Imagine if every mobile device had its own personal fat-pipe ethernet connection — without the CAT5 cable. That’s how Steve Perlman — inventor, entrepreneur, and CEO and founder of OnLive, the games-on-demand system — explains distributed-input-distributed-output (DIDO) technology, an experimental wireless communications system that could render cellular connections obsolete.

If a cell tower today broadcasts on channels that have a capacity of 100 megabits of bandwidth per second, and 100 people connect to that cell tower and share bandwidth equally, each person’s connection will measure roughly one megabit per second. If 1,000 people connect, each will get 100k bits per second. With DIDO wireless signals, everyone within range would get the entirety of the channel.

“I know that sounds impossible,” says Perlman, “but literally if you have a cell that has 100 megabits per second worth of bandwidth in it and you have 100 people, each person gets 100 megabits a second. It’s really pretty amazing; you don’t interfere with anybody else.”

Wired.com caught up with Perlman at the after-conference reception at NExTWORK when he began to describe his quest. There were cocktails — and we had other reasons to be skeptical of what we were hearing since it seemed to defy the laws of physics. But here’s the intriguing pitch he elaborated upon when we sat down with him for an hour or so a few days later.

Amateur radio licenses in hand, Perlman and his team at another of his startups, Rearden Companies, invented completely new radio technology, which he claims is simpler and cheaper than the innards of modern cell phones. DIDO’s feature list almost sounds too good to be true:

• Its “unlimited bandwidth” will eliminate dead zones and dropped calls, even in an urban jungle like New York City.
• The signals will pass through solid objects that block cellular signals at the same frequency and power.
• It doesn’t need tall cell towers — just modest base stations the size of an internet router.
• Those access points will broadcast a signal over a mile, while outdoor antennas can reach 30 miles or more in every direction — beyond the curvature of the earth, brags Perlman. Theoretically, that number will rise to 250 miles once Rearden’s engineers have time to test the tech at a longer range.

Naturally, this didn’t happen overnight — DIDO has been in the works for about 10 years. Everybody wins in this scenario — except perhaps current wireless providers, who, capitalizing on the “scarcity” of wireless bandwith, are all moving to tiered pricing models and only slowly rolling out 4G networks and depending on their data business to make up for their commoditized calling-plan business.

And then there is that whole pesky science thing.

In 1948, mathematician Claude Shannon formulated the concept of channel capacity. Shannon’s Law, as it became known, states that the maximum rate at which error-free data can be transmitted is a function of the bandwidth and the signal-to-noise ratio. No communications system has surpassed Shannon’s theoretical speed limit — until now, asserts Perlman. He says Rearden engineers presently run DIDO connections at 10 times the limit, know they can achieve 100 times the limit, and are optimistic they can push it to 1,000 times faster or more. It’s easily fast enough to run OnLive, the video game, and soon computing, streaming service that operates in the cloud.

“Everyone we called — you know, like professors and Ph.D students — were like, ‘You’re crazy, this’ll never work, we all know that wireless doesn’t work that way.’” says Perlman. “We had another person to whom I said, ‘Look, everyone’s been telling me this can’t possibly work. I just need to know why.’”

Tired of rejection without explanation, Perlman hired the researcher to disprove DIDO. He couldn’t. In fact, he discovered that not only had Perlman and his team at Rearden had done something that nobody else had thought of, but it worked remarkably well. “That was really the first time I got official confirmation that we were not rabidly insane,” laughed Perlman.

Some still doubt that Shannon’s theorem can be violated, given that it has been proved mathematically. “I think there is essentially no chance that there is a mistake there given that it is such a well-studied theorem,” says Kyle Cranmer, assistant professor of physics at New York University. “However, the assumptions that the theorem is based on may be violated, in which case it’s not applicable, not that it’s wrong.”

For others, overcoming Shannon’s Law is old news.

“Multiples of the Shannon limit have been achieved already using multiple-input-multiple-output (MIMO) technology … which is used in the latest WiFi (IEEE 802.11n) and 4G cellular wireless systems,” says Shivendra Panwar, electrical engineering professor at NYU Polytechnical Institute. “Of course, further innovation in this area is always possible. The multiples usually discussed, for practical cases, are like are like two, four or eight, not 10 or 100.”

Perlman hopes DIDO, which he has already patented, will be available to consumers in a few years time. In the long run, he envisions DIDO completely eliminating wired data connections altogether, bringing about a complete transition of computing to the cloud.

“I am as confident that [DIDO] is going to revolutionize communications as I’ve ever been confident in anything I’ve ever designed in my career,” says Perlman. “That doesn’t mean I’m right, but for example I was far less confident that OnLive or MOVA were going to work, and here we are.”