That’s what AMD said, and it’s a big error.
Let me explain why this is so bad. The problem is not that AMD doesn’t have a real netbook processor now, or in 2009. It’s that they’re not going to have a real one by 2011. They’re not even going to try. They just forfeited the next big game in computing.
I’ll put it this way, this is like a major auto manufacturer saying today, “We’re going to ignore hybrids and electrics and all that other weird alternative stuff. We’re just going to stick with the good old internal combusion engine.”
That might not seem to be such a bad idea at the moment with gas prices down, but between environmental concern compounded by the expected explosion in car ownership in emerging economies the next few decades (the number of cars on the road in the world is expected to go from 600 million in 2005 to 2.9 billion in 2050), the typical car a couple decades from now is going to be light, green and cheap. If you’re a car manufacturer, you can’t effectively respond to the change just by adding batteries to your cars, you have to rethink and redo the car, the same old won’t cut it anymore.
The same thing is going to happen in computing, for largely the same reasons; the only real difference will be is that it is going to happen much more quickly than in the auto industry.
Right now, the race is on to build a PC platform that can live in a phone and give the average person “good enough” performance. Maybe it ends up in a phone, maybe it ends up in something a bit bigger with a few more bells and whistle. Where it ends up doesn’t matter, what matters is that the tinyPC platform will power it, and it will be cheaper than other alternatives.
tinyPCs are a disruptive technology, and they will seriously disrupt once the next decade rolls around. And AMD is saying, “What, me worry?”
AMD says its Conesus processor will be meant just for high-powered ultra-portables. What they’re really saying is, “We can’t afford to build a real tinyCPU, so we’ll just cut down our regular CPU.” This is a tactic, not a strategy. It’s bound to fail, because outside of a few luxury notebooks, there isn’t going to be an ultraportable market in a few years. It’s going to get eaten alive by the tinyPCs. Then what is AMD going to do?
You’ll probably say, “AMD can’t afford to design and build a tinyCPU architecture.” Of course they can’t, but that doesn’t make it any less of an error. If you can’t afford the future, what kind of future are you going to end up with? Maybe “error” isn’t exactly the right word to describe this, “failure” may be a better term.
This is how lack of money gets AMD in hot water all the time. Unless they’re especially dense and trapped into thinking strictly inside the PC box, they have to see this coming and know where it’s going. But they don’t have the money to fully fund what they need to do to keep up with Intel in their current markets.
So what does AMD do? What they do is piss on what looks to be the biggest fires at the moment, and let the others go until they become the biggest fire. Then they piss on that one for a while, and so on and so forth. Since there’s only a limited number of pissers who can only piss so much and who know they have a backlog of fires to leak on, there’s always the pressure to gamble, take shortcuts, just pee enough to get the current fire under control and move on to the next blaze rather than put it out. This often fails, and the whizzers have to come back to do the job right the second time. I’m sorry, but that’s how the place works, and I’ve seen enough comments from AMD employees to back up my observations.
This doesn’t work too well in established markets, so you can imagine what kind of priority a task gets in that environment when the task is to build a brand new CPU that has to stand most CPU norms on their head for a market that doesn’t exist yet. Look, AMD hasn’t even built a good notebook CPU yet; it wasn’t all that long ago that AMD’s approach to “making” a notebook chip was to use the desktop chips that came out of the fab which needed the least power and make them notebook chips. They’ve gotten better than that, but the mindset of the culture is still definitely desktop, so a tinyCPU is not just a fiscal, but also a mindset challenge.
What will happen? It is sadly predictable. AMD will of course change its mind once tinyCPUs really start hitting their stride (I’d say sometime in 2010), panic, and those doing Bulldozer and Fusion now will be told to design a real tinyCPU yesterday, so they’ll come out with a rushed, half-assed product in 2011 or 2012.
Does that mean Intel will monopolize this new market, too? No, they’ll have someone else to fight: ARM. Not only are they already in the tinyCPU market, they’ve been in it a whole lot longer than Intel. ARM deserves a full future article on its own, but for now, let’s just say they and their manufacturing friends will prove a lot fiercer competition than AMD. Not only will they fight to defend their phone turf against Intel, they’re also planning to plant a few feet on Intel ground, too.
So there will be competition in tinyPCdom; AMD just won’t be part of it.