On the Texas Instruments eStore it takes only $19.99 to jump into “the Internet of things” by purchasing a Connected LaunchPad unit to bring an everyday device onto the Internet.

Better be patient, however, because they are sold out.

“They are perpetually sold out,” said Hans Rempel, CEO of the Minneapolis start-up Exosite, which will have its data management software on TI’s LaunchPads when they eventually do ship.

Texas Instruments’ failure to keep up with demand for these evaluation devices is just one sign that the Internet of things, the connection of all sorts of regular machines and devices to the Internet, is beginning to realize its long-discussed promise.

Rempel cares very much about how many evaluation units ship. These little gadgets will go into prototypes that engineers and marketers will evaluate to see if something like a trash compactor can be transformed into a far more valuable Internet-connected, smart trash compactor. Rempel thinks maybe one out of 10 evaluation units could lead in six to 18 months to a machine going into production.

That long lead time, as well as reluctance by customers to replace their unconnected but still-working-fine conventional machines, helps explain why the market is still talked about mostly in terms of its potential.

But, oh my, is that potential ever huge. A conservative estimate by leading consulting firms is something like 26 billion devices connected to the Internet by 2020. Cisco Systems, the network-equipment maker, has predicted nearly twice that many.

Cisco even has a real-time “connections counter” on its website. While no genuine data collection seems to be taking place, as of this week Cisco’s number was approaching 12.5 billion devices.

One result of brand-name companies putting out hockey-stick forecasts is that in 2014 the term Internet of things already seems to fall well into the zone of pure hype.

It’s the kind of term that will induce a lot of eye-rolling in an organization if it appears in the subject line of a clumsily written e-mail from the CEO. It will remind folks of last year’s “what’s our big data strategy?” e-mail.

The concept of connecting machines to computers is not even particularly new. One of the most celebrated stories of invention is from more than 30 years ago. That’s when computer nerds at ­Carnegie Mellon University made some unauthorized upgrades to a vending machine so they could learn while seated at their desks if the Coca-Cola inventory was sold out rather than finding out only after a long walk.

But that little project proved the concept’s value — it’s actually good to know if there’s cold Coke.

It took decades for real businesses to be willing to spend real money on similar solutions, but by 2014, “this is hotter than anything I’ve seen as an investor,” said Guy Turner, managing partner of Hyde Park Venture Partners in Chicago. “What does that mean? It means a lot of very large players have looked around … and said, ‘Man, this is big and I need to figure it out.’ A lot of them are looking to start-ups to solve their problem for them.”

The examples of cool ideas tumble out so quickly in conversation that it can sometimes be difficult to tell what customers are actually paying for today. Exosite has even worked on what Rempel called “a better pest control device,” as close to a “build a better mousetrap” plan as can be imagined. That’s why Turner described the market’s development as just starting, as consumers and businesses of all kinds try to figure out which cool ideas are genuinely worth the money.

There are other companies in the business here in the Twin Cities besides Exosite, including at least two others on the radar screen of venture capitalists, the start-up firms Spark Labs Inc. and Grove Streams.

Spark sells a little piece of electronic hardware to be installed in a device and then connected over a Wi-Fi network, but makes money selling the software and services to manage the data. It’s far better known than Grove Streams, in part because it’s raised money on the popular Kickstarter “crowdfunding” site, seeking contributions from any and all.

Turning to Kickstarter suggests that Spark might be something less than a serious technology company, and Spark sure talked a lot about how much pure fun could be had making a Wi-Fi-enabled toy car or a fridge magnet that displays Twitter posts.

But when its last Kickstarter campaign closed last summer, Spark had raised $568,000, and Spark is in fact being taken very seriously these days by professional investors. This week it had 11 job openings posted on its website.

“We are working with a number of larger enterprises on creating large-scale deployments,” said CEO Zach Supalla, “proving that a Kickstarter-funded business is just as capable of appealing to a professional audience as it is to consumers and hobbyists.”

Grove Streams introduced its software only last fall. Users include a South African start-up launching a patient monitoring platform for clinics. “Keep in mind that most African clinics don’t even have computers,” said Grove Streams founder Mike Mills, a former IBMer. “This system relies on sensors and smartphones.”

With software already out in a hot market, Mills has heard from venture capitalists. So far all of the capital has come from his own personal checkbook.

Exosite has progressed much farther than Spark or Grove Streams, and among the applications now heading into broad distribution is a smart trash compactor produced by Minnesota-based Harmony Enterprises.

Exosite has mostly bootstrapped its growth, and is now approaching 60 people with a presence in Hungary and Texas in addition to co-headquarters in Minneapolis and Taiwan.

Rempel said his company faces a choice of whether to accept a big slug of venture capital to grow faster or perhaps find a large company as a partner to help find and develop new users, the option he favors.

Finding capital isn’t one of his challenges. Ducking calls from overly eager venture capitalists is. As he put it, “we have a VC contact us at least once a week.”