Apple to acquire ARM?(theregister.co.uk) |
Apple to acquire ARM?(theregister.co.uk) |
It would take a while for this to have any effect anyway, because I suspect that most ARM manufactures have licence agreements that protect the upcoming Cortex-A9 designs (NVidia Tegra 2, TI OMAP4 etc). The Cortex-A9 should be good enough for a couple of years.
Longer term, there are good alternatives to ARM around anyway. Atom is making good progress, and Loongson (http://en.wikipedia.org/wiki/Loongson) could be an option too. (The idea that Loongson could come to dominate the market is so ironic that I think Shakespeare would rise from his grave and write a play about it)
If Apple is really doing this I see it a play to take on Intel. The ARM instruction set is clean and well-designed, unlike Intel's. This enables ARM chips to draw significantly less power for the same performance. Now, who cares about performance-per-Watt? Anyone who has a cluster does. Anyone building devices that run on a battery does. These two groups are becoming everyone -- it's laptops/tablets and cloud computing all the way, baby. If ARM gets a 64-bit implementation and a fast floating point unit it can compete against Intel in the server market. Apple can fund that with their giant pile of cash. ARM already has the low end market sewn up. Imagine a laptop with 10 hours battery life and Google buying a million 64-bit ARM cores. That's two nice revenue streams to have, and it gets Apple into a big market where they currently have very little presence.
I did a blog post about some of these issues here: http://www.untyped.com/untyping/2010/02/02/is-the-ipad-the-b...
Ironically, if the worst happens (Apple buying ARM, shutting down licensing), Android would become even more the competition as it is already running on Atom and quite portable to other architectures. Intel would lose a major competitor, too.
I'm having a hard time seeing what an acquisition would accomplish for Apple that a nice fat contract could not for much cheaper. It's not as if chip manufacturers tell customers the size of Apple to just get lost if they need a custom chip for some purpose. A "nice fat contract" negotiation would look somewhat similar to acquisition negotiations for rumor mill purposes, too.
When Palm will be bought, i think the resulting company won't build smartphones anymore (the revenue stream) but will only be bought because of Palms patents.
I don't really think this will actually happen, but if Apple buys ARM, it would basically have power over the whole mobile phone market (plus settop boxes, receivers, TVs and almost every other embedded system). It would put the competition to a serious disadvantage for years (until competition is able to switch to a new platform).
Even just owning a core technology like ARM has, in the hands of one company, cannot be good for competition, the market and the customer in any way.
But as i said, i actually don't believe that Apple will buy ARM. That day, i will lose all the last hope i have in the regulation of the market.
Acquiring ARM only for the chance at an end-to-end vertically integrated design seems quite expensive, albeit not beyond the "insanely great" motivations of Mr. Jobs.
It's true that ARM has a compelling and growing market share at the lower end, but continuing the licensing business seems un-Apple-like; at the same time, shutting that business down would only hurt competitors a little - it would mainly serve to make the other CPU makers rush in and try to capture the resulting power vacuum.
Not beyond? I'd take that a step further and say that end-to-end vertically integrated design is very much a Steve Jobs sort of thing to attempt. He has said he thinks of Apple as a software company and, attributing it to Alan Kay, that "People that love software want to build their own hardware." That said, I'm not sure this would require total ownership of ARM -- just sufficiently large stake in it to exert some direction on its future development (sort of like sharing PowerPC with IBM and Motorola).
2) Antitrust problems generally happen when you coerce a company into something using your monopoly power - not when you close down a business (if you believe the bit about them closing ARM down - which I don't). If Apple tried to force ARM licencees to forbid the use of Android (or Flash! :)) on the chips, then maybe there would be a concern (abuse of market power etc).
ARM has a three-tier licensing model (http://www.arm.com/products/buying-guide/licensing/index.php...)
If you believe the wikipedia entry on ARM licensing costs, (http://en.wikipedia.org/wiki/ARM_architecture#ARM_licensees), then the average cost (in 2006) was $0.11 per core, averaged across all cores.
At at 50M iPhones, thats $5.5M (USD). Even if Apple is paying $2/core in licensing costs (effective, the A4 is the first non-Samsung part used by iPhone OS), thats still $100M, so $8B doesn't pencil out.
I do find it incredibly likely though that Apple would 'buy out' their license(s), and negotiate to have a minimal (maybe even zero) cost license to all future ARM IP.
Or they could just buy a controlling interest (probably wouldn't take 51%) and frighten off everyone who competes with Apple.
Just the slow-down for the rest of the market alone (Atom isn't anywhere close to ARM in terms of functionality/watt), it might be worth it.
So Microsoft buying Intel wouldn't generate antitrust problems?
But I wish they would buy Foxconn, so that they can control manufacturing too!
We're at the point where there's only one platform that isn't doing multi-architecture development, and even Microsoft has Windows CE / Mobile running on ARM and x86.
With the way things are going, it's conceivable you could see mobiles running MIPS, Power, and half a dozen other technologies. Apple could lead the way if they bought up a couple smaller companies.
The Company was founded in November 1990 as Advanced RISC Machines Ltd and structured as a joint venture between Acorn Computers, Apple Inc and VLSI Technology
Related question: is there actually public information on who are the major stockholders of a publically traded corporation? I'd love to see that.
Because it's not the ARM implementation we depend on - it's compatibility with its instruction set.
I really hope the acquisition won't get through, I can see Apple doing just that.
But it doesn't make sense - Apple can't stop the existing licensees from stopping producing their chips, and the companies with architecture licenses will still be able to make new, compatible designs to sell to third parties. What it might do is stop innovation in the ARM ISA being made available to other companies, but that's a very long term thing; a new ISA takes many years to get to product silicon.
The whole ARM business model is based around sharing the costs of development (which are substantial) with a large group of companies. Could Apple really justify sustaining that cost to prevent competitors from using future ARM designs, when there are (albeit perhaps inferior) alternatives?
But at a very superficial level: MS & Intel are in different markets, so there is no obvious objection.
IANAL etc etc
But it is funny to speculate as a laymen - I mean what if Google bought AMD instead? That would give them a crazy position to experiment with hardware and keep a check on Apple - after all, Google means serious business with the Nexus one and the mythical tablet. And we all know Google likes to experiment with 3D stuff in browsers and such. Again, I am just blithering- with Google's purchase of AMD (+ATI), they might just have crazy nice open-source 3D tools, maybe even a cloud platform for map-reduce rendering work (take that Pixar - yet another Steve Jobs production).
But on the other hand, Microsoft is the one who could benefit the most from buying AMD. For one thing, MSFT gets ATI for Xbox 360, a platform that MSFT has well executed. And now, Microsoft even make their own chips from win 7 mobile phones to consoles. To really keep Apple's market cap away from them, MSFT panics and makes their own hardware and opens up retail stores :-P
They have a history of participation in the AIM alliance. The Altivec instruction set is practically their baby. The reason they had to switch to x86 is because neither supplier (Motorola nor IBM) was willing to target the performance-per-watt that Apple needed to execute their laptop designs. If there is any truth to this rumor, I would guess it has to do with securing the roadmap so they don't get blindsided by suppliers in the future. That makes much more sense to me than blowing 8 billion dollars to scuttle competitors.
Just the slow-down for the rest of the market alone (Atom isn't anywhere close to ARM in terms of functionality/watt), it might be worth it.
ARM licencees already have next-gen (Cortex-A9; ie: Tegra-2 etc) chips coming this year. Any slow down in new designs from ARM won't effect the market for a couple of years as these chips are much faster than the current designs (Cortex-A8; ie: Snapdragon, TI OMAP 3000, Apple A4)