The home Internet market in the U.S. is broken, and nobody's laying the cables to fix it. Wireless carriers, for their part, have said they don't have the capacity to fulfill Americans' multi-gigabyte home needs. From the folks who brought you Aereo, here comes Starry, which announced today that it's bringing high-speed Internet to Boston in March.

Starry's founder, Chet Kenojia, is really good at transmitting a reality distortion field. And just like he did with failed TV-watching startup Aereo, he's hitting out at some companies that Americans hate: the cable firms, who are also the major ISPs.

According to Starry, ideally, you'll get an antenna to stick out your window and a router to put into your home. Turn 'em on, and you'll have gigabit Internet with no caps, limits, or worries.

The router and service will be sold separately. The $349 Starry Station router, available Feb. 5, will attach to any Internet connection. It has a touch-screen display on the face that shows you the condition of your Internet service, the speeds you're getting, and lets you manage parental controls. A "Starry Wing" extender will help it reach odd corners.

But $350 is a heck of a lot for a router, even one that helps you diagnose Internet service problems. The really hot deal here is Starry Internet, a new ISP that's initially launching in Boston. Starry hasn't announced pricing for the ISP, other than to say it will be "simple" and "consumer focused." The ISP uses very high-frequency, short-range wireless technology to broadcast Internet throughout a city, a lot like wireless carriers do, but with more capacity and a bit of an easier time distributing Internet. It's only connecting to large, window-based stationary antennas rather than tiny antennas that might be at the bottom of your purse.

That makes Starry a wireless ISP, or a "WISP." You may never have heard of such a thing, but it isn't new.

A WISP of a Hope
Kanojia is trying to make it sound like he has radically new technology. But WISPs have existed for decades, especially in rural areas where it's too expensive to lay cable to very broadly spread-out customers.

In cities, you're more likely to find a WISP strategy being used by business ISPs and cell-phone companies, which use them to tie small base stations to a central Internet connection. Towerstream, which operates in 12 cities, has been WISPing its way into businesses since 1999.

Starry intends to distribute very fast, high-density Internet to homes using "millimeter waves" in the unlicensed 38GHz band. At frequencies like that, you get great speeds but very little range. The range gets even shorter when the air is wet. There are a few smaller ISPs working up there; Monkeybrains, which covers a few neighborhoods in San Francisco, is running at 60GHz.

Those waves have trouble penetrating walls, but that isn't the problem here. Starry will have rooftop base stations (which it calls MetroNodes, or Starry Beams, above at left) that link to receivers you put outside your window (above at right), which in turn will link to that station in your home, which will translate the signal into more conventional 2.4GHz or 5GHz Wi-Fi. 

The challenge is that Starry is going to need to place a lot of those rooftop base stations. They'll probably need to place one every quarter-mile to half-mile, in dense cities. Every one of those base stations will need permission from a property owner, they'll need an Internet connection, and they'll need to be maintained and serviced.

(A relevant tangent: New York University has been exploring using 28GHz and 38GHz for handheld devices, and it got about a 600-foot range from its base stations in tests in New York City. Starry should get better range because of its large, fixed-location window antennas.)

Starry has a much easier task than starting a new wired ISP, which requires digging up streets. And it's an easier task than starting a new wireless carrier, which requires buying expensive, licensed spectrum at auctions that only happen every few years. But it's still a long slog of negotiating base station locations and setting up infrastructure, as opposed to with Aereo, which only required one major location per city.

The big start-ups of the past few years have generally been apps or platforms that use someone else's infrastructure. Uber uses drivers' cars and public roads, and Airbnb doesn't own any rooms. But Starry is going to need to build real infrastructure, a physical network of base stations to support its dreams. Google has been able to do that with Google Fiber. The question I don't know the answer to is, does Starry have the deep pockets to actually build its dreams?