The Softbank investment could be a trojan horse...
I doubt Softbank is making a passive investment. They likely want to push Intel to do something... There is a lot of valuable IP that could be carved out of Intel.
Opinion: To make IDM 3.0 a success, Intel must make other companies IDM
- Thread starter delong.height
- Start date
I doubt Softbank is making a passive investment. They likely want to push Intel to do something... There is a lot of valuable IP that could be carved out of Intel.
hist78
Well-known member
I agree that SoftBank is looking for more, and according to news reports, they have held talks with Intel about deeper involvement. But there are several challenges:
- Valuation Risk – With Intel’s shrinking revenue and negative free cash flow, even a “cheap” price might still be too expensive once the risks are factored in.
- Added Value – If SoftBank’s ultimate goal is to buy in bulk and sell in pieces, Intel doesn’t need SoftBank to do that. Intel can execute such restructuring on its own.
- Cash Infusion – Intel urgently needs more fresh capital to sustain its R&D and fund foundry buildouts and equipment procurement. The key question is: how much cash can SoftBank actually provide, and in what format? The answer could either attract or discourage other investors.
- Diminished Asset Value – Through the costly SCIP program, Intel’s most advanced fabs in Arizona and Ireland are now 49% owned by two private equity firms, Brookfield and Apollo. That has significantly diluted Intel’s value.
Last edited:
I started my chip design career in a fabless company (which used to have fabs) by working with two IDMs. LSI Logic and IBM Microelectronics. The relationships worked well. All of this talk about the IDM model being a fundamental problem seems like it comes from people who are merely speculating.Here is a simple response. If you split Intel the design part will not be competitive with AMD, Nvidia, Amazon, Google, etc... Intel Design will be a cow out to pasture. AMD and Nvidia have close relationships with TSMC that Intel does not have. Pat Gelsinger made that happen. Can Lip-BU change that? Maybe but you will not see the changes for years to come. Meanwhile Intel design becomes less relevant.
Apple defined what a close foundry relationship should be 15 years ago with TSMC. TSMC has since expanded that to include other exclusive to TSMC customers like AMD and Nvidia. These companies have first line input into the process and PDK development plus packaging. We now call this "Process-Design Co-Optimization". That is a big advantage!
Intel will do the same for 14A, Intel Design will be included in that inner circle of course but there must be others and that is what Lip-Bu is doing. He has the relationships to make this happen, absolutely.
Samsung is now doing the same with Tesla, Process-Design Co-Optimization. Elon Musk probably has a different name for it... Walking the fab floor?
Bottom line: The IDM model is evolving to better compete with TSMC. How long will it take? I do not know. Maybe 5-10 years but either it will happen or we will only have TSMC which is a very big risk all things considered.
I have been through restructurings before. Even if it's common equity, key investors seem to find ways to get more favorable terms and carve outs. Softbank could be setting themselves to win either way. If Intel turns around they will make money on the investment. If Intel fails, they might be able to find a way to get dibs on the valuable parts.
Like I said, I have seen such things happen many times before.
I also think some folks here are missing a key point about SoftBank's investment. SoftBank recently spent $6.5B acquiring Ampere, which is a datacenter CPU company which had an Arm architecture license. Ampere competes directly with Intel's datacenter group. So all of this speculation about SoftBank's intentions with Intel, if they don't take into account Ampere, are probably missing a significant factor.
IDM or Fabless is all about what makes economic and competitive sense.
For leading edge node I only see two companies where they have both finances and a business that could support an IDM at the leading edge, Nvidia or Apple.
But for both Nvidia and Apple they have risen to that position without the inherent advantage of being an IDM to so I see no reason why the need to now either.
Intel clearly leveraged the IDM for almost decade and quashed many but now are neither a technology, manufacturing nor product leadership and as such really no reason for them not to go the way of dodo except for the monopoly or politics of the current Fabless situation.
The foundry fabless has been a monopoly for two nodes N5 and N3 and I see no case where it is impacting innovation or business, if any it is accelerating. The only reason for the need for another manufacture is really politics that it’s not a western company.
