One of the interesting things about technology is that, at least from the outside, it’s hard to tell what’s actually changing.

That’s not true on the inside, of course, where radical shifts are under way. The next big push in smart phones will be much greater intelligence. In the iPhone, Siri was just the tip of the iceberg. Future versions are likely to be much more interesting. Add to that the ability to more easily navigate between base stations and to intelligently manage power and performance throughout a battery charge and the changes under the covers will be enormous.

All of this requires far more compute power, much faster signal processing and routing, faster memory throughput, advances in coherency, and much more complex power management schemes. It will even require new architectures, which is why we’re headed down the path of 2.5D and 3D stacked die, pre-integrated subsystems and significant changes in software.

At the other end of the spectrum, these changes are equally prevalent in the data center, where improvements in efficiency and performance are significant enough to be line items on an operating expense budget. The emphasis on NVM Express and PCIe are just the start of what will prove to be a massive change in the way data is managed, stored and shared in the corporate enterprise, and it will require much more sophisticated electronics and software to make it all work.

Where electronics have been in more limited use, it’s far easier to spot what’s changing. In automobiles, for example, the shift from mechanical to electrical has opened up broad new opportunities for improving safety (even though some of this stuff is a distraction to the driver), adding convenience, and improving performance and fuel efficiency. What remains to be seen is ultimately how these inroads by electronics will change automobiles. Will they continue to be differentiated by the car vendors, or will they be differentiated by the makers of electronics the way a company like Apple has changed the music industry?

It’s easy to lose sight of the engineering feats in complex problem solving that enable these advances. It’s also easy for teams that accomplish these feats to lose sight of the bigger picture. All of these changes require an increasingly larger view of the system, big systems based on much more complex little systems, and all working together much more seamlessly than in the past. That means far more standards, more awareness of changes implemented on all levels, and more cooperation between groups that have never actually talked before.

So what will this ultimately mean for SoCs? Will it impede innovation as companies standardize on platforms and subsystems, or will it increase innovation as these platforms are called upon to drive more of the functionality in an ever-larger definition of the system? And who will assume the risk as complexity between groups continues to rise in an increasingly complex supply chain?

The changes on the outside may look like small changes to the consumers using technology, but underneath there is likely to be a lot of churn in all directions as technology continues to improve.

—Ed Sperling

Apple’s new iPad is an interesting device not so much because of what it offers to consumers—that’s certainly interesting in its own right—but because of how Apple built the device and why.

Apple has been scouring the market for seasoned semiconductor engineers of late. The process started two years ago when the company hired a team of former engineers from the late Digital Equipment Corp. who migrated first to PA-Semi, aka Palo Alto Semiconductor, and more recently to Apple. This is a rather odd trend, on the face of it, considering most companies have been outsourcing that part of their computer engineering. Apple had abandoned its own chip development efforts in its Mac line because Intel was beating the pants off everyone (which helps explain why Intel has run into so much scrutiny from regulatory agencies).

The folks at Apple haven’t gone entirely mad, however. The first details of its plan to develop its own chip began leaking out when a deal with a large Asian semiconductor company went sour. According to several sources, Apple’s initial iPhone deal called for this Asian company to provide the logic, memory and processor for the iPhone. But after a year of growing iPhone sales, this company also began shipping its own version of a smart phone that had some of the same feature sets as the iPhone.

Word on the street was Apple wasn’t happy. We have no idea what got broken or smashed in this fit of rage, but realize these folks are still reeling from the lawsuit with Microsoft that claimed theft of Apple’s graphical user interface in Windows 3.0. It doesn’t matter that Xerox invented this technology first. Apple brought it to market first, and it put Apple and Silicon Valley on the map. What took much longer was for Apple to establish itself as a brand that cut across business and consumer markets, which is where the iPhone came in.

Rather than risk a repeat performance with Microsoft, Apple began taking its chip development in-house again. It has been competing for engineers with well-known chipmakers in Silicon Valley, and it has been building in the kinds of things that it has been slammed for in the past, like lower power consumption and better utilization of cores.

But will competitive paranoia really drive a re-aggregation trend, or is Apple just so unusual that it will continue to carve its own path? These kinds of trends are best viewed in retrospect, and right now it’s still something happening in the future. The iPad isn’t on shelves yet and so far Apple is still using Intel chips in its Macs.

–Ed Sperling

At the beginning of this decade a writer for a powerful newspaper told me that, come hell or high water, she wasn’t giving up print—no matter how important online got to be. I had to think about that for awhile before answering, “It may not be your choice.”

That newspaper is now a shadow of what it once was, but the statement keeps reminding me of some of the brash claims being made by electronics companies today. No matter how brilliant an idea seems on paper, it can be a colossal flop for unanticipated reasons. And no matter how idiotic something may look to established players, you always have to take it seriously. Who would have thought 20 years ago that you could sell a cup of coffee for nearly $4?

Intel’s entry into all markets doesn’t mean it will succeed in those markets. Likewise, just because the market leaders it is challenging now have dominance in those markets, it doesn’t mean they’ll keep it. IBM invented the PC and lost the market to lower-priced competitors. Apple didn’t invent the MP3 player, but it now owns the market, while its share of the PC market remains small.

What’s getting interesting in the electronics world, though is that battles are no longer being fought by one company anymore. They’re being fought by ecosystems, and how those ecosystems fare against each other is unknown. To some extent it depends on how committed they are to each other. Is this like NATO or is it like the Allies in World War II?

Second, it is uncertain what will win out in technology. Will growth come at the low-end of the consumer market, or will there be enough growth at the high end to sustain more expensive development. To some extent, that depends on how fast developing markets mature and what their consumers are willing to buy, as well as how fast more mature markets recover from a very long and deep downturn.

And finally, it all depends on what sort of business context can be built around all of this. Apple’s genius in the MP3 world was iTunes. What that will be in netbooks, mobile Internet devices, set-top boxes and a variety of new form factors is unknown. And no matter how much money is thrown at solving the unknown, the results may still be unpredictable.

–Ed Sperling