ST to cut up to 2,800 jobs in major manufacturing overhaul

[Daniel Nenni]

Semiconductor manufacturer STMicroelectronics will see up to 2,800 employees leave the company globally over the next three years as part of an overhaul of its manufacturing operations and cost structure.​

The company announced the voluntary workforce reduction in a press release as it detailed a major restructuring initiative aimed at modernising facilities, boosting automation, and increasing efficiency across its manufacturing network.

The job cuts, which will occur in addition to normal staff turnover, are expected to mainly take place in 2026 and 2027. ST says that these will be managed exclusively through voluntary measures, and through dialogue and negotiations with employee representatives – in compliance with national regulations.

“The reshaping of our manufacturing footprint announced today will future proof our Integrated Device Manufacturer model with strategic assets in Europe and improve our ability to innovate even faster, benefitting all our stakeholders. As we focus on advanced manufacturing infrastructure and mainstream technologies, we will continue to leverage all of our existing sites and bring redefined missions for some of them to support their long-term success,” says ST CEO Jean-Marc Chery, in the press release.

STMicroelectronics' manufacturing overhaul has two main goals: to prioritise investment in advanced infrastructure like 300mm silicon and 200mm silicon carbide fabs, and to improve efficiency in existing 150mm and mature 200mm facilities. At the same time, the company will continue upgrading its operations with new technologies, including AI and automation. ST expects the changes to unlock high triple-digit million-dollar cost savings annually by the end of 2027.

Reshaping its manufacturing footprint

STMicroelectronics’ restructuring will affect nearly all of its global manufacturing sites, with a focus on scaling advanced capabilities and reallocating resources for greater efficiency.

In Agrate, Italy, the 300mm wafer fabrication plant will continue to be scaled up, aiming to become the company’s flagship high-volume site for smart power and mixed-signal technologies. Capacity is set to double to 4,000 wafers per week by 2027, with planned modular expansions increasing capacity up to 14,000 wpw, depending on market conditions. The adjacent 200mm facility will shift its focus to MEMS production.

In Crolles, France, the 300mm fab is also set for expansion, targeting a capacity of 14,000 wafers per week by 2027, with the potential to scale up to 20,000 wafers per week. ST plans to repurpose the existing 200mm fab to support Electrical Wafer Sorting high volume manufacturing and advanced packaging technologies, hosting activities not currently present in Europe. Focus will be on next-gen leading technologies including optical sensing and silicon photonics.

In Catania, Italy, ST will maintain its focus on power and wide-bandgap semiconductors. The development of the new Silicon Carbide Campus is said to be on track, with 200mm wafer production expected to begin in the fourth quarter of 2025. The site will shift resources from 150mm and Electrical Wafer Sorting activities towards 200mm silicon carbide and silicon power semiconductor production, including GaN-on-silicon.

Other facilities will also undergo optimisation

Rousset, France will remain focused on 200mm production, with additional volumes reallocated from other locations to fully utilise existing capacity.

Tours, France will remain focused on its 200mm silicon production line for select technologies. Legacy 150mm activities will be relocated, while the site retains its role as a centre of competence for GaN epitaxy. Tours will also take on a new strategic function in panel-level packaging.

Ang Mo Kio, Singapore will remain ST’s core site for high-volume mature technologies, will remain focused on 200mm silicon manufacturing and will also host our consolidated global legacy 150mm silicon capabilities.

Kirkop, Malta, ST’s European test and packaging site, will be upgraded with advanced automation to support the production of next-generation semiconductor products.

ST to cut up to 2,800 jobs in major manufacturing overhaul

Semiconductor manufacturer STMicroelectronics will see up to 2,800 employees leave the company globally over the next three years as part of an overhaul of its manufacturing operations and cost structure.

evertiq.com

 

[Barnsley]

View attachment 3027​

Semiconductor manufacturer STMicroelectronics will see up to 2,800 employees leave the company globally over the next three years as part of an overhaul of its manufacturing operations and cost structure.​

The company announced the voluntary workforce reduction in a press release as it detailed a major restructuring initiative aimed at modernising facilities, boosting automation, and increasing efficiency across its manufacturing network.

