Cox appears to make an especially compelling example of thriving competition. Consumers love its new “Unbelievably Fair” wireless plan and have been signing up in droves. Instead of “rollover minutes,” Cox issues cash refunds for unused cellular minutes at the end of the month. Customers get back a nickel for every unused minute, up to $20 a month. Just last week Cox announced it was expanding its service area again, this time to include Northern Virginia. No wonder AT&T mentioned Cox 58 times in its main regulatory filing laying out the vibrant competition that justifies the proposed merger.

All of which makes the latest news a bit awkward. Cox, which had been slowly backing away from the idea of running its own cellular network, has come out and effectively put a bullet in the whole idea. Cox’s president said this week that the company could sell the whole network to AT&T (T), Verizon (VZ), Sprint (S) or another competitor, or maybe just sell it for parts. A Cox spokesperson adds that while company will continue to resell service over Sprint’s network, as it is doing now, it will stop building its own 3G network (which was never operational) and has no plans to construct a new 4G network.

Cox’s cellular retreat does more than just knock out a major potential wireless entrant that had pieced together over $500 million of prime wireless spectrum. If Cox’s analysis is accurate, then it calls into question the ability of any new entrant looking to compete by building a nationwide wireless network from scratch.

Cox’s experience indicates the problem isn’t customer demand. In fact more customers have been signing up each month for the service than Cox’s business plan anticipated and those customers have run up even bigger bills than Cox had been counting on. Yet even with that great response, Cox still looked at the economics of running a new wireless carrier over the next decade and couldn’t see a way to make it work.

What did Cox see that scared them? The next decade in mobile phones won’t be like the last. The main business of selling voice minutes is finally in decline after three decades of growth. The hyper-profitable text messaging business is also under attack. The big carriers say growing data revenues will more than make up for those declines. Cox apparently didn’t agree. And it doesn’t help that telecom investors have been having a terrible 21st century, averaging significantly negative returns on all sorts of networks.

It’s been a particularly rough stretch for wireless carriers with business plans that call for building brand new networks. Lightsquared has struggled to raise the cash to build out its planned $8  billion network. The company recently got hit with the news that its network was interfering with GPS signals. It’s still possible that the company can fix that problem, but even if it does, it still faces the same market dynamics that scared off Cox.

Clearwire (CLWR) faces similar problems too. No one has stepped forward to put up the rest of the cash to build out the rest of its network. The stock has steadily sunk and the company recently appointed industry legend John Stanton as interim chief executive in an effort to recraft its business model.

The economics of smaller networks are not inherently unworkable. MetroPCS (PCS) and Leap Wireless (LEAP) have both have focused on building regional networks that target urban consumers who are cheap to serve. That targeted strategy can keep costs below the bigger networks, but it does nothing to provide new competiting networks in much of the county.

In April, Ma Bell described Cox as the leading example of how “new non-traditional entrants to the wireless marketplace are further enhancing the competitive landscape.”

Let’s hope not.

Editor’s Note: This story was updated to include an additional comment from Cox on its wireless plans.