Speaking of SerDes, what is your take on Alphawave SerDes? They have 224 on N3P, working on N2P, though I am not an expert to compare them with others. Any chance MTK using Alphawave IP ?
Yeah, I would love to read a piece on Alphawave. There's a lot of promise but I don't have the electrical engineering background necessary to figure out how competitive their IP is. They also don't tend to leak their customer relationships unlike some others in the space.
Well, you are probably still light years ahead of me. Well, from my readings, they have very bold claims, and seems and sounds like the IP is good and legit and very competitive. That said, one of the most shameless engineer-wanna-be-marketer-promoter owner/CEO I have seen which sort of does not bode well with my style, hence kept the position size somewhat small. Stock goes through cliff like sell offs everytime because of some silly action by the CEO. That said, I am genuinely all ears for tech/IP feedback.
Like you, what I've read suggests that their SerDes IP is competitive but it apparently divides opinion because they rely on more advanced DSP than analogue processing. They seem to be second place behind Synopsys in terms of current market share.
They're currently unique in that they offer chiplet IP, including a compute chiplet developed in partnership with ARM (which could potentially be tied to the recent ARM + Meta rumours), and a reconfigurable connectivity chiplet.
What's really exciting is that in the past two years they've transformed themselves from being a pure IP company to a vertically integrated silicon connectivity company. They're now a mini Marvell/Broadcom leveraging their IP for custom silicon and PAM4/Coherent-Lite connectivity products. The recent Siemens partnership announcement is big because now we have a leading EDA company pushing Alphawave's IP and they're jointly engaging customers from spec to silicon.
This comment is already way too long but just on the CEO: I think that's down to the fact that he's more of an engineer than a CEO so I guess he comes across awkward at times but looking at his history you've got to admit that he along with the cofounders (who own around 38% of the company between them btw) have a good track record of previous exits. There have been some significant financial blunders but its nonetheless impressive where Alphawave is now given the company is only 8 years old.
When you say "..on more advanced DSP than analogue processing", what is the read through for someone like me? Is it a smaller SAM? How is it different than Broadcom's?
We own the stock and like the growth prospect of the co, in the overall connectivity and AI chain, and 2026+ earning potential.. It could be easily multi-bagger.
I am a generalist investor in HF world, and interact a lot with management teams. and this guy is a piece of work. He is kind of trying to punch above his weight, maybe he has a shot but well he comes across weird. Why push it? If he were a down-to-earth, normal person, we would not mind owning 20%+ of this co, if not more, easily. But, not with this guy at the helm. I don't like waking up in the morning to some weird unexpected sh1t. I am already taking enough of a risk in my life and investments and it is already too much stress, we can't deal with someone's ego..
Plus, check out the expenses of the co...he and firm spends on company's dime, unreal. Does he really need to attend all those conferences and roadshows, I guess he loves traveling or does not have anything better to do..if not already fed up with engineering.
Hope it tells you how we think about it and thanks for the feedback
the CEO and his CTO have a few successful startup and exits under his belt so comes across a bit cocky seems to be inevitable. Technically, they do have good engineering teams. Getting into more lucrative custom silicon with differentiated IP is the right direction. Competition on custom silicon side though is very tough, and there are just so many hyperscaler out there.
Being 2nd place to Synopsys IP offers growth opportunities for those that want to have alternatives.
Sorry, that should have read "... on more advanced DSP rather than analogue processing". There are are a couple of approaches to SerDes: You can either have an entirely analogue architecture or you can go use an ADC/DSP approach which is what Alphawave focuses on.
DSP based architectures were historically more power hungry than analogue. However, analogue architectures are very process sensitive and therefore difficult to scale down. This means as process nodes keep advancing DSP based architectures can realise the performance and efficiency benefits associated with this. It's also difficult to build analogue based PAM4 so as data rates keep increasing analogue just becomes more and more challenging to do.
What does all this mean? Their SAM may not be as large as a Marvell/Broadcom/Synopsys who have analogue as well as DSP based SerDes but going forward the fastest growing sectors, especially AI, will almost exclusively use DSP as data rates increase.
As for how Alphawave's DSP compares with Broadcom's for example, I have no idea and need someone like OP to answer these questions for me. Again, reading about this stuff is just a hobby so please call me out on any bullsh1t I wrote above.
I'll have to take your word on the CEO, I've never met him but I can see what you're saying about "punching above his weight" but they're a small company competing against industry titans so it's somewhat understandable. I started investing in the company 18 months ago and there's been quite a few nasty surprises. I see the common factor behind these being the way in which they've been forced to pivot away from China with all the new export restrictions. Wisewave being added to the US entity list likely forced them into the equity raise just before Christmas given they were relying on that sale to shore up the balance sheet. The latest update was very good, lets see how things go moving forward, no more excuses.
Marvell 224 has a few papers as well. While indeed FoM comparisons are difficult from papers, the commercial design in historical data might provide who is bluffing more
MTK 2026 story is damn interesting with GB10, AIPC (can't be too bad with nvda name on it) and as you mentioned google TPU. Having said that expectation is sky high...and 2025 could be quite dull with all the smartphone demand being pulled forward
I’m impressed you can infer so much from a bunch of slides. I can’t tell if you’re right or not hahahhaa! Have worked on serdes before? I saw a lot of this stuff in person but it was way out there for me.
could you write about snps serders ip?