The job cuts, which will occur in addition to normal staff turnover, are expected to mainly take place in 2026 and 2027. ST says that these will be managed exclusively through voluntary measures, and through dialogue and negotiations with employee representatives – in compliance with national regulations.



STMicroelectronics' manufacturing overhaul has two main goals: to prioritise investment in advanced infrastructure like 300mm silicon and 200mm silicon carbide fabs, and to improve efficiency in existing 150mm and mature 200mm facilities. At the same time, the company will continue upgrading its operations with new technologies, including AI and automation. ST expects the changes to unlock high triple-digit million-dollar cost savings annually by the end of 2027.

Reshaping its manufacturing footprint

STMicroelectronics’ restructuring will affect nearly all of its global manufacturing sites, with a focus on scaling advanced capabilities and reallocating resources for greater efficiency.

In Agrate, Italy, the 300mm wafer fabrication plant will continue to be scaled up, aiming to become the company’s flagship high-volume site for smart power and mixed-signal technologies. Capacity is set to double to 4,000 wafers per week by 2027, with planned modular expansions increasing capacity up to 14,000 wpw, depending on market conditions. The adjacent 200mm facility will shift its focus to MEMS production.

In Crolles, France, the 300mm fab is also set for expansion, targeting a capacity of 14,000 wafers per week by 2027, with the potential to scale up to 20,000 wafers per week. ST plans to repurpose the existing 200mm fab to support Electrical Wafer Sorting high volume manufacturing and advanced packaging technologies, hosting activities not currently present in Europe. Focus will be on next-gen leading technologies including optical sensing and silicon photonics.

In Catania, Italy, ST will maintain its focus on power and wide-bandgap semiconductors. The development of the new Silicon Carbide Campus is said to be on track, with 200mm wafer production expected to begin in the fourth quarter of 2025. The site will shift resources from 150mm and Electrical Wafer Sorting activities towards 200mm silicon carbide and silicon power semiconductor production, including GaN-on-silicon.

Other facilities will also undergo optimisation

Rousset, France will remain focused on 200mm production, with additional volumes reallocated from other locations to fully utilise existing capacity.

Tours, France will remain focused on its 200mm silicon production line for select technologies. Legacy 150mm activities will be relocated, while the site retains its role as a centre of competence for GaN epitaxy. Tours will also take on a new strategic function in panel-level packaging.

Ang Mo Kio, Singapore will remain ST’s core site for high-volume mature technologies, will remain focused on 200mm silicon manufacturing and will also host our consolidated global legacy 150mm silicon capabilities.

Kirkop, Malta, ST’s European test and packaging site, will be upgraded with advanced automation to support the production of next-generation semiconductor products.

ST to cut up to 2,800 jobs in major manufacturing overhaul

Semiconductor manufacturer STMicroelectronics will see up to 2,800 employees leave the company globally over the next three years as part of an overhaul of its manufacturing operations and cost structure.

evertiq.com

I think tough times at STM.

Interesting that the new Catania FAB moving away from the 150mm Silicon Carbine line as the Ang Mo Kio [AMK] have been doing a lot of tapeout for that product of late

 

[Barnsley]

ROME : The Italian government opposes the CEO of STMicroelectronics, the economy minister said on Wednesday, as the Franco-Italian chipmaker faces a sustained downturn in its key automotive and industrial markets.

"The behavior of the Italian shareholder will be one of criticism and opposition," Economy Minister Giancarlo Giorgetti told reporters during a press conference on Italy's multi-year budget framework.

Rome is increasingly unhappy with STMicroelectronics Chief Executive Jean-Marc Chery and wants Paris to back an effort to replace him, an Italian official previously said.

The company, in which the Italian and French governments own a combined 27.5 per cent share through a holding company, forecast a 28 per cent drop in first-quarter revenue on January 30.

Giorgetti said the government's position "reflects the behavior of management itself, which sold STM shares it held the day before reporting the negative results."

