There are few industries with higher barriers to entry than wireless networks. You either have to be crazy or have insanely deep pockets to compete against the likes of AT&T , Verizon , T-Mobile , and Sprint .
Social Capital’s Chamath Palihapitiya, a former Facebook executive, said on Tuesday that he intends to build an American mobile carrier called Rama. It’s good timing—the window for building a fifth major American telecom is closing, and soon, it may not be possible anymore.
But before Rama sells a single phone or data plan, it must first spend billions of dollars to buy spectrum at auction.
A full-fledged mobile operator needs exclusive rights to certain radio frequencies for its network to work, and the United States is running out of underused frequencies. In the wireless industry, the rights to those frequencies are called spectrum and are administered by the Federal Communications Commission. You can think of spectrum like real estate—it’s valuable, and they’re not making any more of it. For example, the block of spectrum currently used for mobile phones is already divided and claimed by the big carriers.
Next year, the Federal Communications Commission is going to auction off blocks of spectrum previously used for TV broadcasts. They’re likely to be last pieces of spectrum well-suited for cell phone networks to go on sale in the United States for a long time, and Rama plans to bid for them.
Palihapitiya believes Rama will spend between $4 billion and $10 billion to win the spectrum it needs to run its network. He reportedly already has big-named investors lined up to provide the funding, although he declined to name them.
But Rama is going to have to compete with some deep-pocketed companies, including the carriers he wants to take on. AT&T, Verizon, and T-Mobile have already declared they plan to bid on the spectrum, and they could be joined by technology firms like Google, Charter, Dish, and Comcast as well.
Next year’s auction is shaping up to be a bidding war for the radio frequency equivalent of beachfront property, and spectrum isn’t cheap to begin with—the last FCC spectrum auction closed at nearly $45 billion this past January.
Even if Rama ends up winning the spectrum it needs, it faces an uphill battle. First, the company will have to start the slow and expensive process of actually building the towers and infrastructure that make up a wireless network. Palihapitiya said he wants to use “microcells,” or tiny cell towers installed on people’s homes, to help build the network quickly and provide better coverage.
As for what Rama would do differently than the incumbent carriers, it would revolve around software from LotusFlare. Currently, LotusFlare makes software that lets people in developing regions monitor their data usage, and it could be a major improvement over the customer service and billing options currently offered by the big carriers. Earlier this year, Google reportedly tried to acquire LotusFlare.
Rama could also sell itself on a simple billing structure, like Google’s Project Fi, which transparently charges customers based on the amount of data they use.
“There’s a whole bunch of issues that carriers aren’t trying to solve. All these things are now just totally broken,” Palihipitiya told Business Insider.
But before Rama starts building its “new, modern carrier,” first it has to win the spectrum auction. If it doesn’t win, it’s going to be nearly impossible to build a national mobile operator. Palihapitiya’s plan to take on AT&T and Verizon might be a long shot, but it’s now or never.
For more about wireless carriers, watch this Fortune video: