“We’re going to be building them as fast as we can and that still won’t be enough,” Gedmark says.

Astranis is one of a host of companies, ranging from major players like SpaceX and Amazon’s Project Kuiper to startups like Swarm Technologies and Sky and Space Global, that are racing to loft satellites to provide Internet connectivity. Some of the schemes, like SpaceX’s Starlink system, aim to offer Internet worldwide and compete with traditional telecom and cable services on the ground. Even in Astranis’s niche to serve currently underconnected regions, there are multiple players, such as Google’s Loon balloon Internet service and startup GapSat. The race for funding already has forced some to drop out, like startup LeoSat, which shut down last year.

All of the companies are trying to take advantage of the plunging cost of both building satellites and launching them into space. Increasingly capable and tiny computers can fulfill functions that required larger, more power-hungry analog components on older satellites. And new commercial launching companies like SpaceX and Blue Origin have reduced the cost of sending satellites into orbit. That’s helped ignite a new space race, with startups raising almost $6 billion last year for space-focused businesses, up 38% from 2018.

“The capability that we have now in space has crossed the chasm to do things that we just couldn’t do before,” says Ethan Batraski, a partner at Venrock who also sits on Astranis’s board. “It’s now opening up a whole ecosystem and making the space economy a real thing.”

At Astranis, the strategy is a hybrid of old and new. For decades, major satellite communications companies like Hughes and Viasat have lofted bus-sized craft into high orbits, 23,000 miles up, that remain stationary over one point on the ground. Newer entrants like Project Kuiper, Starlink, and SoftBank Group’s OneWeb plan to use hundreds or even thousands of smaller, cheaper satellites in orbits that are only a few hundred miles up to blanket the planet with coverage.

However, Astranis is building smaller, cheaper satellites but then putting them in high, geosynchronous orbits. That means each satellite can provide coverage to a specific country or region at a much lower cost than a typical geosynchronous satellite. And the company doesn’t need to put up hundreds of small, low-orbiting satellites to open for business. 

“This is one of the world’s most epic problems–we have four billion people not online,” Gedmark says. Reducing the cost “is what you have to do if you’re going to get four billion people online. Many of those people are in lesser-developed countries and can’t afford the kinds of prices we’re paying for Internet today.”

The business model is straightforward. Each contract with an Internet service provider or other customer covers the cost of the satellite dedicated to the specific customer’s area of coverage. The first Astranis satellite, going up on a SpaceX rocket at the end of this year, will bring in tens of millions of dollars of revenue under contract with Alaskan Internet service provider Pacific Dataport, Gedmark says, declining to provide more details. Pacific Dataport will sell service to consumers and businesses, so Astranis doesn’t have to do retail sales. Under this model, Astranis never has to build satellites and hope customers will come later—the deal is made before the satellite goes up.

Andreessen Horowitz led Astranis’s previous fundraising round in 2018. It’s the only space play in the famed venture capital firm’s portfolio. That’s because most space startups are doing more hardcore rocket design and other hardware endeavors that fall outside of Andreessen’s focus on Internet and software startups, says general partner Martin Casado. But Astranis relies heavily on custom software for the radios on its satellites and is attacking a market, communications, that the firm knows well, he says. 

Astranis was also the first to go after the opportunity of putting small satellites in high orbits when almost every other company was going for low orbits. “We love contrarian views because that’s where we believe differentiation is,” Casado says.

More must-read stories from Fortune:

—The strange tale of Jeff Bezos’s $16,840 parking ticket bill
—Post-Brexit U.K.’s surveillance practices could spell problems for business
—Governments deploy surveillance tech to track coronavirus victims
—How marketers are increasingly using A.I. to persuade you to buy
—Predicting the biggest tech headlines of 2020

Catch up with Data Sheet, Fortune’s daily digest on the business of tech.