The company, whose ADRs trade in the U.S., was hit with a proposed U.S. shareholder class action lawsuit last year that alleged STMicro leaders misled investors between January 25 and July 24, 2024, by failing to accurately disclose a deterioration in the company's business.

Chery and other insiders sold shares during that period, according to U.S. regulatory reports.

Italy also faces resistance over the appointment of Marcello Sala, a leading official at the economy ministry, as a member of STMicrolectronics supervisory board.

https://www.channelnewsasia.com/business/italy-says-it-opposes-ceo-stmicro-5055696

https://finance.yahoo.com/news/french-industry-minister-says-supports-183449275.html




I guess the reality kicks in during the difficult times

 

[Jert]

ST is more or less facing the similar dilemma as Intel is, albeit at much smaller scale, that is, IDM vs Fabless+Foundry. The choice is even more difficult for ST due to Europeans more stringent labor laws.

 

[nghanayem]

ST is more or less facing the similar dilemma as Intel is, albeit at much smaller scale, that is, IDM vs Fabless+Foundry. The choice is even more difficult for ST due to Europeans more stringent labor laws.

No it isn't. TI is an IDM and has FAT margins (like TSMC margins level good and WAY higher than peer firms like UMC tower or GF). Meanwhile the few fabless equivalents to TI and STM practically strugle to pull in single digit margins! As long as you can FtF! (Fill the Fabs!) IDM has always and will always be the more profitable business model then being a foundry or fabless buisness. That is a mathematical fact. STMs problem is manufacturing technology. TI has a bigger 300mm footprint and doesn't bother with this 150mm nonsense. TIs new 300mm fabs vs the old 200mm STM fabs, all else being equal, have die costs that are 1/2 to 1/3rd due to the extra dies per wafer and greater automation. To make matters even worse those shinny new TI fabs will be using newer higher throughput and more reliable machines too as 200mm equipment options are very limited. For 150mm we are talking a 6x die cost increase vs 300mm only accounting for the fewer dies per wafer. And that 6" equipment really is crap from a process capability, cleanliness, variation, throughput, and reliability perspective. You also literally can't even get parts for this stuff. If you told me that an identical power IC at TI Richardson cost actually 1/10th the cost of the same chip made at a 150mm fab, I wouldn't bat an eye for even a picosecond. In a downclycle like that part of the market is in right now those with a more expensive product will struggle to move units and will suffer drastic profitablity reductions. Meanwhile those with strong cost structures will pull through less battered than the rest. The results of the financials speak for themselves on this one. This fabless vs IDM myth, is just that. IDMs no longer have any magical or unique design cooptimization advantage, and being fabless doesn't make you inherently more profitable or more innovative than IDM. All of it is hogwash. But semiconductor manufacturing nay manufacturing fundamentals in general are universal. Continuously improve. Always adopt the newest manufacturing techniques/technologies. Otherwise you WILL become cost ineffective and die.

 

[osnium]

No it isn't. TI is an IDM and has FAT margins (like TSMC margins level good and WAY higher than peer firms like UMC tower or GF). Meanwhile the few fabless equivalents to TI and STM practically strugle to pull in single digit margins! As long as you can FtF! (Fill the Fabs!) IDM has always and will always be the more profitable business model then being a foundry or fabless buisness. That is a mathematical fact. STMs problem is manufacturing technology. TI has a bigger 300mm footprint and doesn't bother with this 150mm nonsense. TIs new 300mm fabs vs the old 200mm STM fabs, all else being equal, have die costs that are 1/2 to 1/3rd due to the extra dies per wafer and greater automation. To make matters even worse those shinny new TI fabs will be using newer higher throughput and more reliable machines too as 200mm equipment options are very limited. For 150mm we are talking a 6x die cost increase vs 300mm only accounting for the fewer dies per wafer. And that 6" equipment really is crap from a process capability, cleanliness, variation, throughput, and reliability perspective. You also literally can't even get parts for this stuff. If you told me that an identical power IC at TI Richardson cost actually 1/10th the cost of the same chip made at a 150mm fab, I wouldn't bat an eye for even a picosecond. In a downclycle like that part of the market is in right now those with a more expensive product will struggle to move units and will suffer drastic profitablity reductions. Meanwhile those with strong cost structures will pull through less battered than the rest. The results of the financials speak for themselves on this one. This fabless vs IDM myth, is just that. IDMs no longer have any magical or unique design cooptimization advantage, and being fabless doesn't make you inherently more profitable or more innovative than IDM. All of it is hogwash. But semiconductor manufacturing nay manufacturing fundamentals in general are universal. Continuously improve. Always adopt the newest manufacturing techniques/technologies. Otherwise you WILL become cost ineffective and die.

