Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ The Five Percent Solution ]   |  The Eyes Have It  |   [ What Goes Around Comes Around ] >>

Weekly Column

The Eyes Have It: Last week's Apple mystery is all about video chips (I think)

Status: [OPEN] comments (11) | add a comment
By Robert X. Cringely

The Pulpit Poll

Bob seems to be out on a limb on this one. Is he right about Apple's upcoming H.264 hardware support?


Skip this one and see results

So little real information leaks out of Apple these days that we tech pundits tend to jump on any crumb we can get and munch it to death. That's certainly the case with this week's story about Apple possibly dumping Intel chipsets for the new MacBooks expected to be announced in September. What's funny to me is that the answer to what's REALLY happening has been in front of us all for more than a year.

Here's how this mess of a story got started. On Monday, July 21st Apple Chief Financial Officer Peter Oppenheimer dropped a bomb on those listening to Apple's quarterly conference call on earnings for Wall Street analysts. He said that gross margins for the coming quarter, and possibly beyond, would be lower for three reasons: 1) a back-to-school special; 2) a one-time charge related to a contract manufacturer, and; 3) "a future product transition that I can't discuss with you today."

That's it. That's all he said. And from that sprang a zillion stories about what that future product transition could possibly be. It settled eventually on the idea that Apple might be abandoning Intel processors, later downgraded to Intel chipsets, for the new MacBooks and beyond. Of course Apple's recent purchase of PA Semi got folded in as pundits wondered if Apple was going back to PowerPCs after all.

I know how these stories develop, having written more than a few of them myself over the last 20 years. You start with one fact, get the usual suspects to speculate on what that fact could mean, throw those speculations into print, then look for an official denial of the parts that are wrong. Once that denial comes through we rinse and repeat with the goal of eventually converging on something close to the truth. It's not a very elegant way to do journalism, but that's the way it happens in the tech trades, which now include everything from blogs to the New York Times.

But what's REALLY happening here? Stepping back from the carnage we can see that Apple has a "product transition" coming up that will hurt margins in the near term, but Oppenheimer also said it was dramatic and would definitely HELP margins in the long term. That's all we really have to work with from Apple, but it is really quite a bit if you parse the data carefully.

First is the product transition, which quite specifically DOESN'T mean a new product. If Apple was announcing something completely new as they did last year with the iPhone and Apple TV, then Oppenheimer would have referred to it as a new product. As CFO he has fiduciary and legal responsibilities that could land the guy in hot water with the SEC, so language on these calls is important and never by chance.

Second is the margin hit that will go away, which smart readers right away saw as a change of chips, because they start expensive and become very cheap over time. By making an aggressive semiconductor move Apple would be trading profit margins for technical market advantage knowing that in a few months the new chip process would come down and margins could return to normal. THAT's why all the smart money went immediately to speculating about Intel, then backed off somewhat as official denials began filtering through back channels from Cupertino and Santa Clara.

As of today people are just left scratching their heads. Apple isn't changing CPU families and evidently they also aren't dumping Intel chipsets for those of Nvidia. But SOMETHING is happening because Peter Oppenheimer gets no pleasure predicting lower margins that he knew would drive down Apple's share price, if only temporarily.

So now the pundits are wasting even more packets wondering what Apple is planning, at the same time generally admitting that they (the pundits) don't really have a clue.

Regular readers of this column may well have an idea what's up, because I wrote about it more than a year ago. Before I drop my own bomb, though, I should say that I have no new information and what I am about to predict is based solely on my earlier reporting. Here's what I THINK Apple is about to do.

I reported more than a year ago and repeated in this year's predictions that Apple would be adding H.264 hardware support to its entire line of computers. The chip they are adding comes from NTT in Japan and was developed in cooperation with Japanese broadcaster NHK. The chips began sampling a year ago and should now be available in volume, though Apple may be paying as much as $50 each for early production.

This would be a major blow to gross margins because, unlike all the speculation covered above, this wouldn't be a matter of replacing one chip with another but of adding a new chip to the mix. That'll be an extra $50, thank you, with no savings from eliminating other parts.

The fun part is figuring how this all fits into Apple's strategy as not just a maker of computers but also as a seller and distributor of entertainment content.

The NTT chip is not just an H.264 decoder, it encodes, too, which is what makes it so special. The last I heard NHK was claiming the chip could compress a 1080p video and audio stream into four megabits per second, down from the 20 megabits normally required. If we assume Apple will apply the same kind of wink-wink, nudge-nudge transcoding to 1080p that they've already applied to 720p in the Apple TV, then it is within reason to expect they'll claim to distribute 1080p over iTunes in two megabits per second.

As the dominant technology platform in television and movies today, it makes good sense for Apple to put this H.264 hardware capability into the Mac Pro line, and maybe even into the MacBook Pros for professional use, but darned if I can immediately see why such powerful and expensive compression capability is required in a MacBook, iMac, or Mac Mini, yet I was told long ago that the chips would be applied "across the entire line." We'll see.

Of course this is all about taking command of the 1080p video market. Apple's strategy with iTunes will continue to evolve, but for the moment having a unique real-time 1080p capability will suck a lot of early adopters back into the Apple stores and give Apple's emerging content competitors like Netflix something new to worry about.

When Apple marketers sit down to talk about the competition they discuss Netflix and MAYBE TiVo, but that's it. Hulu is something iTunes could emulate overnight so it doesn't matter and none of the other video distribution channels are seen as having the potential to achieve critical mass.

What really excites me as a content creator is the amazing potential of real-time HD. Video and games are by far the greatest consumers of cycles on modern PCs. By embracing a dedicated H.264 chip THAT IT MAY WELL HAVE EXCLUSIVELY FOR A YEAR OR MORE, Apple is taking an out-of-the-box approach that will frustrate its competitors in both software and hardware. While the H.264 chips are expensive, they'll enable Apple to save money elsewhere by having slower computers that run faster video. Though it is doubtful that many will use it, you can be sure Apple will trumpet the ability to support 720p video in iChat.

So why am I the only one writing this?

It's because I could be wrong, of course. But I don't think so. I'll just have to take a chance and see.

Comments from the Tribe

Status: [OPEN] read all comments (11) | add a comment

Perhaps the product transition is simply Steve Jobes stepping down for health reasons.

dale sykora | Aug 01, 2008 | 6:19PM

Why not just put a high(er) end GPGPU into all Mac? Most if not all GPGPU can decode H.264 already. I'm sure with a bit of software it can do a pretty good job encoding too. With OpneCL coming next year, a hundred GPU can be really useful for a lot of thing. OTOH, a hardware encode/decoder can only do 2 things...


pH7 | Aug 01, 2008 | 6:25PM
Peter | Oct 20, 2008 | 8:27AM
[OPEN] read all comments (11)


Ground rules for posting comments...

  1. No profanity or personal attacks, please.
  2. Please restrict your comments to the subject of the column and directly relevant topics.
  3. Be more funny.
  4. Those who violate these ground rules will have their comments deposted (which is a bit like being deported, only you don't have to leave the country).



NOTE: Your email address is for internal purposes only and will not be published, shared or sold to other entities.

url (optional):

Comment (br and p tags are not necessary for line breaks)