Blog Post

Could this battery-powered, Wi-Fi camera show chip startups how to find a market?

Stay on Top of Enterprise Technology Trends

Get updates impacting your industry from our GigaOm Research Community
Join the Community!

Immedia Semiconductor, the maker of several image processing chips used in devices today, has launched a Kickstarter campaign for what could be a sweet new device for the smart home. The campaign is for a Wi-Fi camera and corresponding hub that costs between $69 (one camera) and $269 (five cameras) and is powered by two CR123 Lithium batteries. This is a pretty slick device, but what’s slicker is the silicon making it possible.

Immedia was formed in 2009, by silicon designers who were familiar with developing image compression techniques and video codecs. The proprietary chips they have built do three things really well, so the cameras can go a full year on a set of batteries.

This is a tough problem

The first has to do with image processing. Video files are huge, because each frame has tons of information. But not all of that information needs to be kept. Lower resolution files dump some of the clarity, while other compression techniques might only track the differences in each frame and dumping all that stuff that stays the same. So when a chip processes images it generally sends frames to memory and calls them up as needed for comparisons and processing. The Immedia chip doesn’t do that. The lack of DRAM means it uses less power. Instead, its algorithms compress in real time using a proprietary silicon design and software.


The second thing any wireless, Wi-Fi camera has to deal with is the radio. Wi-Fi is a power-intensive protocol, able to send huge data files quickly because it has a lot of capacity. Immedia’s has integrated its image processing chip with the lowest power Wi-Fi chips around and has also tweaked its software so the Wi-Fi radio only sends the video files when the video is done recording. This doesn’t stream data — instead it sends it in a quick, powerful burst and shuts back down to save power.

And that leads to the third challenge Immedia faced — getting the camera awake quickly and able to transmit its video as soon as possible. Blink is a security camera, so if a burglar pops into your domicile and triggers the motion sensor the camera needs to come awake and start recording quickly in order to catch an image of the intruder. That’s actually tough to do — waking up a chip and recording video within a second or less.

A new path for silicon startups

Yet Don Shulsinger, VP Marketing/Sales of Immedia, said that’s what the Blink camera does. The Kickstarter campaign will deliver the camera in March 2015, so we’ll have to wait and see if this works. However, since Shulsinger told me that the reason the company is turning to Kickstarter is to show off the new tech that its silicon makes possible, I have to think they’d deliver on their promises.

Shulsinger explained in an interview that crowdfunding lets Immedia bypass the far-longer vetting and design cycles of large consumer electronics giants who are the primary customers for most silicon. Generally a company designs a chip and shows it to OEMs, who then show it off to their end customers or otherwise devise tests to determine if the market wants a device with the new capabilities, and it can be 18 months or more before a prospect becomes a significant order.

Given how little money venture firms want to put into semiconductor startups, this long lead time and caution can frustrate a technologist who feels like her silicon can change the world — or at least the current device landscape. Thus, the ease of building new hardware and markets like Kickstarter have led chip firms to go direct to the end buyer with novel technology in hopes of getting the large original equipment manufacturers on board.

The Adapteva Epiphany 16-core chip.
The Adapteva Epiphany 16-core chip.

Immedia isn’t the first chip firm who has turned to a crowdfunding site to find a market. In 2012 Adapteva, a company that was building a massively multicore chip that could offer supercomputer-like performance on a tablet or cell phone, started a crowdfunding campaign after its venture backers backed off. It ended up meeting its crowdfunding goals and attracting more venture capital once it seeded the market with its chips.

Immedia has raised $20 million and already has plenty of OEMs buying its chips, but it thinks that with Blink it can showcase its silicon in ways that current customers aren’t yet trying. I’m pretty sure if the camera works, there’s a significant market for a small, battery-powered Wi-Fi camera.

4 Responses to “Could this battery-powered, Wi-Fi camera show chip startups how to find a market?”

  1. Blair MacGregor

    “Given how little money venture firms want to put into semiconductor startups, this long lead time and caution can frustrate a technologist who feels like her silicon can change the world”

    I try to stay away from arguments that begin with “VCs ought to be putting money into ‘X’ instead of consumer web, etc.” But this is kind of alarming. I’m not sure if the 18 month timeline is typical for chip companies to see revenue but plenty of consumer facing apps that have gotten funded from angels/VCs have gone a lot longer than that without seeing a profit.

    • But chip companies need a lot more capital up front compared with a web service. For example, Calxeda, which flamed out for a variety of reasons, raised over $100 million before shutting down. That’s a painful loss for an investor.

    • Ken Feldman

      It’s not 18 months. If I am reading this article correctly it was 5 years and $20M, and then they decided to bring a consumer product out, which will take another 18 months. In the mean time, they now have a reference design and some market success they can point at. It will help them sell their chip to OEMs.

      That being said, I firmly believe that Silicon Valley needs to get back into SILICON. Without the underlying chips to make things happen, the underlying infrastructure, what’s all that social / consumer web crap going to run on? Sadly chip companies appear to take too long and have too much “risk” and not enough multiples for VCs and their Limited Partners. A very shortsighted view.

      • Gary Dare (@GaryDare)

        Agreed on the last paragraph, Ken! The reference design by Immedia with its capabilities would not be possible under the common practice in embedded systems of just trying to program one stock board or another, never mind the next step of trying to figure out how to translate it into the small silicon form factor needed in the end product. What is needed is an optimal partitioning of hardware and software for the system, where the hardware is not necessarily stock (i.e., a new ASIC or at least a FPGA).