32 billion dollars for Arm

Murph

Well-Known Member

Reaction score: 181
Messages: 297

As a former small shareholder of Acorn, who made some money off the original ARM bubble, before it was spun off, I'm saddened to see the company essentially lose its independence. I struggle to see how this deal actually provides anything useful to the future of the technology, as the company is already quite successful and innovative on its own, appearing to be both quite financially healthy and with sufficient resources to keep pushing forward at an acceptable rate.

As far as royalties going up is concerned, my guess would be that the existing major licensees likely have their royalties mostly locked in somewhere close to current rates with only modest increases over time. Aside from that, if SoftBank want to achieve their stated volume goal, a cash grab on royalties would be very much the wrong direction. Right now, there's considerable potential for 64-bit ARM to make a major impact into Internet servers (particularly in some areas of the cloud space, where there is much less architecture lock-in to x86), and a significant part of that potential comes from the low cost of the CPUs. To me, the sensible route is to keep the cost model down at its current quite modest level, so that it is a very attractive proposition to produce and deploy those chips in the huge volumes talked about.

It does cost relatively big money to keep rolling out new updates to the architecture, but ARM Ltd has a per-unit cost of zero on the actual chips being sold. It is very much in their interests to see huge volumes being shipped for a fairly small royalty, as that gives them a greater chance of eating away at the market share of much more costly chips (which also cost much more per design iteration). Part of that huge volume has to come from large volumes of low cost devices (including the current Internet of Things buzzword / concept), which essentially makes a very low cost CPU a hard requirement.

Bottom line: A major part of ARM's success is down to the low cost model. I can't see changing that model as being a good thing for anyone involved in it (ARM Ltd, SoftBank, the chip vendors, or the people deploying the chips).
 
Top