Electronics Weekly Electronics Design & Components Tech News
That hoary old chestnut, a possible takeover of ARM, is up on the billboards again.
The received wisdom on this is that there are three reasons why it would never happen:
1. A takeover of ARM would be such a disaster for the whole industry that the industry wouldn’t let it happen.
2. A purchaser of ARM would have to pay about $10 billion to buy a company with $600m of annual revenues.
3. The effect of a takeover would be to destroy ARM’s value because all its licensees would jump ship and look for another processor architecture.
On point 3, there’s a snag: potential ship-jumpers might find it extremely tricky to use ARM code on a different processor architecture, and it might be too late to economically write enough code on a new processor to replace existing ARM code.
On point 2: it would only be worth it to someone who wanted to destroy ARM as a potential rival e.g. Intel
On point 1, ARM co-founder Tudor Brown hinted at the power of the argument recently when he said: “There are a lot of powerful people out there, a lot of big companies — the Qualcomms, the Samsungs, the TI’s, the Nvidias, the Motorolas. If any one of those companies made a play you can imagine the other ones creating quite a sort of rear-guard action.”
One wonders how this might work but, in the event of a bid, presumably some sort of consortium to put up a knock-out counter-bid could be put together.
But there’s a silver lining in every cloud and, if there was a change in the dominant mobile processor architecture, there’d be a bonanza for software writers.
I, like many, would like to see ARM remain independent. Their cores are everywhere, the company should be a monster, but with only 600m in revenues they are giving away their IP. With that much market penetration the discussion should be who Arm is buying.. but its not. Their problem is that if they raise their licensing will people jump to another core? Who? Mips? Im not sure but I think at some point they are going to have to do it, or else whoever buys them is going to try it anyway.
I do agree with the general conclusion, but I don’t believe porting software to another architecture is a non-starter. There has always been a lot of fear-propaganda from ARM, trying hard to lock-in. Put another way, port Android to a new architecture and you are done… at least for the large part of ARM devices that are used in phones and mobile products. Hypothetically speaking, if ARM disappears into a bigger entity, Google will have no hassles porting Android to a new architecture. After all, given Microsoft’s recent announcement on how much they love non-Intel architectures…why would Google be wedded to ARM?