Should French ST Split From Italian ST?

The time may have come to turn the clock back 27 years to when SGS Ates of Italy and Thomson Semiconducteurs of France were separate companies.

In 1986, SGS’s CEO Pasquale Pistorio and Thomson’s CEO Jacques Noels began talking about merging their companies. In 1987 they sealed the deal.

Now, a generation later, the French and Italian parts of what became STMicroelectronics seem destined for different fates.

The Italian half has a profitable, slow-growth, low-ish margin business making discretes, analogue, MEMS and automotive ICs. It’s not a business that needs cutting edge technology, it does not need vast capex, it is highly cost-sensitive and it is reasonably protectable in the light of ST’s long experience and IP position.

The French half is engaged at the leading edge of the digital IC business with a facility in Crolles which is not far behind the best of the rest of the world in IC process technology. The French half has the promise of growing a huge business in digital ICs but its capex and R&D needs are also huge.

The French side of ST has aspirations to turn the IC world on its head by getting FD-SOI adopted as a mainstream process technology in competition with the general worldwide consensus to go with finfet processes.

Finfet is proving to be a difficult process but it promises high performance ; FD-SOI is a simpler process which delivers low power.

Subject to the approval of the China government, ST has inked a deal with the China foundry SMIC to provide volume FD-SOI wafers.

It may also do the same with Samsung which is currently one of ST’s foundries for bulk.

The French government is behind the French side of ST in its aspirations to become as good as the best in the world at IC process technology, and the EC support plans for IC technology are also backing the French side of ST in its ambitions.

The only snag is that the Italian side of ST doesn’t want to go down this French route of growing a digital business on the back of a process technology which has little surrounding ecosystem.

Italian ST gives priority to shareholders; French ST wants to build a cutting edge technology business.

Can these twin aspirations sit together in the same company?

Can a CEO be committed to both ST’s shareholders and to building a leading technology company?

Should we go back to 1987 and undo the SGS-Thomson merger?



  1. Wise words Grax, ‘Go West young man’ is as true today as it was 150 years ago. The Italians are in safety-first, look-after-the-pension, don’t take a risk mode But things can change and the Italian government is selling its stake in ST. If the French take control maybe they can bring some ambition and élan to ST.

  2. in France, the problem rises from the fact that Bull is partnering on large data center project with Intel and Avago; after the failure of STE and the expulsion Didier Lamouche, ST Italy management are -dominating- STm RD and product strategy. ST Italy will NEVER invest in FDSOI in the future case they are afraid of stock market response after their mediocre financial performances; today they just wait for French gov public investment in semi (nanoinnov) together with CEA.

    The solution to this problem? … build a startup developing ARMv8 server.

    this proposal has been submitted couple of years ago to STm top management … and rejected cause italians were afraid to loose they dominance in digital and no BIG french manager was supporting it because they risk was involved.

    At the end: italian keep the POWER (of an empty digital departement about to die) and french didnt assume any risk.

    Typical italian, typical french.

    Lesson for young engineers: do you want to work on big project inside motivated teams, following the dream to compete at international level?? the EMIGRATE TO USA, where food and wine isnt good indeed, but MANAGEMENT is by far better as well as your salary

  3. Well, Contrarian, I don’t think Intel, Samsung, Toshiba, TI and the DRAM guys would agree with you and they’re a hefty chunk of the industry. I like the ST strategy because it’s a good old-fashioned semiconductor play – get an edge in process technology and you win the day.

  4. Besides the valid considerations exchanged above, the key issue is that ST is maintaining a costly admin infrastructure to manage the internal and external manufacturing. The transition to become a lean manufacturing IDM is hitting on a great layer of admin jobs who are probably redundant once ST is adopting a focused manufacturing of the key technologies and outsourcing the rest. My opinion in observing the market is that the semiconductor business game is shifting towards the accretion of a large IP inventory and super fast execution of new products. Manufacturing is a secondary but very costly and defocusing issue.

  5. Ah yes Keith, I seem to remember that when Plessey Semis won the Queen’s Award for exports in 1997, it had an astonishingly high proportion of foreign sales. The figure of 80% springs to mind but that seems far too high. Mind you Plessey had previously won a Queen’s Award for exports in 1978 because it was designed into so many Plessey systems used overseas. And, of course, all the TV companies used Plessey’s ECL dividers at fancy prices. As for political ineptitude, the UK government pretty well cut Plessey out of the European R&D process technologies programmes by insisting that UK companies wanting to get involved had to have under 250 employees.

  6. I worked for Plessey in the 80’s and the process they had (initially the 2um VB process, which stood IIRC for Vic Brown, its inventor) and the subsequent 1.4um VJ process were very good – we’d opened up a LSI Logic competitor’s chip and LSI’s metallisation was not a patch on ours – but LSI had the sales and marketing muscle which Plessey di not have, preferring to get jammy contracts from the MOD and BT, which of course dried up over time.

