same satellite day and night.

But that system doesn’t work well for internet service. Geostationary satellites require data to travel the 20,000 miles and back whenever a user clicks a link, which can cause frustrating lag times.

SpaceX and its competitors want their internet constellations orbiting much closer to the ground. In Starlink’s case, thousands of satellites would circle 340 miles overhead, working to deliver internet speeds rivaling the fastest ground-based services.

How cheap do they need to be?

Bill Milroy is the chief technology officer at US-based antenna maker ThinKom, which makes computer-controlled antennas — called “phased array” — that steer themselves to maintain a connection with a satellite as it streaks across the sky. The company’s antennas are currently used in commercial airplanes and are used to relay in-flight Wi-Fi to passengers.

Milroy said it’s impossible to know exactly how expensive it would be to mass produce the kinds of user terminals that SpaceX will need.

It’s a classic chicken-and-egg problem: You won’t know how cheap they will be to manufacture until you know how many you’re building, and you can’t gauge demand until you’ve priced the terminals.

Milroy said that if Elon Musk called him tomorrow and told him to build user terminals as cheaply as possible, ThinKom could probably get the cost down to about $1,000 each, about as much as an iPhone 11.

At those prices, Starlink might be able to win over some Americans unhappy with their current internet providers.

But, Milroy added, he doesn’t believe SpaceX will attract customers in the mostly poor areas that still lack internet access. The company would probably need to get user terminal prices down to about $150 a piece, or what “a cheapo phone in India costs,” to make a significant impact, Milroy said.

“We don’t think anybody’s ready to hit those kinds of price targets,” he added.

SpaceX doesn’t plan to buy user terminals made by others. In typical fashion, the company will keep design and production in house.

Shotwell, the SpaceX COO, said a team of engineers have started a prototype production line at the company’s headquarters in Hawthorne, California. They “still have a lot of work to do,” she said.

SpaceX has proven on more than one occasion that it can get the job done. And Milroy, the ThinKom CTO, echoed what many others in the industry have said after expressing skepticism about Starlink: “I certainly don’t want to bet against SpaceX.”