You are right about the merits of 300 mm fabs vs 150 mm fabs from an automation and tool quality perspective. One downside of going to 300 mm all the way is the ability to launch new product lines on new materials. For example, SiC technology first started on 4 and 6 inch wafers due to wafer growth challenges. GaN on Si at TSMC has started on 6 inch wafers. If you go 8 and 12 inch exclusively, you won’t be able to be the first mover on new materials, which are increasingly important for analog, power and mixed signal.

 

[meo9725]

Frankly, ST is an afterthought, especially with their French and Italian government oversight. No regrets here.

 

[Paul2]

No it isn't. TI is an IDM and has FAT margins (like TSMC margins level good and WAY higher than peer firms like UMC tower or GF). Meanwhile the few fabless equivalents to TI and STM practically strugle to pull in single digit margins! As long as you can FtF! (Fill the Fabs!) IDM has always and will always be the more profitable business model then being a foundry or fabless buisness. That is a mathematical fact. STMs problem is manufacturing technology. TI has a bigger 300mm footprint and doesn't bother with this 150mm nonsense. TIs new 300mm fabs vs the old 200mm STM fabs, all else being equal, have die costs that are 1/2 to 1/3rd due to the extra dies per wafer and greater automation. To make matters even worse those shinny new TI fabs will be using newer higher throughput and more reliable machines too as 200mm equipment options are very limited. For 150mm we are talking a 6x die cost increase vs 300mm only accounting for the fewer dies per wafer. And that 6" equipment really is crap from a process capability, cleanliness, variation, throughput, and reliability perspective. You also literally can't even get parts for this stuff. If you told me that an identical power IC at TI Richardson cost actually 1/10th the cost of the same chip made at a 150mm fab, I wouldn't bat an eye for even a picosecond. In a downclycle like that part of the market is in right now those with a more expensive product will struggle to move units and will suffer drastic profitablity reductions. Meanwhile those with strong cost structures will pull through less battered than the rest. The results of the financials speak for themselves on this one. This fabless vs IDM myth, is just that. IDMs no longer have any magical or unique design cooptimization advantage, and being fabless doesn't make you inherently more profitable or more innovative than IDM. All of it is hogwash. But semiconductor manufacturing nay manufacturing fundamentals in general are universal. Continuously improve. Always adopt the newest manufacturing techniques/technologies. Otherwise you WILL become cost ineffective and die.

I think they only use 150mm for discreetes, and power electronics

 

[nghanayem]

You are right about the merits of 300 mm fabs vs 150 mm fabs from an automation and tool quality perspective. One downside of going to 300 mm all the way is the ability to launch new product lines on new materials. For example, SiC technology first started on 4 and 6 inch wafers due to wafer growth challenges. GaN on Si at TSMC has started on 6 inch wafers. If you go 8 and 12 inch exclusively, you won’t be able to be the first mover on new materials, which are increasingly important for analog, power and mixed signal.

