LTE Handset in 2010 a Longshot Despite 4G iPhone Hope - Telephony

LTE Handset in 2010 a Longshot Despite 4G iPhone Hope: Most chip vendors won’t have the silicon ready to support an LTE phone next year, and those that do just don’t see the business case
Telephony
By Kevin Fitchard

May 12, 2009

Speculation is crescendoing that Verizon Wireless and Apple are not only on the verge of locking down a deal for an iPhone, but that the two plan to launch a 4G device on Verizon's new long-term evolution (LTE) network next year. But a poll of chipset makers supplying the crucial LTE silicon reveals just what an unlikely proposition that is.

A high-end concept phone with an LTE chip emerging next year isn't entirely out of the realm of possibility, but the chances of a mass-market LTE handset such as a 4G iPhone are practically nil. The release schedules of the wireless silicon providers just don't support it. Chipset makers have announced their first-generation LTE platforms, many of which will sample this year and begin shipping in volumes early next year, but almost all of those platforms are purpose-built for laptop computing, designed to be embedded in USB dongles, PC cards and eventually PCs themselves.

ST-Ericsson, the new semiconductor joint venture between Ericsson and ST Microelectronics, has already begun sampling its first LTE chipset, the M700, but Chief Technology Officer Jörgen Lantto said the platform has been specifically developed and cost-optimized for USB dongles and other plug-in modems. Making a handset isn't the simple matter of slipping the M700 into a phone form factor, Lantto said. A handset would require an integrated application processor, a smaller chip form factor and lower power consumption. Perhaps most importantly, it would need an integrated 3G radio so the device can fall back on the much more pervasive high-speed packet access (HSPA) or CDMA EV-DO network when LTE coverage isn't available.

While laptop card makers also will need to fall back on 3G, they don't necessarily need a hardware-integrated chipset, Lantto said. Embedding separate 3G and LTE chips into a USB dongle would be far cheaper than provisioning an integrated chipset while LTE volumes are miniscule, Lantto said. Only when device volumes reach high levels—around 100 million units—does an integrated solution for an external modem make economic sense.

That said, Lantto isn't ruling out entirely the possibility of a vendor making a handset with M700, or any of its counterparts from ST-Ericsson's competitors. An enterprising device maker could embed separate 3G and LTE chips as well as application processor, integrating them all through software. In fact, Lantto said ST-Ericsson has received inquiries from customers about doing just that. But any device using piecemeal components would likely be extremely expensive, highly inefficient and power-leaching design.

"There will be an interest in developing super-high-end devices that have compelling super-high-end data rates, but those won't be high-volume sellers," Lantto said. "They will be used more to show off."

BUILDING THE LTE iPHONE
If Verizon and Apple were to launch an LTE iPhone in 2010, they likely would be dependent on one company to get them the necessary silicon: Qualcomm. As the world's primary supplier of CDMA chips, Qualcomm would need to supply the integrated CDMA 1X, EV-DO and LTE chipsets that would allow the iPhone to make calls on the Verizon network and access both the 3G and 4G data networks. Though Qualcomm isn't commenting on any possible deal with Apple or Verizon, it doesn't see much likelihood of any Qualcomm-powered LTE handset emerging in 2010.

Qualcomm is making only dual-mode platforms—integrating LTE and either HSPA or EV-DO in the same chipset—so its products will hit the market slightly later than those competitors making single-mode LTE chips. Right now, Qualcomm is scheduled to sample its first data card chipsets, the Mobile Data Modem 9xxx family, in the current quarter, but its first handset chipset, the Mobile Station Modem 8960, won't get into vendors' hands until mid-2010. Given the time lag between receiving a chip and designing, building and testing the device, launching an LTE handset using a Qualcomm chipset would be virtually impossible.

