Life After Smartphones

By Frank Ferro
Don’t let the title confuse you. Smartphones are not going away anytime soon. In fact this year’s smartphone shipments have exceeded feature phones for the first time, with a total of 216 million units in Q1, according to IDC, and the overall mobile phone market is expected to grow 4.3% in 2013. This volume represents an increase in smartphone sales of 42% from Q1 2012.

There is concern, however, that the high-end smartphone market is beginning to show signs of saturation, as previously optimistic sales expectations are being lowered. The market picture is not that simple, with high-end smartphones sales in the United States and developed countries still doing well, but the battle is now over brand dominance and not necessarily new growth (i.e. swapping back and forth between Android and iOS).

Market growth for smartphones is expected to come from developing countries with midrange or value-based smartphone products. This shift in product mix has the potential to disrupt the current market leaders in both the handsets and semiconductors industries. Concern from market leaders is evident as rumors surfaced about lower-cost iPhones, along with the recent announcement by Qualcomm of upgrades to Snapdragon 200, which is being targeted for entry-level phones in China and emerging regions. This tactic by Qualcomm is clearly in response to growing competition from both Asian and U.S. chip manufacturers targeting low-end smartphones.

Given that smartphones and tablets have been fueling the growth and innovation for the semiconductor industry, it will be interesting to see how the shift in focus to low-end smartphones will affect SoC development. As a consumer, I believe this is a good thing for the SoC industry, because up to this point semiconductor companies have been focused on the ‘race to performance and features.’ This race, without a doubt, has driven much of the SoC innovation, including multi-core CPUs, higher frequencies (>2GHz), more powerful GPUs, along with many new features. These features and functions were necessary as applications processor companies were fighting for market share. They were also necessary because consumers could not get enough new features in their smartphones.

Today, high-end smartphones are maturing to the point where differentiating one product from another is difficult. Adding additional CPU cores or speeding up the clock may not provide a discernible benefit for the user. In fact, it may have the opposite effect by increasing cost and power consumption. So the battle for low-end smartphone market share will require SoC manufacturers to focus on cost, power, product efficiency and time-to-market.

To reduce the overall development cost a better SoC design methodology is needed. At the recent Design Automation Conference (DAC), Nvidia chief scientist, Bill Dally had some very interesting comments on this topic, saying that it is “just scandalous” that SoC design takes as long as it does. He also noted that it may cost $50 million to take a “relatively simple SoC” to first prototype. A key element of the solution he proposed was a better IP ecosystem with hard IP blocks that conform to a standard network-on-chip (NoC) interface, so that IP can be easily connected together. Although I am not completely in agreement on the practicality of creating all these hard IP blocks, I do believe that complete IP subsystems, connected via an network-on-chip, will go a long way toward reducing development time.

Power management also will need to be a key focus if SoCs vendors want to be successful in the low-end smartphone market. Many design teams for the current application processors on the market had the best of intentions to aggressively manage power, but schedule pressure and the state of power management technology often left these solutions falling far short of their intended power saving goals. For the next generation of SoCs, power management has to start at the architecture level, including more power domains with finer grain control of these domains. Hardware control is needed for much faster on/off times, thus minimizing the need for CPU and software intervention.

Clearly smartphones will be the ‘hub’ of our activity for a long time, and like most consumer products, growth in the early phases of deployment means price reduction through improved cost. Advanced SoC development will certainly continue as requirements for the tablet market will lead high-end mobile SoCs. For the smartphone market, however, expect to see less influence on SoC design from the high-end segment, which is driven by performance, and more influence from the low-end, which is driven by cost and power consumption. Life after smartphones? That’s a story no one can write yet.

—Frank Ferro is director of product marketing at Sonics.