I think they only use 150mm for discreetes, and power electronics

I acknowledge that. But at this point (2025), that isn't as big of a problem as it was say 10 years ago. 8" GaN and SiC have been in HVM for a little while now and based on the rapid progress from various papers I have seen, 12" GaN is literally on the cusp of commercialization. Now granted I don't think InP or InGaAs are coming to 300mm anytime soon, but those are very niche substrates even in the discretes/power electronics/defense world. At the end of the day, bulk is what 95+% percent of the market and bulk+SiGe+GaN+SiC is like 99.99% of the semiconductor industry. For STM to have been operating not just 1 but 3 150mm fabs when they don't use substrates that can't be done on 8" and will soon be done on 12" is IMO hard to make a business case for. I think even the current plan is way too reliant on 200mm capacity. What they should be doing is taking that euro chips act money to get a full 300mm build out done like TI has been doing and make sure 300mm GaN R&D is not resource constrained. Dump the 6" fabs and maybe install some of the equipment in one of the 8" fabs if some of the products lines cannot move to 200mm wafer making equipment. Then phase some of those 200mm fabs as the 300mm fabs come online and transfer the old personal over to said new fabs (starting with the oldest 200mm fabs first since they will have the highest spares cost/the least automation/the least capable equipment). It will be expensive, but it is necessary if STM wants to be relevant. As TI continues to move production to 300mm, gets closer to 95% internal Si, and once they start having 8" GaN and SiC they won't just be kicking butt in a downturn, they will start BURYING all of these guys who are clinging onto their 8" and 6" fabs alive. Even folks like OnSemi have already closed up 6" fabs and are putting all of their growth on expanding their 300mm presence (a capability they totally lacked until only recently). I've seen this, "well we already paid for this equipment 90s so we might as well keep it running since new capex is expensive" argument before. But if the TVO of your equipment and the products it produces are lower, then I don't care if your old jalopie has already paid for itself. You are losing potential profit by not using something that isn't obsolete. Now granted I when you are doing the type of stuff STM does that pressure to not spend capital and to milk the existing assets longer is more logical than when say a company who is trying to push the boundaries of "Moore's law" tries to keep milking obsolete manufacturing equipment. But even outside of semiconductors, I have seen how the chickens always come to roost sooner or later. Bear with me because it has been a long time since I saw the quote, but I believe Andy Grove (but it might have been a different prominent member of the "Fairchildren") once said something to the effect of depreciation expenses on accounting books are meaningless. When you buy your equipment, you paid for it then. The money has already been spent, it isn't really spent on diminishing payments over a 5-year period. The losses from deprecation are purely on paper. And when managers and accounts forget this fact and think about their investments in this way, it biases their judgment against obvious business opportunities for the sake of saving the company money on paper but not in practice.

 

[osnium]

I acknowledge that. But at this point (2025), that isn't as big of a problem as it was say 10 years ago. 8" GaN and SiC have been in HVM for a little while now and based on the rapid progress from various papers I have seen, 12" GaN is literally on the cusp of commercialization. Now granted I don't think InP or InGaAs are coming to 300mm anytime soon, but those are very niche substrates even in the discretes/power electronics/defense world. At the end of the day, bulk is what 95+% percent of the market and bulk+SiGe+GaN+SiC is like 99.99% of the semiconductor industry. For STM to have been operating not just 1 but 3 150mm fabs when they don't use substrates that can't be done on 8" and will soon be done on 12" is IMO hard to make a business case for. I think even the current plan is way too reliant on 200mm capacity. What they should be doing is taking that euro chips act money to get a full 300mm build out done like TI has been doing and make sure 300mm GaN R&D is not resource constrained. Dump the 6" fabs and maybe install some of the equipment in one of the 8" fabs if some of the products lines cannot move to 200mm wafer making equipment. Then phase some of those 200mm fabs as the 300mm fabs come online and transfer the old personal over to said new fabs (starting with the oldest 200mm fabs first since they will have the highest spares cost/the least automation/the least capable equipment). It will be expensive, but it is necessary if STM wants to be relevant. As TI continues to move production to 300mm, gets closer to 95% internal Si, and once they start having 8" GaN and SiC they won't just be kicking butt in a downturn, they will start BURYING all of these guys who are clinging onto their 8" and 6" fabs alive. Even folks like OnSemi have already closed up 6" fabs and are putting all of their growth on expanding their 300mm presence (a capability they totally lacked until only recently). I've seen this, "well we already paid for this equipment 90s so we might as well keep it running since new capex is expensive" argument before. But if the TVO of your equipment and the products it produces are lower, then I don't care if your old jalopie has already paid for itself. You are losing potential profit by not using something that isn't obsolete. Now granted I when you are doing the type of stuff STM does that pressure to not spend capital and to milk the existing assets longer is more logical than when say a company who is trying to push the boundaries of "Moore's law" tries to keep milking obsolete manufacturing equipment. But even outside of semiconductors, I have seen how the chickens always come to roost sooner or later. Bear with me because it has been a long time since I saw the quote, but I believe Andy Grove (but it might have been a different prominent member of the "Fairchildren") once said something to the effect of depreciation expenses on accounting books are meaningless. When you buy your equipment, you paid for it then. The money has already been spent, it isn't really spent on diminishing payments over a 5-year period. The losses from deprecation are purely on paper. And when managers and accounts forget this fact and think about their investments in this way, it biases their judgment against obvious business opportunities for the sake of saving the company money on paper but not in practice.

