Oracle wants to buy chipmaker, ARM shares rise | Hacker News Reader

Oracle wants to buy chipmaker, ARM shares rise

bloomberg.com

37 pointsdmillar43 comments

AMD is mentioned as an alternative target and that makes a lot more sense than ARM.

* ARM is not a chip manufacturer, it is a CPU architecture design house.

* Oracle runs on big to medium iron: x86, Power (IBM), SPARC, and HP (Itanium, PA-RISC). They don't run on ARM and I would speculate that it would be a major effort to port to it. Oracle already owns SPARC.

ARM is not a chip manufacturer, it is a CPU architecture design house.

So is AMD after they spun off Global Foundries.

They're still not in the same category. AMD designs chips that are fabbed by GF. ARM designs cores, then hands them off to TI/Nvidia/Samsung/etc. to design into chips, which are then fabbed by yet another company.

"Oracle to possibly buy ARM" is a pretty strange summary of the article, which actually says (1) Ellison says that Oracle might buy some chip-making companies, (2) stock in ARM rose, apparently in response, but (3) one of the two sources cited by the article says that it's very unlikely that Ellison had ARM in mind and the other goes no further than that it's possible that he had.

You're right. And the last thing we want right now is to see another company go down in patent flames.

Agreed, updated.

it just goes to show the stupidity of the market. ARM takeover was speculated for months. Now that someone with a lot of cash actually mentions that they want "chip companies" the market goes crazy and pushes up it's first choice even more. Nevermind that buying ARMH makes zero sense for ORCL. Though stranger M&As has happened...

Edit: though judging by today's jump MIPS and AMD seem more of a choice

I certainly hope that it is just speculation. I've never seen a CPU that was as much a joy to code for. I've done some assembly for various architectures, and nothing at all compares to the elegant, powerful simplicity of the ARM.

Based on Oracle's treatment of Sun, it would be a crying shame for ARM's design team to share a similar fate. No one needs a world where "movsne r0, r1" is going to violate some Oracle license agreement.

This makes even less sense than Intel buying McAfee. Sure, there are ways that ARM could be integrated with Oracle's products, but they don't need to buy them to do that.

They didn't need to buy Sun to run on their servers, either. But they clearly have ambitions to own the whole stack.

They didn't buy Sun for their servers

Why on earth would Oracle buy ARM, they're not exactly the kind of processors that Oracle would benefit from owning.

AMD, maybe, but that might come a bit too dear even for Oracle.

If there is one party that I think might buy ARM at some point it would be Apple. Or some other mobile devices manufacturer.

ARM is low power. For applications which scale horizontally, Oracle could build a server with lots of ARM's inside. There is already a startup (Smooth-Stone) which proposes to do exactly this.

Also, an Open Solaris port for ARM exists. Oracle may plan to extend/rebuild it.

http://hub.opensolaris.org/bin/view/Project+osarm/WebHome

It's a lot harder to stick a database like oracle on a mini farm of very small processors than it is to throttle down a larger CPU during idle moments.

For the most part database servers that run Oracle are are not 'underutilized' that I would expect a huge pay-off from that.

The systemZ port of open solaris has already been killed by Oracle, I would not be much surprised to see them do the same with the ARM port (as far as I know it's not exactly the most popular platform for open solaris).

If there is change on the part of Oracle with respect to open solaris on ARM I'd expect it to be of the terminal variety.

I would argue that ARM is the biggest threat to Intel/AMD in the datacenter over the next ten years. Datacenter design these days is all about computation/joule/$, and ARM has a big lead by that metric. Smooth-Stone is already working on ARM-based server designs, and I'm sure more are in the works. Seems like exactly the kind of processor Oracle would want to own.

Databases are not about computation, they're about throughput, IO, vast gobs of memory, it goes much further than being simply compute bound.

Oracle is a database vendor first, OS vendor second, I can't make the case for them becoming a chip manufacturer, but who knows, I've been wrong before.

They may be a database vendor first, but they also have a big applications business. They already have a good processor for databases (sparc), and ARM is exactly the kind or architecture you want for your app servers. Btw, I'm just playing devil's advocate. I sincerely hope that ARM doesn't join the Oracle graveyard of once-great technologies.

> I sincerely hope that ARM doesn't join the Oracle graveyard of once-great technologies.

And maybe part of me is so scared of that happening that I see all kinds of reasons why Oracle really shouldn't buy ARM.

Those big companies are like black holes, stuff goes in and it's anybodies guess what will come out.

What's with all these rumours about companies planning to buy ARM? ARM is a famously profitable, 30 year old company that is on the verge of world domination. What makes anybody think that they want to sell?

Seems like buying ARM could set off antitrust allegations. Oracle already strong-armed itself into the chip business with the purchase of Sun, going for ARM, which aside from Power is the only credible x86 competitor gets on to pretty shaky ground.

