Oversimplifying the process of board design- very challenging to design and test the VRs under all environmental and load conditions. With that said, one of the vendors you mentioned already crossed the moat.
I'm curious if you could clarify what is/was the moat you allude to? Is there one other than the challenge of doing a really good design validating all conditions as you mention.
Something else I would like to know is, once you have done the design and a product is shipping/used without any issues, what is the point of designing it out? When you are selling a product worth 10s of thousands of dollars, how much money are you going to save with parts that are maybe hundreds of dollars max if anything? Why not use the best part and if you really need to then you can do the redesign when needed. So, there is some switching cost at the very least. As long as the part vendor knows it and keeps you happy, why bother?
Moat is how secure a design win is; a single source device has a big moat. A multi-source device does not. Power lies somewhere in the middle. Your question- " once you have done the design and a product is shipping/used without any issues, what is the point of designing it out?" If there weren't any problems, you would still get multiple sources for price leverage and continuity of supply. With that said, your assumption that there are not any problems is not the case.
What about the fact that they use their own variant of a BCD (Bipolar-CMOS-DMOS) process at third party fabs that supposedly gives them an advantage over other power/analog semi mfrs using their own older fabs?
Also what explains the 20+% rolling returns for the past 10 years (long before the AI hype)? That could indicate some kind of moat.
Yes it probably got overhyped but It has always had a high PE historically. Interesting fact that it’s the youngest company in SMH. Second youngest is NVDA! Just shows how hard it is to break into the bigger league in semiconductors.
I don’t know a whole lot about analog/power semis, but if this kind of sheep like selling continues, it sure looks tempting…
I mean it looks oversold now. MPWR products are generally better than the competition. There is a reason why MPWR publishes excellent public datasheets while Infineon does not.
Right. As far as I can find out, one of the other reasons for their superior products is this -
MPS’ supposedly proprietary "BCD Plus" process technology as implemented at standard CMOS foundries, integrates BiCMOS signal/digital transistors with an efficient DMOS power transistor. So MPS is a fabless analog/power semi mfr working with third party foundries with its own tech. Other analog mfrs might use standard process technologies on their own older process nodes but MPS works with foundries using its BCD Plus process technology in their facilities supposedly for use only by MPS.
What prevents ADI/TI from following a similar path? I'm going to discount STM which even though invented the BCD process, is a slow moving not very agile European company IMO😝But at least TI had been doing things like expanding 300mm wafer capacity. No idea how much of a "Plus" there is for MPWR. I think they use fabs like VIS which gets some of its tech from TSMC and is also ramping 300mm fabs now.
Anybody here has more knowledge about the fabrication related technical competitive dynamics in this space?
Great analysis! I get more and more the impression that this newsletter was started in part out of exasperation from reading analyses by "theory" guys who have not the experience and probably have not worked with the hardware.
Oversimplifying the process of board design- very challenging to design and test the VRs under all environmental and load conditions. With that said, one of the vendors you mentioned already crossed the moat.
I'm curious if you could clarify what is/was the moat you allude to? Is there one other than the challenge of doing a really good design validating all conditions as you mention.
Something else I would like to know is, once you have done the design and a product is shipping/used without any issues, what is the point of designing it out? When you are selling a product worth 10s of thousands of dollars, how much money are you going to save with parts that are maybe hundreds of dollars max if anything? Why not use the best part and if you really need to then you can do the redesign when needed. So, there is some switching cost at the very least. As long as the part vendor knows it and keeps you happy, why bother?
Moat is how secure a design win is; a single source device has a big moat. A multi-source device does not. Power lies somewhere in the middle. Your question- " once you have done the design and a product is shipping/used without any issues, what is the point of designing it out?" If there weren't any problems, you would still get multiple sources for price leverage and continuity of supply. With that said, your assumption that there are not any problems is not the case.
Yeah, that was the assumption. Theres all kinds of insinuations about issues without any concrete details. What exactly is the issue?
What about the fact that they use their own variant of a BCD (Bipolar-CMOS-DMOS) process at third party fabs that supposedly gives them an advantage over other power/analog semi mfrs using their own older fabs?
Also what explains the 20+% rolling returns for the past 10 years (long before the AI hype)? That could indicate some kind of moat.
Yes it probably got overhyped but It has always had a high PE historically. Interesting fact that it’s the youngest company in SMH. Second youngest is NVDA! Just shows how hard it is to break into the bigger league in semiconductors.
I don’t know a whole lot about analog/power semis, but if this kind of sheep like selling continues, it sure looks tempting…
I mean it looks oversold now. MPWR products are generally better than the competition. There is a reason why MPWR publishes excellent public datasheets while Infineon does not.
Right. As far as I can find out, one of the other reasons for their superior products is this -
MPS’ supposedly proprietary "BCD Plus" process technology as implemented at standard CMOS foundries, integrates BiCMOS signal/digital transistors with an efficient DMOS power transistor. So MPS is a fabless analog/power semi mfr working with third party foundries with its own tech. Other analog mfrs might use standard process technologies on their own older process nodes but MPS works with foundries using its BCD Plus process technology in their facilities supposedly for use only by MPS.
What prevents ADI/TI from following a similar path? I'm going to discount STM which even though invented the BCD process, is a slow moving not very agile European company IMO😝But at least TI had been doing things like expanding 300mm wafer capacity. No idea how much of a "Plus" there is for MPWR. I think they use fabs like VIS which gets some of its tech from TSMC and is also ramping 300mm fabs now.
Anybody here has more knowledge about the fabrication related technical competitive dynamics in this space?
Great analysis! I get more and more the impression that this newsletter was started in part out of exasperation from reading analyses by "theory" guys who have not the experience and probably have not worked with the hardware.
I give thanks :) !