News (I guess not unexpected) this morning is that Microsoft is cutting staff, Intel is closing down underutilized resources and cutting staff, IBM is cutting staff, and we heard yesterday from John at InsideHPC.com of more cuts at AMD.
Having been on the wrong end of RIFs before, I know what it is like. I empathize with those effected. Having run a company for 6+ years, I know the abject terror of the other side of this. It is very sobering. When credit is available, you can generally borrow to keep paying people to help ride out the economic soft spots. Credit is not available.
In all these cases, these companies have core businesses, and HPC programs or business units, or groups. None of these businesses have HPC as a core business. Which means that the magic words you need to be listening for are “focusing upon core business”. That could have some significant (negative) implications for HPC at these companies. I know of two of the list here where it already has.
Put another way, the accountants and management at each organization has to ask some very hard questions, such as if we invest $X in this group, will we get $X+$Y with $Y positive going forward, and how large is $Y as compared to $X? For companies developing product and looking to break into a market, $Y is negative. The rate at which you are spending $Y is your “burn rate”. Startups know this all too well. After $X is gone, you either need a new $X, profit from ongoing operations ($Y) to feed into $X (after taxes, costs, …). It usually takes a while to get the earnings ($Y) up to the same level as $X.
The decision management must make is, if we commit this $X here, can we maximize $Y? Moreover, they need to look at alternatives, including, if we reduce $X by $C (removing costs/expenditures … e.g. people and facilities), do we stand a better chance of maximizing $Y, or even simply stabilizing it in a difficult economy.
While HPC has been growing in dollar volume year over year, and has been driving server business worldwide, even in the face of declining IT server expenditure, it is still quite a bit smaller than the base IT market. This suggests that it may be a tempting target to cut investment in.
Especially when profit margins for HPC clusters are so slim. Customers simply don’t want to pay more for HPC. They want to pay less. So they demand lower margins, bigger discounts, etc.
This has an impact. A less profitable business unit, or even a loss making business unit is prime for getting cut in hard times. A startup with no foreseeable revenue going forward will either shut down, or trim back to skeletal dimensions to try to weather the economic storm.
There are vendors in the market now with unsustainable cost structures. They will be culled. The bloodletting will likely be brutal. Look at your server rooms … imagine say 10-20% of the product vendors who supplied things for your HPC systems going away.
The bloodletting is not done by any stretch.