Apple should buy it...
I was discussing this on Ars Technica about it given Apple has hired an ex JPMorgan guy to find and buy companies with it's $40b cashpile. Personally I've always thought that Apple should've stepped in and bought SGI even when SGI first started experiencing trouble.
Anyway, I thought it'd be an interesting discussion to repost here:
I think there are 2, possibly 3 good options for Apple: SGI ($280m) and Autodesk ($6b). So for $6.25b, they get a metric sh*t ton of capability, and *still* have $34b in the bank!
Why?
- Apple would inherit the architecture (AutoCAD) and 3D (Maya, 3DS Max) markets overnight and complement their market leadership in video (Final Cut Pro, Quicktime) and audio (Logic, Digidesign) production software.
- Mac OS X would benefit immeasurably to the point it would probably be respected as the best UNIX OS out there. SGI's IRIX has XFS, far better than HFS+ (and given time can probably better ZFS), invented OpenGL which will complement Apple's OpenCL efforts, and if Grand Central Dispatch is the best threading optimisation Apple can do, the mind boggles at what SGI can do
- SGI know big iron. A Mac Pro is a nice machine but if you're in the market for a $25k workstation, a Mac Pro is a little anaemic. SGI's latest Octane is an enclosure housing 10 dual socket blades: 8 cores in a Mac Pro versus 80 cores in an SGI Octane. Think that's funny, you then have the servers: Try comparing an 8 core XServe with a 2,048 core Altix UV supercomputer! And they both use Intel.
- SGI know all about graphics. Both companies have a good relationship with Nvidia, but if Apple wanted to, SGI know how to make GPUs, especially pro ones. Apple of course probably wouldn't want to, but developing game-friendly hardware as a side-effect wouldn't be a big deal either. Did I mention SGI invented OpenGL? Needless to say it might be the right way to give Apple the kick it needs to provides decent graphics iron in their hardware. I mean A GT120 even being offered in a Mac Pro is just embarrassing!
- SGI has the envy of the geek in the same fashion as Apple is the envy of the consumer. While SGI's supercomputer business is totally irrelevant to Apple, supercomputers are profitable and the R&D windfall is worthwhile. Think of supercomputers as a racing team to a car manufacturer... let alone getting intimate knowledge of how spooks work (and helping the Apple secrecy cat and mouse chase And remember, Apple are developing OpenCL, and the primary beneficiaries of the unleashed GFLOPS from GPUs are precisely the same as supercomputer customers...
- Corporate IT depts would take ages (like a decade) to be convinced of using Apple hardware, even with SGI's credibility in big iron. But if Apple can sustain a business providing big iron regardless, then corporates might begin to swing in their direction over time. Instead of actively chasing it, Apple should nurture an ecosystem for corporate software to flourish on a combined SGI-Apple platform. They'd need virtualization (partner with Parallels or VMWare already closely here). They can offer storage through SGI, (but EMC who owns VMWare is better as a friend than an enemy), and they need to have hardware credible for a database. From those a corporate IT agenda might develop.
- In buying SGI, Apple would make me want to buy even more of their hardware :(