The startup’s stock price was languishing around $36 on April 10 when AT&T swooped in with an offer to buy the company for $95.63 per share. Then, Verizon—without having its name officially disclosed—topped AT&T’s bid and eventually closed the deal at $184 per share–five times the starting trading price.

But don’t get distracted by the seemingly gaudy bids. Straight Path, which was spun out of telecommunications carrier IDT (IDT) in 2013, owns a huge amount of spectrum around the country in the 28 GHz and 39 GHz bands. And when analysts calculated how much Verizon had paid for each potential customer covered, the amounts were tiny compared to the value of spectrum licenses used to offer current 4G LTE service.

For example, under AT&T’s original bid, Straight Path’s licenses were valued at just under one cent per megahertz covering each potential customer. Even after Verizon’s top bid, the spectrum was worth only 1.7 cents per person.

Compare that to the recently completed Federal Communications Commission auction for airwave rights currently held by TV broadcasters in the 600 MHz band.

Even though it could take up to three years for the TV stations to clear the rights—and even though no current smartphones yet operate in the 600 MHz band—winning bidders paid an average of 88 cents per megahertz per potential customer, and $1.25 in the 40 largest metropolitan areas, according to consultant and strategist Frank Rayal. And that was after a disappointing auction that raised less than $20 billion, half to one-third of what some on Wall Street predicted last year.

Get Data Sheet, Fortune’s technology newsletter.

Why is Straight Path’s (STRP) hot sale still valuing its licenses at pennies on the dollar compared to 4G band licenses?

For one, the economic case for 5G service in the 28 GHz and 39 GHz has yet to be made. AT&T and Verizon are conducting trials to offer super-fast Internet and video service to homes and businesses using the high frequency bands, also known as the millimeter wave bands. But despite being able to carry a lot of data, transmissions in those bands are much more susceptible to interference and don’t travel as far as lower band broadcasts.

That could mean that consumer services won’t be reliable enough, or won’t be cost-effective if the carriers would need too many cell sites to reach a critical mass of customers. While the wireless broadband technology is being tested, cable companies are rolling out even faster and higher-capacity fiber optic wired connections to homes and businesses, cutting into the promised speed and cost advantages of 5G.

The ghosts of telecom bubbles past may also be haunting the millimeter wave spectrum. Almost all of the licenses, including those owned by Straight Path, were gobbled up by startup carriers in the 1990s telecom infrastructure bubble. Almost none succeeded, burning through $1.25 billion of wasted capital. Straight Path’s former parent, IDT, grabbed the licenses from bankrupt wireless pioneer Winstar in 2001.

Other licenses Verizon (VZ) has rights to use were acquired by Carl Icahn from XO’s bankruptcy in 2002. FiberTower, a small company acquired by AT&T (T) in February, spent a decade trying to get off the ground until it filed for bankruptcy in 2012 and had many of its licenses stripped by the FCC.

So some analysts and investors with long memories are wary of the promise of using the millimeter wave rights for business cases that sound eerily familiar, despite the obvious advances in technology that have occurred over the past 20 years.

Sprint and T-Mobile have taken a very different approach, emphasizing the use of lower bands for 5G and seeking to offer fast connections to mobile devices, rather than homes and businesses. T-Mobile (TMUS) said it plans to use some of the 600 MHz rights it just won at auction for 5G, while Sprint (S) this week announced a strategy relying on 2.5 GHz licenses. T-Mobile does have some 28 GHz rights from its 2013 acquisition of MetroPCS, while Sprint leases millimeter band rights.

But until some of the current trials turn into more compelling commercial services, the value of the high bands remains unproven.