Could Samsung merge with Intel?

[Brady]

Yes, Moore’s Law Is Dead has some sketchy rumor track record. But, he does most certainly have sources deep in Intel and has been right about them often.

So for the purposes of this thread, could it happen? Could it be a Joint Venture for just manufacturing? Uncle Sam probably would never allow a merger or takeover by the much larger Samsung.

Interesting times for Intel. Revitalization in server, leadership in ultrabooks, abysmal results in desktop, barely a glimmer in AI and GPUs. And given Samsung’s performance in semiconductors, aside from massive cashflows from the Samsung conglomerate I can only think of bad outcomes from something like this.

 

[delong.height]

Yes, Moore’s Law Is Dead has some sketchy rumor track record. But, he does most certainly have sources deep in Intel and has been right about them often.

So for the purposes of this thread, could it happen? Could it be a Joint Venture for just manufacturing? Uncle Sam probably would never allow a merger or takeover by the much larger Samsung.

Interesting times for Intel. Revitalization in server, leadership in ultrabooks, abysmal results in desktop, barely a glimmer in AI and GPUs. And given Samsung’s performance in semiconductors, aside from massive cashflows from the Samsung conglomerate I can only think of bad outcomes from something like this.

I do believe that some form of joint will take place because Samsung lack the R&D and internal capacity to scale up while Intel lack the dollars to scale out.

Any forms of merger or takeover just takes too long and require lengthy approval.

 

[Daniel Nenni]

I’m one of the industry veterans that he speaks with. I had not heard the Apple rumor and do not think it makes sense at all. The Apple TSMC relationship is much too important for that to happen. I had heard the QCOM rumor but also do not feel it is feasible, especially with the Arm litigation that’s happening. The Samsung rumor makes much more sense but as a licensing deal where Samsung uses Intel process technology (GAA) moving forward and cuts their logic R&D and CAPEX which is a very big number.

Given the current geopolitical nonsense with China I do not see any big semiconductor acquisitions being approved so Intel will still be Intel for years to come. I do however feel an Intel pivot is coming or maybe they will do the executive shuffle because clearly the future of Intel and IDM 2.0 is in question.

The coming Intel investor call should be entertaining to say the least.

 

[DepthFirst]

The Samsung rumor makes much more sense but as a licensing deal where Samsung uses Intel process technology (GAA) moving forward and cuts their logic R&D and CAPEX which is a very big number.

Could you elaborate how you see this working? Aren’t Samsung and Intel fighting for the not TSMC market? How much does licensing make sense for Intel especially when so many already passed on 18A? I understand your point (stated elsewhere) about how the combined volume of the two combined would make for a viable leading edge fab, but I just don’t see how you get there.

 

[samwilde]

Could you elaborate how you see this working? Aren’t Samsung and Intel fighting for the not TSMC market? How much does licensing make sense for Intel especially when so many already passed on 18A? I understand your point (stated elsewhere) about how the combined volume of the two combined would make for a viable leading edge fab, but I just don’t see how you get there.

It is fairly simple. Intel does not have the money to build all the capacity they need. So they could just license their process perhaps in exchange for some capacity at Samsung's fabs, which are currently running way below the capacity they could be if their process yielded worth a damn.

 

[Brady]

It is fairly simple. Intel does not have the money to build all the capacity they need. So they could just license their process perhaps in exchange for some capacity at Samsung's fabs, which are currently running way below the capacity they could be if their process yielded worth a damn.

I hear some outspoken folks on this forum say that even if Intel had the capital, it lacks the wafer demand to fill the shells. Would Samsung even bring any customers to fill those fabs? Their own Exynos?

I think it’s most possible for Samsung to license Intel’s tech because Samsung process development appears to be in a near state of crisis. If Intel’s GAA+BSPDN is in fact executing and cost effective to become HVM in 2025, they would be years and years ahead of Samsung. And how would Samsung processes development survive as a far 3rd place? Everything I know about the leading edge economics says that’s not economically viable, even with the Mother Samsung bankrolling.

 

[Daniel Nenni]

Could you elaborate how you see this working? Aren’t Samsung and Intel fighting for the not TSMC market? How much does licensing make sense for Intel especially when so many already passed on 18A? I understand your point (stated elsewhere) about how the combined volume of the two combined would make for a viable leading edge fab, but I just don’t see how you get there.

Intel can create custom processes for other foundries such as UMC and Tower for external customers using Intel fabs, they are already doing this. For Samsung Intel can create custom GAA processes for Samsung to use internally and externally since Samsung owns the Korean ASIC market and has other customers like IBM. Samsung makes a lot of chips for their own products so it is not just the phone business. At some point in time they will need GAA.

Intel only needs to break even on their manufacturing costs with these licensing deals to make it worthwhile. Intel can be the king of the NOT TSMC business and hopefully grow the market, absolutely.

 

[BruceA]

Intel can create custom processes for other foundries such as UMC and Tower for external customers using Intel fabs, they are already doing this. For Samsung Intel can create custom GAA processes for Samsung to use internally and externally since Samsung owns the Korean ASIC market and has other customers like IBM. Samsung makes a lot of chips for their products so it is not just the phone business. At some point in time they will need GAA.

Intel only needs to break even on their manufacturing costs with these licensing deals to make it worthwhile. Intel can be the king of the NOT TSMC business and hopefully grow the market, absolutely.

“Can” I don’t think the can-will do is in their current leadership.

 

[delong.height]

Could you elaborate how you see this working? Aren’t Samsung and Intel fighting for the not TSMC market? How much does licensing make sense for Intel especially when so many already passed on 18A? I understand your point (stated elsewhere) about how the combined volume of the two combined would make for a viable leading edge fab, but I just don’t see how you get there.

Here is the thing, Intel want another foundry to build its chips because building factory is just one of the most nasty thing. And it would be awesome if the foundry, in this case Samsung, uses their PDK and their process. You don't have to have design team to deal with two PDKs. One TSMC, one Intel. There's more agility. It takes Intel 4 years to see its Arizona fabs to come online, and 2027/2028 for Ohio. That is just far out. And the original goal is to have 100 billion usd to be invested in the US. But given how US government has been slowly moving, it will never happen. We need to see what will change when Trump got onboard, but even him, he didn't elaborate on industrial policy, he just keep mention tax, tariff. But where 's the other side?

I hear some outspoken folks on this forum say that even if Intel had the capital, it lacks the wafer demand to fill the shells. Would Samsung even bring any customers to fill those fabs? Their own Exynos?

With all the money it saved up from not need to build more fab. It can attack Nvidia + AMD to win on supply just like how it usually did in the past. The outsourced capacity is still running on 18A, which for most if not all future Intel products, that's the node that they are going to be running on. So it is quite significant volume to be needed. And here's the thing, Qualcomm or Nvidia now has the motive to switch because Samsung will have a decent PDK worth trying. And for MediaTek and Qualcomm, I believe Samsung would want to pressure the two to put in some order at Samsung foundry because they control where the chips that going into samsung phone be allocated from. If Qualcomm still not come back, they can very well switch back to Exynos, and Arm will be very willing to help in this regard.

 

[Barnsley]

Intel looking to break into the Korean Shipbuilding market or maybe Samsung want to build some hospitals in the US?