# Not surprised … IBM pulls plug on Blue Waters

I say I am not surprised for their reasoning … not that I had an inkling that they would do this before hand.
Basically they pulled the plug because the costs were growing far faster than they planned, and they couldn’t afford to deliver the machine at the requested price. Which makes perfect sense to a business that has to consider profit and loss, but maybe not so much sense to research groups that want things.
I’ve said many, many times … that really bad deals can kill good companies. Not that IBM would be brought down by one bad deal, but fundamentally, IBM has to tell its owners (stock holders) what it did in its business and explain to them their results, every quarter.
Imagine, having to explain a deal that cost you $300M and you only received$30M for it. Or even \$60M.
Yeah, someone high up would get fired for that. No, lots of people would get fired for that.

And again, I’ve argued this many … many times … name cachet means effectively squat in most cases in the market. It adds no value to your bottom line. It doesn’t really help you win more business. No one really makes a purchase decision for a new supercomputer based upon what UIUC bought. No one would give this any more credit than it deserves … which is fairly little.
Clusters of APUs (accelerator processor units) will likely be the delivery mechanism for Blue Waters … if its not canceled outright. I hope not.
Like other projects, conceived during a different era, funds may not be available to fully fund this effort now. I’d argue that HPC is a critical strategic asset and should not be cut deeply. But we do need to make sure we are getting good value for our money. Being able to learn how to build and program such machines is IMO, critical for business as well as for academia.
But we also now live in a world where we can no longer afford everything, we have to pick and choose more carefully. We have to use our more limited resources more effectively. Which is in part why I think IBM had to pull out, they realized that the the cost benefit equation had drastically changed … in part due to technological considerations, but in larger part due to business considerations. They just couldn’t afford this, and they faced up to it.
This is why they’ve been in business 100+ years, and why they will likely remain so far longer than this.
We as a country need to prioritize our spending, and our cutting. Drastic and draconian cuts are on the horizon. I’d hate to see Blue Waters cut. But it is possible. The benefits need to be clearly delineated so as to outweigh the costs. Moreover, the costs will need to be reduced somewhat to deal with the smaller amount we will likely be able to spend.
Its not pretty, but if we handle this right, it shouldn’t be terrible either.

### 6 thoughts on “Not surprised … IBM pulls plug on Blue Waters”

1. I’ve heard a different version. One that’s still driven by business decisions balancing loss and gain. I don’t have cite-able details and won’t pretend to “truth,” but your version above is a bit one-sided considering who paid to develop the technology in the first place… They never would have built that interconnect or modified the Power7 without the sponsors.

2. @Jason:
Quite possibly one sided … I don’t have all the details, and I made specific assuptions. There is no doubt that this is a complex situation.
But I don’t fault IBM for walking away. Its often very hard to do the right thing, or even know what that is. In the case of a business, you can’t simply give away money. This just doesn’t work. Its a recipe for going out of business or making a loss.
Its arguable that a CRADA mechanism would have made more sense than an acquisition. Seriously, for jointly developed technologies, this mechanism just makes more sense.
I hope more details will emerge, and I hope that Blue Waters can be built with a different technology. I wonder if they can hit the same price point. No vendor in their right mind will buy this business. They can’t afford it.
We’d be happy to provide storage for them … siClusters built on our DeltaV are extremely cost effective, and quite fast. But if we were asked to give our units away, or subsidize the machine … we have to ask what the net to our business would be. And the answer would be a loss. This isn’t the way to run a business.
Honestly, I do expect to see changes in the way vendors respond to research/edu procurements going forward. The days of massive discounts to beat the other guy, profit be damned … I think they may be over.

3. also, I should point out that we are asked for this with some regularity. And we simply reject such calls. There’s no benefit to us to give things away. There are no additional sales we will get by doing so. Our investment in the hardware, software, and service will simply result in a loss. So asking us for this is a waste of time, as we will say no.
But it still happens. With some regularity.
Usually accompanied by the tired old statement that today’s students will be purchasing tomorrow when they are profs. Which fails to note that they will all ask for the same deal. Because they do.

4. very true about walking away – captain Joe even gave me a hard time over the T’s and the C’s.
Dude got the SGI kool aid “today?s students will be purchasing tomorrow” and just spat it out.
Like the Zohan – you just don’t mess with any three letter company.
Be it I, B, M or J, O, E 😉

5. @James
Owie !
I used to hate (and well, I still do) all the T&C bits. We had really simple, what I thought were … practical … ones.
And then we got taken advantage of for the first time.
So we added a clause to prevent that.
And then a different way the second time.
And so on.
I used to hate these things. And I still do. But I understand them. And we don’t get taken advantage of so much any more.
Its not research/IT staff thats the problem. Its purchasing staff.
My all time favorite are still the one at a local university that wanted us to teach our competitors how to build what we were building so that they could more effectively compete with us, because she didn’t want to buy from us. We aren’t selling to that university (at all), and are simply waiting for her to retire.

6. Looks like a little more info is emerging. Short version, execs at IBM don’t want profitless or loss making revenue any more.