"We would expect multi-mode LTE data cards in the market in later 2010, with multi-mode LTE handsets in 2011," Qualcomm CDMA Technologies senior director of product management Peter Carson said in an e-mail interview. Qualcomm claims it will have the first multi-mode LTE chipsets in the market, which puts it on target to deliver the first LTE phones. While a handset maker could use a single-mode platform to make a device, Carson said there would be little point. "Our multi-mode solutions will support LTE as a standalone mode, however, 3G multi-mode will be necessary to ensure a smooth LTE introduction and a seamless mobile broadband user experience."

Another option for Apple would be to eschew Qualcomm for LTE and provision separate radio and processing components, which, while not integrated on a single chip, could each be optimized for the handset. Specialty 4G chip maker Altair Semiconductor is proposing these sorts of implementations with its product line, which consists of standalone radio baseband and radio frequency transceivers. Altair is focusing primarily on the embedded device market and has designed a small low-power chip that can be used in both handsets and data cards, said Eran Eshed, vice president of marketing and business development for Altair.

That baseband chipset will ship to customers in the fourth quarter, and a handset maker like Apple could feasibly have it embedded in a handset along with a CDMA chip, RF components and a standalone baseband processor as early as the first quarter of 2010. But Eshed finds that scenario unlikely for any vendor. New wireless technologies follow an orderly progression from modems to handsets as networks deploy and the potential for volumes increase, Eshed said. No single vendor, no matter how aggressive, is interested in upending that system.

"In the first quarter of next year, Apple or anyone else will be able to get a chipset small enough and low-power enough to go in any device," Eshed said. "The challenge won't be availability. LTE will take time to mature. Why would Apple be the first to take the beating?"

DO WE REALLY NEED LTE 'PHONES'?
Speaking at Ericsson's Capital Markets Day analyst event in Boston last week, Ericsson Chief Technology Officer Håkan Eriksson said he's puzzled by the intensifying interest in LTE phones. "We often get the question 'When will phones be ready for LTE?'" he said. "Phones are not the target for LTE."

Phones are primarily voice-centric devices, which are supported quite efficiently on today's 2G and 3G networks, Eriksson said. While smartphones do have powerful data capabilities, their small form factor makes 3G networks perfectly adequate to support most mobile applications. LTE, Eriksson said, is much better suited for devices that traditionally run on the wireline broadband network: laptops, MIDs, multimedia streaming devices, etc.

Bandwidth, however, isn't the only justification for LTE. 4G technologies don't just deliver data faster; they deliver it far more efficiently than 3G networks. Eventually operators will want to use LTE for any data application because operationally it's much cheaper to deliver a kilobyte of data over an LTE network than a 3G one, said ST-Ericsson's Lantto. Before that can happen, the LTE ecosystem has to reach a balance, he said. The cost of putting an LTE phone into the hands of every subscriber would far outweigh the cost savings of shipping all data through a 4G network. Once that balance shifts, the market will shift to LTE handsets, Lantto said.

Some operators, though, might see the advantage of shifting to all-LTE sooner rather than later. As LTE chipset prices drop, they'll become cheaper than EV-DO chipsets, Lantto said. When that happens, a CDMA operator with an extensive LTE footprint could procure LTE phones cheaper than it could EV-DO phones, allowing it take advantage of both equipment and operational savings. Verizon Wireless is certainly entertaining the possibility: VZW chief technology officer Tony Melone said last week Verizon's 3G network may not make it another 10 years once LTE is launched.

"I could imagine a CDMA operator really aggressively moving its smartphones toward LTE because they can take advantage of LTE's global economies of scale," Lantto said.

Operators, however, can't simply wait for scale to magically appear before they start deploying handsets, said Marty Smuin, Americas president for wireless business consultancy Aircom. Data cards and embedded consumer electronics will be sizable niche markets, but ultimately handsets are the only devices that can drive mass adoption of a new wireless technology.

"The biggest roadblock to the commercial adoption of LTE is not building out the radio network or core, although they are major factors, but rather the availability of LTE-enabled handsets," Smuin said.  "Until handsets support LTE, are the same size or smaller than the current generation of handsets, offer a similar level of battery consumption and are price-protected, mass market adoption of LTE will not occur."