All valid arguments wrt to GaN and SiC today but what happens 10 years down the line when researchers have Gallium Oxide or diamond technology ready for commercialization? 12 inch heavy companies will have no choice but to engage later with newer technology.

 

[osnium]

I acknowledge that. But at this point (2025), that isn't as big of a problem as it was say 10 years ago. 8" GaN and SiC have been in HVM for a little while now and based on the rapid progress from various papers I have seen, 12" GaN is literally on the cusp of commercialization. Now granted I don't think InP or InGaAs are coming to 300mm anytime soon, but those are very niche substrates even in the discretes/power electronics/defense world. At the end of the day, bulk is what 95+% percent of the market and bulk+SiGe+GaN+SiC is like 99.99% of the semiconductor industry. For STM to have been operating not just 1 but 3 150mm fabs when they don't use substrates that can't be done on 8" and will soon be done on 12" is IMO hard to make a business case for. I think even the current plan is way too reliant on 200mm capacity. What they should be doing is taking that euro chips act money to get a full 300mm build out done like TI has been doing and make sure 300mm GaN R&D is not resource constrained. Dump the 6" fabs and maybe install some of the equipment in one of the 8" fabs if some of the products lines cannot move to 200mm wafer making equipment. Then phase some of those 200mm fabs as the 300mm fabs come online and transfer the old personal over to said new fabs (starting with the oldest 200mm fabs first since they will have the highest spares cost/the least automation/the least capable equipment). It will be expensive, but it is necessary if STM wants to be relevant. As TI continues to move production to 300mm, gets closer to 95% internal Si, and once they start having 8" GaN and SiC they won't just be kicking butt in a downturn, they will start BURYING all of these guys who are clinging onto their 8" and 6" fabs alive. Even folks like OnSemi have already closed up 6" fabs and are putting all of their growth on expanding their 300mm presence (a capability they totally lacked until only recently). I've seen this, "well we already paid for this equipment 90s so we might as well keep it running since new capex is expensive" argument before. But if the TVO of your equipment and the products it produces are lower, then I don't care if your old jalopie has already paid for itself. You are losing potential profit by not using something that isn't obsolete. Now granted I when you are doing the type of stuff STM does that pressure to not spend capital and to milk the existing assets longer is more logical than when say a company who is trying to push the boundaries of "Moore's law" tries to keep milking obsolete manufacturing equipment. But even outside of semiconductors, I have seen how the chickens always come to roost sooner or later. Bear with me because it has been a long time since I saw the quote, but I believe Andy Grove (but it might have been a different prominent member of the "Fairchildren") once said something to the effect of depreciation expenses on accounting books are meaningless. When you buy your equipment, you paid for it then. The money has already been spent, it isn't really spent on diminishing payments over a 5-year period. The losses from deprecation are purely on paper. And when managers and accounts forget this fact and think about their investments in this way, it biases their judgment against obvious business opportunities for the sake of saving the company money on paper but not in practice.

Also the EU decided that they would rather give money to TSMC in Dresden rather than invest more in domestic champions like ST and Infineon.