The Misery Of The Downturn


The most miserable thing about downturns is job losses. If you keep your job through a recession, the recession doesn’t hurt you. If you lose it, it hurts like hell.

As usual, the harshest cuts seem to be happening in the US where bankers’ analysts publicly cheer on CEOs who cut jobs.


Last year, Silicon Valley lost nearly 12,000 jobs in what can only be regarded as pre-emptive strikes by CEOs who, while largely unaffected by bad trading figures, wanted to make sure their figures would look OK if a downturn took hold.


This year, things are different. The downturn is taking hold and companies are being hit by bad figures.


This week Intel, after a 90 per cent profits drop last quarter and projections of a possible loss this quarter, said it will close three back end plants and two fabs with a potential loss of up to 6,000 jobs though it is said that it will try to re-deploy some people.


Motorola, which sold 19 million handsets in Q4 compared with 25.4 million in Q3, and 28 million in Q2 has said it would lay off 4,000 people, 3,000 in the handset business.


Microsoft is to lay off 5,000.  TI will cut 3,400 jobs. Sprint Nextel will lay off 8,000.


Advanced Micro Devices has laid off 3.300 people in the last twelve months.


Freescale said it would cut ten per cent cut of its workforce during 2009 which will see 2,400 employees losing their jobs.


ON Semiconductor laid off 1,500 workers, imposed unpaid time off, and temporarily closed plants.


 In Europe there has been some attempt to mitigate the effects of the downturn on workers. About 1800 workers in Infineon’s Dresden fab will reduce their weekly working hours for six months, starting in February.


ST closed its fabs in Catania and Agrate over Christmas, but recently announced it would lay off 4,500 people this year.


Philips, however, is laying off 6.000 employees. 


Taiwan seems to take the view that it’s best to look after valuable fab workers rather than ditch them at the first signs of trouble, with TSMC and UMC announcing unpaid holiday programmes for the year.


Hopefully the new US President’s advocacy, in his inauguration speech, of cutting hours across the board rather than sacking people, will catch on.





  1. Big Softie, you make an interesting point when making that argument about SoCs, rather than process technologies. Discussions are often limited to Moore’s Law and how such a phenomenon would mean the process cycle would not necessarily halt, but at least gradually slowdown.
    In the case of SoCs, the same is indeed true. However I think there’s an interesting point to be made for them: they are made of many small parts, and every generation does not require you to redesign every part. Therefore what will happen is that there will be more and more reuse between generations of SoCs.
    A logical extrapolation of that is major tool changes will be minimized even more aggressively in the fab world, with things like EUV being put off as long as possible – and probably even further, if you look at how long 193nm dry litho has survived. Perhaps the same will be true for immersion technology. I would tend to believe that this kind of slowdown of innovation’ isn’t as bad as a lengthening of the refresh cycle…

  2. Big Softie. Gosh that’s a relief! I’m glad to be wrong about your state of good-cheerfulness. Sorry to miss your point. Now I understand it I appreciate you are absolutely right.

  3. David,
    Fortunately I am brimming over with good cheer, and I didn’t intend my posting to appear pessimistic 🙂 I was attempting to look at the history of the industry and the facts objectively, and extrapolate forward, though I accept any future view is subjective by definition, and open to significant influence by external market factors….such as arrival of some new killer whizz-bang end-user product that no-one foresaw.
    My point was emphatically not about the costs of developing next generation processes, I specifically said next generation semiconductor products. i.e. the scenario where Chipmaker A is developing and supplying SoC platforms for e.g. wireless devices or STBs together with all the embedded software and design-in costs that we discussed in another thread. My point (question) is whether in a fiercely competitive market the profit from this generation of product will be sufficient to develop the next generation platform, which will demand even more video processing, more audio processing, more software, more design-in cost etc. Analysis shows that R&D costs of overall semiconductor product development including all the software and design-in cost is increasing exponentially, just as the ASP is reducing exponentially courtesy of Moore’s Law. Logic suggests that with a modest market growth (the population of the planet), such price erosion of end-user products cannot be perpetually sustained against increasing development costs….and price erosion appears driven more by competition and historical expectations based on manufacturing costs rather than a real value-based calculation.

  4. My God Big Softie you’re as pessimistic as a stockbroker’s semiconductor sector analyst. Now I know why Infineon shares cost under a Euro, ST shares are at 4Euros, and Intel’s at $13.
    Your last sentence is apocalyptic. If developing next-gen processes costs more than ther collective proft of last-gen’s products, then is this the end of Moore’s Law?
    Well I don’t know the answer. The economics of this industry have always baffled me. Somehow there always seems to be someone to pick up the tab, someone with a new idea which rescues the industry when on the brink of disaster, and a new killer product which send everyone off to the races again.
    Will there never be another sucker (I mean government wanting to invest in the industry)?. Never be another great idea? Never be another killer product?
    Somehow I don’t think so.
    Cheer up Big Softie

  5. A stronger industry will emerge from the ashes
    Sorry, I don’t believe this will happen. The semiconductor industry has always been cyclical, and swung back and forth between supply chain shortages (high prices, high profits) and over-supply with correspondingly lower prices and margins. Historically this was a function of manufacturing capability and capacity; the industry had to wait for the next step down of wafer processing to catch up to supplying the market; new fabs had to be built and commissioned ready to fuel the next wave of lower cost mass market killer products that needed the increased supply chain capacity, and the attendant lower chip ASPs to develop into a profitable market.
    This time it’s different. We are already in over-supply mode, and coincidentally an economic recession has come along. There are no new killer products on the horizon, and in general we all have more than sufficient PC processing power, flat screens, STBs, and mobile handsets in our homes and lives. There are still markets to win in the less developed areas of the world, but bottom line is we have a global over-supply situation with no immediately obvious outlet to take up the capacity once the economic situation improves.
    Costs and prices must have almost bottomed out. You can move your manufacturing offshore, followed by your business management, and your R&D. Many already have. But once you’ve located all your operations in the cheapest place on the planet there’s no way to go any further down without taking pain. In trying to encourage Joe Public to buy more products that we don’t need, the whole chain from semiconductor vendors to OEMs has ferociously driven prices of mass market products to ridiculously low levels.
    I believe we have reached the crossover point where the inexorable increasing costs of developing the next generation of semiconductor products has exceeded the profit levels available from the current generation to fund it. Maybe the endgame of such fierce price competition is stagnation of innovation?

  6. The State of Saxony may come up with a new deal. it is very keen to see Qimonda survive.
    But the DRAM industry is a collective madness. It is losing huge amounts of money and is over-supplying the market. Clearly the industry is best served by some suppliers dropping out. But that’s no consolation for those who drop out.

  7. Not forgetting Qimonda, currently in Protective Insolvency in Germany. At the very least there will be serious job losses, worst case it will close completely. It is unfortunately not a terribly auspicious time to be searching for major new investors, as they are trying to do now. We can only hope a metaphorical rabbit can be pulled from the hat, but will that be good for the DRAM industry as a whole? The biggest tragedy of the Qimonda story though is that their considerable technological lead has not saved their bacon.

  8. Friedmanite, Yes but the scale and suddenness of it still come as a shock. Only a few months ago people were saying how can a crisis in the housing market affect tech? But now it’s happening

  9. Twas ever thus. It’s called ‘Creative Destruction’. A stronger industry will emerge from the ashes, let’s just think about the people getting hurt in the meantime.

Leave a Reply

Your email address will not be published. Required fields are marked *