Absolutely love your work . Technically potent and funny 😄
Speaking of SerDes, what is your take on Alphawave SerDes? They have 224 on N3P, working on N2P, though I am not an expert to compare them with others. Any chance MTK using Alphawave IP ?
Yeah, I would love to read a piece on Alphawave. There's a lot of promise but I don't have the electrical engineering background necessary to figure out how competitive their IP is. They also don't tend to leak their customer relationships unlike some others in the space.
Well, you are probably still light years ahead of me. Well, from my readings, they have very bold claims, and seems and sounds like the IP is good and legit and very competitive. That said, one of the most shameless engineer-wanna-be-marketer-promoter owner/CEO I have seen which sort of does not bode well with my style, hence kept the position size somewhat small. Stock goes through cliff like sell offs everytime because of some silly action by the CEO. That said, I am genuinely all ears for tech/IP feedback.
Like you, what I've read suggests that their SerDes IP is competitive but it apparently divides opinion because they rely on more advanced DSP than analogue processing. They seem to be second place behind Synopsys in terms of current market share.
They're currently unique in that they offer chiplet IP, including a compute chiplet developed in partnership with ARM (which could potentially be tied to the recent ARM + Meta rumours), and a reconfigurable connectivity chiplet.
What's really exciting is that in the past two years they've transformed themselves from being a pure IP company to a vertically integrated silicon connectivity company. They're now a mini Marvell/Broadcom leveraging their IP for custom silicon and PAM4/Coherent-Lite connectivity products. The recent Siemens partnership announcement is big because now we have a leading EDA company pushing Alphawave's IP and they're jointly engaging customers from spec to silicon.
This comment is already way too long but just on the CEO: I think that's down to the fact that he's more of an engineer than a CEO so I guess he comes across awkward at times but looking at his history you've got to admit that he along with the cofounders (who own around 38% of the company between them btw) have a good track record of previous exits. There have been some significant financial blunders but its nonetheless impressive where Alphawave is now given the company is only 8 years old.
When you say "..on more advanced DSP than analogue processing", what is the read through for someone like me? Is it a smaller SAM? How is it different than Broadcom's?
We own the stock and like the growth prospect of the co, in the overall connectivity and AI chain, and 2026+ earning potential.. It could be easily multi-bagger.
I am a generalist investor in HF world, and interact a lot with management teams. and this guy is a piece of work. He is kind of trying to punch above his weight, maybe he has a shot but well he comes across weird. Why push it? If he were a down-to-earth, normal person, we would not mind owning 20%+ of this co, if not more, easily. But, not with this guy at the helm. I don't like waking up in the morning to some weird unexpected sh1t. I am already taking enough of a risk in my life and investments and it is already too much stress, we can't deal with someone's ego..
Plus, check out the expenses of the co...he and firm spends on company's dime, unreal. Does he really need to attend all those conferences and roadshows, I guess he loves traveling or does not have anything better to do..if not already fed up with engineering.
Hope it tells you how we think about it and thanks for the feedback
the CEO and his CTO have a few successful startup and exits under his belt so comes across a bit cocky seems to be inevitable. Technically, they do have good engineering teams. Getting into more lucrative custom silicon with differentiated IP is the right direction. Competition on custom silicon side though is very tough, and there are just so many hyperscaler out there.
Being 2nd place to Synopsys IP offers growth opportunities for those that want to have alternatives.
Sorry, that should have read "... on more advanced DSP rather than analogue processing". There are are a couple of approaches to SerDes: You can either have an entirely analogue architecture or you can go use an ADC/DSP approach which is what Alphawave focuses on.
DSP based architectures were historically more power hungry than analogue. However, analogue architectures are very process sensitive and therefore difficult to scale down. This means as process nodes keep advancing DSP based architectures can realise the performance and efficiency benefits associated with this. It's also difficult to build analogue based PAM4 so as data rates keep increasing analogue just becomes more and more challenging to do.
What does all this mean? Their SAM may not be as large as a Marvell/Broadcom/Synopsys who have analogue as well as DSP based SerDes but going forward the fastest growing sectors, especially AI, will almost exclusively use DSP as data rates increase.
As for how Alphawave's DSP compares with Broadcom's for example, I have no idea and need someone like OP to answer these questions for me. Again, reading about this stuff is just a hobby so please call me out on any bullsh1t I wrote above.
I'll have to take your word on the CEO, I've never met him but I can see what you're saying about "punching above his weight" but they're a small company competing against industry titans so it's somewhat understandable. I started investing in the company 18 months ago and there's been quite a few nasty surprises. I see the common factor behind these being the way in which they've been forced to pivot away from China with all the new export restrictions. Wisewave being added to the US entity list likely forced them into the equity raise just before Christmas given they were relying on that sale to shore up the balance sheet. The latest update was very good, lets see how things go moving forward, no more excuses.
Marvell 224 has a few papers as well. While indeed FoM comparisons are difficult from papers, the commercial design in historical data might provide who is bluffing more
MTK 2026 story is damn interesting with GB10, AIPC (can't be too bad with nvda name on it) and as you mentioned google TPU. Having said that expectation is sky high...and 2025 could be quite dull with all the smartphone demand being pulled forward
I’m impressed you can infer so much from a bunch of slides. I can’t tell if you’re right or not hahahhaa! Have worked on serdes before? I saw a lot of this stuff in person but it was way out there for me.
His day job is related high speed links.
Thought you might like this as a fellow Broadcom investor: https://x.com/Jukanlosreve/status/1893657026393493843
I'm from Taiwan...just bought some MTK lol