This year’s Computex also underscored the changing dichotomy of notebooks and PCs, according to an industry observer.
notebook sales are clearly being affected by the strong growth of
tablets, “the volumes of desktop PC have been sold on the same basic
trajectory that they were on before the tablet phenomenon came into the
limelight,” the source said. That’s not to say that desktops are growing
in overall volume. But at least “The decline in annual unit sales is
consistent with the forecasts,” which were in place before tablets
disrupted the notebook segment, he added.
Some observers appear
to believe that because desktops are less affected by tablets, several
in the ODM community in Taiwan are now finding it more appealing to work
on desktop enhancements in order to maintain or grow their market
shares, despite declines in total unit counts for desktops. One observer
described it as “Sort of like what happened before when PCs started to
4. Logic and memory trading places
trend at Computex was the new phenomenon of logic and memory switching
places, according to Deutsche Bank’s research note.
SoCs with memories supplied to the smartphone space, the Deutsche Bank
research analysts wrote, “Memory vendors are likely to generate better
return on invested capital (ROIC) from low-end devices than SoC
vendors.” It’s because unit growth is not offset by an increase in
supply,” the research note said.
More specifically, in DRAM,
vendors continue to prioritize mobile and specialty DRAMs, which have
higher die sizes on average but this added cost is more than offset by
higher blended average selling price and gross margin.
to larger and higher resolutions displays is also helping DRAM
suppliers, according to Deutsche Bank, “given memory bandwidth is a key
constraint on display and resolution scaling.” As low-cost smartphones
and tablets continue to take over the market, DRAM suppliers will have
more to gain, while traditional smartphone SoC suppliers are likely to
suffer in the intensifying price battle.
5. ARM in servers? No so much
it’s time for investors on ARM SoC companies to face up to reality.
Despite some expectations that ARM SoCs will start moving into high-end
average selling price markets such as notebooks and servers, Deutsche
Bank flatly pointed out, “We found that the facts support the opposite
view.” The research note said, “We note very little interest in Windows
RT devices and with the exception of Marvell’s storage win at Facebook
we saw little traction for ARM Server SoCs.”
I am not so sure. They seem to be still years away before gunning for IPO. Actually, I am more concerned about how sustainable their business is. More M&As, licensing or acquisition of technologies must happen before Rockchip an Allwinner can start truly leading the market.
MTK has already emerged as a formidable player in the smartphone apps processor market. The question is whether the growing white box market will continue to be left alone for all the China fabless companies to take. Or will we ever see leading chip companies n the west to move in that space?
Junko - Very informative, thanks. A few observations:
- Not to take anything away from them, but I wonder how long the China-based semiconductor companies such as Rockchip, Allwinner...etc will last. I wish them well, but they seem like one-trick pony, shanzai semiconductor companies, here now with a big splash, causing chaos with their price-crashing strategy, but gone tomorrow. Remember RDA? Where are they now?
- It is possible an MTK or an M-star will emerge from these, potentially threatening the likes of Qualcomm/Broadcom etc., but if SOC prices are crashing (relative to memory prices, as your #4 point says), then maybe there's more room to play in the memory space. Interesting, because the innovation cycle in memory is longer, whereas in the SOC it is shorter.
David Patterson, known for his pioneering research that led to RAID, clusters and more, is part of a team at UC Berkeley that recently made its RISC-V processor architecture an open source hardware offering. We talk with Patterson and one of his colleagues behind the effort about the opportunities they see, what new kinds of designs they hope to enable and what it means for today’s commercial processor giants such as Intel, ARM and Imagination Technologies.