Artemis Networks, demo, Code Conference

Asa Mathat


With the rapid spread of high-end mobile devices and streaming services to feed them, the cellular networks built to carry data in the United States are becoming congested. Wireless spectrum is limited, and the bandwidth delivered by cell towers must be shared by many people, whose devices basically take turns grabbing some of the spectrum.

So even high-speed LTE networks are slowing down — especially in densely populated areas.

At the Code Conference today, startup Artemis Networks showed off a new technology called pCell (which stands for “personal cell”) that it claims solves this problem by creating tiny, virtual cells around each mobile device. The company claims these cells deliver fast, unshared bandwidth to each smartphone, tablet or laptop, even in packed places like stadiums.

The pCells are generated by small transmitters that look sort of like a larger version of a home wireless router and can be placed all over a building or a city. If adopted, they could even one day replace cell towers, according to Artemis founder Steve Perlman, who formerly worked at Apple and Microsoft and helped invent products like QuickTime and WebTV.

On the Code stage, Perlman and his team created a wall of iPads playing high-speed video over a cellular network with pCell technology, something that very likely would have caused endless stuttering and buffering in normal circumstances.

“It’s an interference avoidance game.”

Artemis Networks founder Steve Perlman on the status quo of internet access in areas with many devices.

The pCell system works with standard LTE phones, so no new handset technology is required. And they can coexist with the current cell phone transmission system.

In a February Re/code post, one analyst, Richard Doherty of the Envisioneering Group, was quoted as saying: “[The] pCell is the most significant advance in radio wave optimization since Tesla’s 1930s experiments and the birth of analog cellular in the early 1980s.”

But there’s a catch: Deploying them will require the active cooperation of the major wireless carriers, who may be suspicious of relying on a radical new technology from a small startup. And Perlman hasn’t always succeeded. His last venture, a high-speed game streaming service called OnLive, failed to catch on.

Another option: pCells work with Wi-Fi as well, so the technology could be deployed without the cooperation of cellular carriers. But the company’s main goal is to get the system adopted by the carriers as a way of solving their congestion problem.

Perlman said Artemis is trying to launch in both developed and developing countries.

“We’re overwhelmed, and we’re trying to grow as fast as we can,” he said. “We’ve literally virtualized radio.”

Additional reporting by Eric Johnson.

More Posts From the Spotlight Demos at the Code Conference 2014


“It’s an interference avoidance game.”  without significantly change iPad's modem SW, how can one get reliable channel feedback to avoid interference?  This violates all my knowledge about MIMO and interference. 

If Artemis truly believes that its technology is much superior to the intellectual property of Qualcomm/Ericsson/Huawei combined, which they seem to want us to believe so. Why not buy out all the put options of these companies and then disclose their patents? The market will reward them immediately if their technology is truly so magic. As of today, Qualcomm's $70 put option for 9/20 is only 1 cent.


I'd rather see the specifications and use a spectrum analyzer during a demo.  Either these guys are violating the laws of physics, or are lying, or their technology is being wrongly reported, or some combination of the foregoing.

Sure, you can have almost unlimited bandwidth if you don't transmit very far. But, in that case, your phone won't have long reach, and the real estate costs of all those microtransmitters will eat you alive.

Perhaps this is reporting akin to the breathless reporting of those various "perpetual motion machines" that never pan out? "Unlimited bandwidth machine" probably fits into that category of "science."


Is there a video of this presentation?


J. S. Greenfield
J. S. Greenfield

@JMWJMWThey're not violating the laws of physics, they're not lying, and the reporting has not generally been inaccurate -- though it often does not provide meaningful detail, as in the above.  (And note, the "tiny, virtual cells" that the article references are not tiny cells around the system transmitter, but tiny zones around a client device, created by a device-specific signal generated using adaptive arrays and constructive interference.)

That's not to say it's assured that they can actually succeed in doing what they say they will do -- it involves a computational task of huge complexity, and their claims that they can track (fast) moving devices adequately remains to be demonstrated.

But fundamentally, they are simply using antenna arrays to generate a separate effective signal for each device.  In the end, it is a logical extension (or perhaps better described as a logical extreme) of adaptive array signal procession techniques, and in particular, MIMO.

You can google to find a number of good discussions of the technology.  You can find a particularly detailed technical discussion here:

ElizCrane moderator Re/code

@milan03 Yes! Full videos will be appearing on the site in the days following the conference.


@J. S. Greenfield In the real world of radio frequency engineering (in which I work and for which I am qualified) you create more unwanted interference than you do usable signal when you use phased arrays. It's called the d/u (desired to undesired) signal ratio (to whittle it down to a nutshell.)

Unlike you, I can only ask questions at this point, and without testing myself, I remain VERY skeptical.  I'm sorry, I'm a professional, I don't want to read what chimps discuss.  Before I or any other RF engineer listens to anybody in such matters, we determine the technical competence of the speaker.

Frankly, Steve Perlman and nobody else who I've ever seen comment on this, are unqualified to talk about their technology.


@JMWJMW First, we're very happy to speak to anyone about pCell technology (or have them come to the lab for demos). Antonio Forenza, PhD is an Artemis co-founder, and if you look up his published work in wireless, I'm sure you'll find his credentials more than qualified in terms of technical competence.

One of the goals of the pCell Code Conference demo was not only to show that pCell scales to any number of concurrent devices, but also make it very clear this demo would be impossible to do using conventional LTE (or phased array antennas for that matter). Our purpose is to advance the discussion to where pCell is recognized as something new, and then focus the discussion on the new things we can do with it.

We went to great lengths to make sure that this demo showed that all of the iPads were unmodified, had no new software and worked out-of-the-box with just a SIM card, and had WiFi or Bluetooth off. Also, the iPads were all located very close together, so if we had been using conventional LTE they all would be in the same cell, sharing the exact same spectrum.

The iPad Air is a 2-antenna LTE Rel. 9 device. We showed 20 of them streaming HD video because this went far beyond LTE's peak spectral efficiency to a 2-antenna device, proving the demo would be impossible to achieve, even in a lab, using any known wireless technique.

But, pCell does this demo effortlessly, and it can just as easily stream HD video to 50 iPads in the same spectrum. It's not magic, and it's not breaking any laws of physics. It's a new radio architecture (and lots of new math) exploiting a previously unexplored area of wireless: interfering space.

Also, regarding handling fast motion (Doppler), pCell works just fine. The video posted by Brian Hart below shows me shaking a Nexus 5 while HD video is buffering. We're well aware that other spatial techniques like MIMO and CoMP can't tolerate rapid motion. pCell is something different and it works great at high speed.

Again, we're happy to talk with you or meet with you. Please contact us, and you can talk to Dr. Forenza and/or you can come to our SF lab for a demo (and bring a spectrum analyzer if you'd like). Then, once you can set aside your skepticism, we'd be happy to talk about the cool stuff we can do with pCell.


Steve Perlman, Founder & CEO, Artemis Network, 


Get every new post delivered to your Inbox.

Join 300,896 other followers