    And that’s the reason we don’t have a European semi industry today – the demand from system houses is supplied by the US, Korea and Japan. BT would rather buy from Huawei than a local supplier, and so they do. The government is not going to spend any money on reviving the industry, it would rather throw away £50bn on a stupid train set so satisfy some idiotic minister’s ego.

  7. That’s what happens when you hand the whole thing over to those old tarts RONA, ROCE, ROI and ROE Mike.

  8. Yes 3-4 years is more like it. It was ridiculous that we went from world leadership to nowhere in such a short time !

  9. I was wrong saying seven years Mike, Europe was more like three to four years behind the Americans and Japanese in 1984. Siemens brought out its first 64k DRAM in 1981 which was four years behind IBM and most of the American companies and three to four years behind the Japanese. But my point was that Europe, now on 28nm at Crolles, is not so far behind the leading edge in 2014 as it was in 1984 – and that point holds good. I don’t know much about Mitel’s ISO-CMOS process but I know Plessey were royally pissed off when BT told them they had to use it. Plessey thought it was piss-poor and wanted Toshiba’s process which was, at that time, the best CMOS process in the world. And Toshiba offered it to Plessey for free. Later on, of course, Siemens went to Toshiba for the CMOS process with which they eventually caught up with the world in process technology.

  10. In the mid-70s Mitel had the densest CMOS 3 micron process in the world. They didn’t keep up with the US and Japan on the 1,5 micron node but to loose 7 years in 7 years is rather unlikely.

  11. In 1983, when Siemens and Philips started talking about The Megaproject, Europe’s digital IC industry was in far worse shape than it is now, APM. In 1983, the best European semiconductor process technology was seven years behind the Americans and the Japanese. But Megaproject, JESSI and MEDEA helped Europe catch up with the best of the rest of the world. Nowadays we have 28nm in limited volume production at Crolles – so European semiconductor process technology is not nearly as backward compared to the rest of the world as it was in the early 80s.

  12. Yes David, but the world was a very different place in the 1980’s. The entire high volume digital customer base is gone in Europe and it is not coming back in our lifetimes.

  13. I think you’re probably right about the digital target price, APM, but back in 1987, the target price for the merged SGS-Thomson Semi wouldn’t have been much more. IMHO it would be a fine thing to have a cutting edge, digital semi firm in Europe. NXP and Infineon aren’t going to do it, ST is the only hope and the European politicians seem to have oodles of money to back them. Political will saved the European car industry in the 80s and resurrected IC manufacturing in the late 80s. It can be done.

  14. David, I am not sure such a split would create shareholder value, but it would indeed generate hedge fund alpha. Short French Digital ST (hopelessly squeezed between top notch US system level vendors such as QCOM and BRCM and Asian low cost suppliers such as MediaTek and Spreadtrum), hedged by going long Italian Analog ST would be a high return, low risk strategy. By the way, the target share price for French Digital ST would be zero.

  15. I agree, Don’t Agree, but in Catania? It’s not so easy.

  16. With this discussion about how ST is aimlessly circling the drain I can’t help but think of those poor Micron employees that were (will be) taken back by ST … if I were them, I would try really hard to find employment anywhere else.

  17. I think you’re right ExST, having every part of the solution in-house is a major plus – so long as the management is savvy enough and strong enough to utilise all its strengths. That depends on wise technical judgments from the top. And for that, ST needs management change. Of course the financial analysts would like to see French ST give up its digital ambitions and Italian ST feels aggrieved that ST’s shares are marked lower by the markets because of French ST’s ambitions. So there’s loads of stresses there. And to handle those stresses needs strong leadership which means change.

  18. David, This is what the current management is doing, unfortunately. And it will be bad for shareholders from both sides of the Alps. The semiconductor market, i.e. the customers, the guys who pay the bills, is not thinking in terms of Analog and Digital anymore (Italian and French, if you want to use this ST analogy). Customers think in terms of Solutions. The market is about SmartGrid, Car Entertainment, Car Safety, IoT, Wireless, DataCenters, etc… Customers (OEMs, ODMs, …) want solutions for their needs. Companies that understand this very well, like Maxim, Broadcom, Qualcomm, and many others will thrive and capture the big share of those growth markets. For sure, a commodity market will continue to exist, where TI is the super-power and there is room for a bunch of small niche players. So, splitting ST in 2 actually means going back to pre-Pistorio, pre-Merge days. Remember Thomson and SGS ? 2 sub-scale companies, loosing tons of money. Yes, this is where ST is heading and it is bad for shareholders. It is so bad that they will, of course, do something to protect their interest. They will sell bits and pieces of the company to maximise shareholder value before it’s too late. And, then, what Pistorio did will be undone, as you say. Of course, this assumes the current management stays. We’ll see.

  19. I was thinking of the Supervisory Board, Jack

  20. True Dave. Those people supported the ideas were ousted. That’s why we saw few “high level mgmt” turn over last year.

  21. Yes, Jack, I’m not sure it’s the right thing to do. I understand Pistorio is against it, but there are people on the board who support the idea.

  22. With Bozo is extending his term, most likely they would scrap the chance of splitting up.

Leave a Reply

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