It was nice to see our stock price going up today while the rest of the tech sector was falling; the cause was an upgrade from an analyst at Lehman Brothers. Apparently, he likes our new CFO; the CFO’s last name perhaps has something to do with that, but the analyst also thinks that the new CFO will signal a round of cost cutting. I confess to being a bit taken aback by the following quote from the analyst:
Last week, [the new CFO] indicated that all areas of Sun are being reviewed for reduction including the research and development department which had previously been relatively protected from cuts; however, given that most of the new products have been introduced, R&D may also be targeted.
I know we’ve been introducing a lot of good stuff recently, but why does the analyst think that we should stop? I think the Rock guys would have something to say about that; and while I’m a fan of our new Opteron servers, I don’t think it would be a good idea to rest on our laurels after introducing a pair of two-socket servers. Don’t get me wrong – if I wanted a 1U or 2U two-socket Opteron server with lots of maintenance and reliability features, they’d be what I’d buy (and, as it is, I’ve bought a couple of our slightly lower end one-socket Opteron boxes) – but surely we’re not going to stop there?
After all, AMD doesn’t just sell one-way and two-way Opterons: they also sell an eight-way model. So wouldn’t it be more fun if we, say, added an eight-socket Opteron server? At two cores per socket, you have quite a lot of processors to play with; I haven’t had a chance to use one of the prototypes myself, but I was pretty amused when I got e-mailed a boot message from one of the machines. I will say that using a 4P (all single-core) Opteron box does nicely speed up compilations; I’m not sure exactly how a 16-core server would do on our builds – the bottleneck might well have shifted away from processing power by then – but doubtless it’s the bees’ knees for many applications. (E.g. high-performance technical computing, if Tokyo Tech is to be believed.) If you need lots of threads to serve up web pages, or similar tasks, Niagara might be your machine; if you need a broader class of computing power, these guys should do nicely.
The article also mentions a forthcoming blade server. Which is, honestly, an area I don’t know much about, but I’m pretty sure that Andy and the rest of the NSG brain trust have come up with some nice designs there, too.
And meanwhile, the Itanium bad press just keeps on appearing…
Post Revisions:
There are no revisions for this post.
The key phrase is “most of the new products”. Your analyst obviously believes (unconsciously) that ideas for new products are in a bucket somewhere, and that the bucket is now almost empty, and there won’t be anything you can do to fill it up again.
3/10/2006 @ 5:18 am
I think they are talking about cutting core R&D (ie Sun Labs), not product development efforts like the Galaxy product line.
3/10/2006 @ 8:59 pm