At that point you would just have AMD/Intel on x86 vs. Oracle. Not a monopoly, but I wouldn't expect that to be a good situation.

ARM is really ubiquitous in any device that isn't big iron or a personal computer, so you can bet the FTC would be all over any attempt at purchasing ARM Holdings, no matter who is buying.

I think the only reason we haven't seen serious antitrust litigation against ARMH is because they're very good about licensing all the technology they produce to anyone that can pay and generally avoiding anti-competitive behavior. The FTC is going to be very nervous about that liberal licensing scheme going away if someone does purchase ARMH.

Doesn't look plausible to me -- Oracle's Sun business uses Sparc and x86 processors. If the chip-maker purchase is supposed to complement Oracle's existing products, they should rather buy a x86 or Sparc company. Unless they plan to throw away the existing Sun designs in order to replace them with efficient ARM-based servers (then why buy Sun).

They could buy ARM just for the dividends, but then they could also buy a fast-food chain.

Agreed. This Didier Scemama ought to be laid off!

I agree with you, this story makes little or no sense.

But he seems a serious guy that knows what he's talking about: http://uk.linkedin.com/in/didierscemama

[edit: typo]

Someone will surely buy them some day, but doesn't seem logical that Oracle would. My money is still (literally) on Apple buying them in the next couple of years.

I think it's unlikely for Oracle to buy ARM. They are more likely to buy chip companies in $30-300MM range, not multi-billion companies. But, I see why Ellison would like to buy ARM:

* Buying ARM will give Oracle more leverage over Google in their Android/Java litigation.

* Combining their SPARC and Server expertise with ARM Design expertise will give them competitive advantage in Low Power High Density Datacenter plays.

I would expect that Apple, Google, Samsung, TI, and Qualcomm would do a little bit of a panic and try to put in competing bids. ARM works because it is a neutral party, and Oracle would be regarded as anything but a neutral party. This could have a large impact on a lot of companies, and some of those companies have much deeper pockets than Oracle.

Give Oracle already has SPARC, perhaps he meant GLOBALFOUNDRIES?

I don't know how much panicking Apple would do; Jobs and Ellison are good buddies.

Apple has an architecture license for ARM, so they don't really care what happens to ARM the company. It could shutdown tomorrow and that would not impact future iPhones in any way.

Rumor aside, I'm surprised in general that ARM's stock has risen so much in the last year. I suppose the increase in price is due to the current, and anticipated, use of ARM designs in mobile devices. Still, in the long run would ARM really be able to maintain it's lower-power chip dominance over Intel? I know Intel isn't there yet, but they seem to be a fearsome R&D competitor.

There are a LOT of ARM processors in use. It would be interesting how much money ARM gets from the smartphone use compared how much it gets from all the other applications.

This is completely rational if you consider that Larry Ellison's goal is to build TJ Watson's IBM. Oracle wants to be a systems company instead of a software license company.

http://www.zdnet.com/blog/btl/ellison-wants-to-model-new-ora...

Why was this downvoted? It's perfectly obvious Ellison wants to be an Apple of enterprise computing. You don't guy and buy megabytes and gigabytes from apple, you buy something that just works.

Ellison wants that too. You pay $bucks, plug in a power cord and an ethernet cord and your erp/accounting/db setup is done. If I were a decent sized company I'd pay beaucoup bucks for that.

Personally, I commend him. IT is a cost centre for most companies.

Doesn't it look disturbing that premier financial data company, the source of business & financial news, needs to describe ARM Holdings Plc as "designer of chips that power Apple Inc.’s iPhone"?

Can someone comment on the implications of this for Google, Apple and Microsoft if such a thing happens? All major software companies in the non-pc space are mostly dependent on ARM.

It all comes down to the royalties ARM will charge the chipmakers. If the price of a TI OMAP or Freescale i.MX jumps 100%, that seriously hurts the price of phones and netbooks.

Google, Apple, and Microsoft really don't get hurt on the software end, just the hardware pricing end.

ARM's average royalty per chip is currently around 4.5c. http://www.arm.com/about/newsroom/arm-holdings-plc-reports-s...

Of course, that's a blend of the royalty for all their cores, but at that level a 100% hike won't have a material effect on a $400 smartphone, even with 2.6 cores/phone (again, average from above source).

Ew, Please no.

You may take our Java, but you will never take our ARM!

I see Larry needs low-power navigation equipment for his new boat...

Oracle seems to be too focused on acquisitions and litigation.

Why?

Power density in the datacentre.

Buy ARM to ruin it? ^_^

Jokes aside, seems like the target is AMD. They're in trouble, like Sun.

Reply on news.ycombinator.com