G.fast Chip Startup Raises $10M

Israeli semiconductor startup Sckipio has gained $10 million in funding to speed the development and deployment of chips for new G.fast broadband modems.

The International Telecommunication Union (ITU) just completed work on the G.fast standard, which supports 1 Gbit/s speeds over copper phone lines at a range of up to 250 meters. So Sckipio is now coming out of stealth mode to promote its own G.fast solution. Michael Weissman, vice president of marketing for Sckipio, told Light Reading that the company believes the first G.fast modems will arrive in 2014.

The initial investors in Sckipio include Gemini Ventures, Genesis Partners , Amiti Capital, and Aviv Ventures. The company, which was founded in January 2012 by former executives from chipmaker CopperGate (which was sold to Sigma Designs Inc. (Nasdaq: SIGM)), currently has about 25 employees.

The G.fast standard has the potential to help telecom companies counter the top cable broadband speeds made possible by DOCSIS 3.0 and 3.1 technologies. G.fast requires carriers to drive fiber deeper into their networks than many have in the past, but the option is still far less expensive than running fiber all the way to the home. Telcos can pull fiber to a distribution point near a consumer residence and then use G.fast over the last stretch of copper phone lines to deliver gigabit speeds.

"The solution that a lot of people have been going towards is fiber to the home, and the challenge with fiber to the home is it is very expensive to deliver to the home," Weissman said, noting that it can "easily add up to well over a thousand dollars per subscriber."

In contrast, Sckipio says it can slash the cost to deploy "fiber-like" speeds by as much as 90 percent. That would presumably bring the cost down to $100 to $200 per subscriber.

Weissman said that G.fast improves network performance by using DSL-like vectoring to eliminate cross-talk. He said G.fast also offers flexibility to network operators, allowing them to choose the upstream and downstream speeds they want to deliver -- symmetrical or asymmetrical.

Because the standard was just completed and hasn't even been officially announced by the ITU yet, there are no G.fast products on the market today. But that hasn't stopped companies from trying out pre-standard versions of the technology. For example, BT Group plc (NYSE: BT; London: BTA) said recently that it's testing out Huawei Technologies Co. Ltd. 's pre-G.fast solution in its research and development center near Ipswich, UK.

Unlike Sckipio, Broadcom Corp. (Nasdaq: BRCM) has not been very forthcoming about its G.fast plans. But the chip giant will undoubtedly join the fray in the coming months. (See BT Trials Huawei's G.fast for FTTx.)

Meanwhile, Sckipio says its engineers contributed more than 20 percent of the work to the new G.fast standard, and Weissman noted that the company is solely focused on G.fast technology. While Sckipio won't share the names of its potential modem hardware partners, Weissman noted that his firm's team has worked with all of the major vendors in the past and continues to maintain those relationships.

G.fast is backed by top telecom operators globally, including AT&T Inc. (NYSE: T), Orange (NYSE: FTE), and Deutsche Telekom AG (NYSE: DT). In a press release, Oliver Johnson, CEO of the research firm Point Topic Ltd. , said, “G.fast is the best hope for telcos to affordably deliver 1Gbps to consumers. We expect G.fast to be broadly deployed quickly because telcos need an effective alternative to cable and G.fast appears to have everything these telcos need to succeed.”

— Mari Silbey, special to Light Reading Cable

albreznick 12/12/2013 | 2:34:41 PM
Re: G.fast fast enough? Then it sounds like G.fast may not really be an equal rival for DOCSIS 3.0 and 3.1, which or will be mass-maeket services. Do the telcos have anything else up their sleeves to boost broadband speeds significantly for the masses? 
brookseven 12/12/2013 | 12:20:11 AM
Re: G.fast fast enough? Alan,

The places that very high rate DSL have been reasonable topics are places like Singapore where population densities are very high.  From a US perspective, FTTC has been around for over 10 years in BellSouth and Las Vegas.

It just fits a few applications and has not become mainstream.  And at lower housing densities, FTTH is actually cheaper as an ONU per home is a really bad economic deal.


albreznick 12/11/2013 | 5:55:05 PM
Re: G.fast fast enough? So it's more of a niche market play then, not a mass market opportunity? Sounds like a lot of trouble to go thru for just a fraction of your market. 
albreznick 12/11/2013 | 5:53:09 PM
Re: G.fast fast enough? Yeah, and that's no walk in the park either. The cable guys don't have to worry about that with D3.1. No fiber-pulling required there.  
albreznick 12/11/2013 | 5:51:46 PM
Re: G.fast fast enough? Sounds like you don't buy the claim. What do you think might be the real top download speeds? And what kind of upstream speeds are we talking about? 
brookseven 12/11/2013 | 10:06:27 AM
Re: G.fast fast enough?  

Okay, well here is your problem.

So, now you have decided to do G.fast to save the last 750 feet of fiber construction.  In apartment buildings, maybe that is a good idea.  In single family home construction, the former Bellsouth's FTTC plant had a similar reach and would pass 8 - 12 homes per ONU.

So, now you have to do fiber construction, deploy lots of very small DSLAMs, and figure out how to power them (doing AC drops with batteries all that many locations would be insane).

So, I can see the possibilities in MDUs.  But for SFH - no frickin way.


msilbey 12/11/2013 | 9:50:38 AM
Re: G.fast fast enough? Alan- Easy enough to replace the modems, but telcos will also have to pull fiber deeper in a lot of cases.
msilbey 12/11/2013 | 9:49:53 AM
Re: G.fast fast enough? Right?! Seems so incredibly unlikely. But that's what they claim.
chuckj 12/11/2013 | 3:02:14 AM
Re: G.fast fast enough? 1G on twisted pair you must be kidding.
albreznick 12/10/2013 | 6:15:57 PM
G.fast fast enough? So the telcos aim to catch up with cable by deploying G.fast? That should be very interesting. I guess one big question is how easiy will it be to roll out G.fast? Will it be as easy as rolling out DOCSIS 3.1 promises to be for cable operators? And what will be the cost of deploying the new technology and equipment associated with it? This will make for a fascinating comparison with D3.1.     
Sign In