This week’s issue of Tech Watch is filled with stories about new products, many of them semiconductors, SOCs to be more specific.
SOCs are tricky devices to design, and get right. Almost no one got the first one or two right. Temperature, performance, battery consumption, and of course price are just a few of the macro factors that can undermine such a product. Usually companies have to sneak up on the SOC, designing a smaller device first with limited functionality. Then one day taking the plunge and hoping they got right the hundreds of items that are interlinked.
Also there is an awesome supply chain, certifications, and in some cases government qualifications.
An SOC takes about three years to de-velop, depending on where in the process you start counting. That means the de-signers and their suppliers have to try and figure out what will be needed at least five years out. In a fast moving industry like CE, PC, or mobile that seems close to impossible—and yet they do it.
All SOCs have some non-native IP in them. SOCs just have too many functions and features for any one company to be able to design the best of class for all the items – and do it in one’s lifetime.
Add to that the IP suppliers are on their development cycle and the probability of getting all the moving parts synched is almost incalculable. So how then does anyone get such a complicated thing as a SOC ever done? Aside from the complexity, building an SoC is labor and computationally intensive. A $20 SOC probably takes 1,000 man-years and a trillion gazillion computer cycles.
That of course takes money, and it ab-solutely demands a coherent, consistent, and dedicated team willing to stick it out for the long haul.
How in the face of the worst world-wide recession in modern times do you work that balancing act if you’re the CEO of an SOC company? Well for one thing you take a page out of crusty old Craig Barrett’s book— “You can’t save your way out of a recession—you have to invest your way out.”
And when he was running things he did, pouring billions into fab construction that put Intel in the most powerful position to win business when the recession was over.
Seems quite a few companies, certainly all the ones written about in this issue, got the memo and did just that.
Now maybe it’s coincidental that all the SOC announcements in the past three or four weeks happened around CES and soon GSMA, or maybe it’s synched to economic meltdown of the second half of 2009. But if, when the recession’s early warning signals were being flashed, several smart CEOs and their executive teams said—batten down the hatches boys we’re heading into rough waters, and then invested their precious resources for when they’d come out of the storm, you’d probably gets the results we have today.
And that’s what I think happened.
Now it’s always easy to pat yourself on the back after the battle has been won, and we all know the victors get to write the history books. But I can’t find another logical explanation for the rush of great new products.
2010 will probably go down as the most boring year in the decade. It was supposed to have growth approaching 20% and ended up with a mediocre 3% gain for Q4 year to year.
Good-bye 2010, glad to see you go. But 2011 – we’re going to have a new iPhone, a new Nintendo 3D game Boy, a new slew of tablets from major bands plus a new iPad, a new Sony PSP, new Internet TV, and some with 3D. Our cars are going to get smarter, our electronic books brighter, our phones lighter, and recharge cycles fewer.
It’s a renaissance of technology that will be met by pent-up demand and a (fi-nally) growing economy.
At CES I saw pico projectors that could light up an entire room, power-line Ethernet adaptors that could toss an honest 100 Mbits/second over the mains, and Internet radios that not only got Pandora but looked so cool I wanted to take one home.
The SOCs are going into everything from POS and kiosk machines to anything and everything that uses or creates media and runs on batteries. And in a year or so they will also go into a new class of PCs tentatively known today as Smartbooks.
Traditional x86 have become a bit of an SOC themselves as they take on the mantel of HPU and EPG.
Naturally the semiconductor fab boys are going to take credit for all this, point-ing out how none of it would be possible if they hadn’t made transistors so tiny you can’t actually see them and just have to assume they’re there.
And speaking of seeing things the new crop of 3D photo sensors, and photo sensor technology for everything from game gadgets to road lane detectors in side mir-rors, tele-presence and intrusion detection will forever change how we interact with the world, and our neighbors.
It was Nietzsche who said whatever doesn’t kill you makes you stronger. The recession of 2009 followed by the dol-drums of 2010 have certainly made the electronics/computer/CE stronger, and better, and a hellofa lot more exciting.