For leading edge node I only see two companies where they have both finances and a business that could support an IDM at the leading edge, Nvidia or Apple.
But for both Nvidia and Apple they have risen to that position without the inherent advantage of being an IDM to so I see no reason why the need to now either.
Intel clearly leveraged the IDM for almost decade and quashed many but now are neither a technology, manufacturing nor product leadership and as such really no reason for them not to go the way of dodo except for the monopoly or politics of the current Fabless situation.
The foundry fabless has been a monopoly for two nodes N5 and N3 and I see no case where it is impacting innovation or business, if any it is accelerating. The only reason for the need for another manufacture is really politics that it’s not a western company.
I don't know what Gelsinger really said.
I don't even know if it's true or not, but it's just as expected... As expected, I think it's disrespectful to Gel Singer.
Intel has no such value
That's right, I'm sorry
Except for political reasons, no advanced Fab is required except for TSMC
Incompetence is not required
As NVIDIA's monopoly is good, TSMC's monopoly is very effective in developing humanity, not the semiconductor industry.
It's what humans need It's what mankind needs, the monopoly of TSMC
Other FABs are a nuisance, a cancer that interferes with the development of the semiconductor industry.
Unless you buy or invest more stocks, Softbank won't have the right to speak up.
I agree
And so did I, I've worked with/for both fabless and IDM -- but those days are gone, because the cost of developing a leading-edge process has gone up exponentially, and it's got much harder with DTCO and needs far more and varied products pushing through the fab to optimize both yield and PDK.
Before this being an IDM had advantages, you could tightly optimize the process and IP for your own designs (e.g. Intel CPUs) with things like custom rules/layouts/PDK and get PPA advantages over foundries who had to use more standard rules/PDK/IP to support multiple customers -- and it didn't cost too much and could easily be funded out of your own product revenue.
Now the boot is on the other foot -- having lots of volume and multiple designs actually helps with DTCO/PDK/yield, rules are extremely restrictive so it's far harder to do "specials" for your own designs, and the fab/process cost is enormously expensive and has to be amortized over many customers. Also chips are much more complex with a lot more IP embedded, which can only come from a large design ecosystem not in-house like an IDM.
That's not speculation, it's fact -- the world has turned away from IDMs and towards foundries, and more specifically today, TSMC. There's no going back, if anything the situation for IDMs is going to continue getting worse as nodes evolve further. No amount of enthusiasm for Intel is going to halt or reverse this process.
In theory it's still possible for a second foundry set up and run on the TSMC model to compete, even if they only get 20% of the market and TSMC get 80% -- but it'll be hard given that TSMC have 4x more money to invest on process development/fabs, and the biggest issue is how to get from here to there -- where do the customers come from who are willing to risk going with "non-TSMC" instead of TSMC, when a delayed/failed process can risk your entire business?
Probably the only way forward is for "non-TSMC" to have guaranteed backing and/or business from a big government -- for example in the USA -- which is willing to act as a financial backstop/guarantor and push business towards it, either by subsidies or tariffs against TSMC or tax breaks which encourage companies to use "non-TSMC". But this would be enormously expensive and need huge commitment over many years, and I'm not sure even Trump is willing to do this... :-(
tooLongInEDA
Moderator
The other point to add to this is the effect of the huge ecosystem around TSMC reinforcing its established advantage. Any IP or EDA company has to make a decision which foundry to support first for any new technology node, major tool release or new IP. And that's always going to be TSMC while it has > 50% market share and is the clear leader. In theory, they could co-develop IFS and Samsung options in parallel with TSMC, but they don't have unlimited resources and time and TSMC is the high reward, low risk option. It will take a lot to shift this equilibrium.
Even though it is an IDM, it may have changed if we had gathered customers from the outside as early as possible and made chips of the customer in our factory as early as possible.
I just made the wrong choice...
If IDM does something like a foundry at an early stage...
In short, it's a hybrid, it fits the times