tip to my friend Peter for forwarding Testosterone Chips: The End of an Era – Business Week – Drew Lanza to me this week.
This is yet another article that addresses what I call the “The Curse of Moore’s Law” and the possible new directions that the semiconductor industry could take.
From a VC point of view he nails the point that the cost of developing chips is getting higher ( he notes $20-$30M) and exit prices have not changed in a decade.
From a Comm Semi vendor point of view I would add that the end markets have not grown much in that time as well. Some areas have died and some new areas have emerged, but in the end it is not great when it costs more to play in the same size sand-box.
He notes that the possible opportunity is in Internet appliances where “smaller, cheaper, and low power” are king. I agree with this and have noted it in a previous post where I add that there may also be a great opportunity in developing chips to for “serving up applications” for these appliances. There is more to it than Intel and AMD CPUs. To support my point refer to Eric Schmidt’s comment in his recent podcast at iInnovate — “Google is a capital intensive business”.
The only gotcha that I see with this new direction is that chip development costs for appliances are also very high. For example, if one uses $20-$30M per chip development (mentioned in article) then the iPhone is leveraging ~$100M in chip development. The next stage of iPhone will be integrating the current devices into 1 or 2 chips. That will not be cheap and I’m guessing that this ~$100M pales in comparison to development cost of all the f/w and s/w the surrounds iPhone.