Apollo acquires 49% of Intel's Fab34

[delong.height]

"I think what's important to note is, this is likely the last SCIP deal that we are contemplating. And so we did a SCIP1, we did a SCIP2. It really helps us protect the balance sheet and prosecute the strategy, during a period of time where we know we're playing catch-up and our P&L and cash flow is under-earning."

Intel desperately needs cash to save its deteriorating financial situation. For example, to continuously pay dividends while in fact it can't afford to do so. Another problem is Intel is required to match grants US, Germany, Ireland, and Israel government give to it to complete those new fabs. It's a huge amount of cash that will keep biting into Intel's financial health.

"Fab light? Asset light? I still do not get it."

One thing for sure, it's heavy in financial engineering.

It's not financial engineering, it's a pivot in how capital should be allocated, and it's a norm. I think it's a great investment strategy vs TSMC's. TSMC is going down the old path of Intel, reusing equipment that were for older node and retooling them for newer node. And I think Intel's approach will open up a pandora. TSMC is much cheaper than Intel, so what? who cares? Intel is going to spend half of the CapEx to have a same level of capacity along with government credit and support when TSMC is raising its prices thanks to Nvidia, but its American fab won't be cheap either. So I think to Pat's credit, one of the biggest hurdle (cost) to compete with TSM is gone.

And even when Intel may need to give up half of their profits on fabs. It doesn't matter because those are just wafers, there are still many steps involved to make a fully functional system. And that's where Intel shine. Whenever we talk about Intel, remember it's still not a pure-foundry company like TSMC. It's Intel foundry and Intel products. What's lost in foundry will be made up in product until the older node will make profit on a non-depreciating basis. So if they have to give up half of their profit to double their internal capacity than they could do themselves, I think it's a good strategy overall. They can pull back all their design back to in house, ultimately hurting TSM along the way, and I think Pat loves to see. Pat is a competitive guy, but clearly he doesn't want his competitor to get unfair advantage against Intel, this includes AMD, Nvidia, Apple, Qualcomm, Arm, and TSM. This means a decoupling of TSM and its strong ecosystem. Also Intel will make money at packaging and design.

 

[delong.height]

One sentence in the press release caught my eye:



Apart from the given explanation on the financial terms, could such a move be used to provide credibility to IFS´ claim to be an independent company? While Intel will still control operations, giving a 3rd party such an extensive stake might convince external customers about their intentions.

I would consider such a move the probably most risky and expensive marketing move of all time, yet I am curious to hear your opinions on the consequences for IFS reputation on the market.

I think all these SCIP will pave a way for an eventual split for IFS, where SCIP partners will get shares of the new IFS

 

[delong.height]

The biggest question is what's the return and other guarantees Intel obligated to this deal and Apollo. These information are confidential and we won't be able to find out. But one thing for sure is the true direct and indirect cost will be higher than the Intel-Brookfield Arizona deal.

In the Intel-Brookfield Arizona deal, the fab and equipment are owned by the "JV". But in the Apollo case, the new "JV" doesn't own the fab or equipment at all. Apollo's $11 billion got the "right" to manufacture Intel 3 and Intel 4 products with undisclosed amount of guaranteed profit. It doesn't own the fab and equipment at all.

The returns here doesn't matter, as long as it's positive. Intel will make back on Intel product. And that's the thesis until 2030

 

[delong.height]

Unfortunately, the Intel-Brookfield and Intel-Apollo financing deals make the Intel product division harder to compete. Because the wafer volume commitment and profit/return guaranteed by Intel, Intel product division will be forced to use the Intel fabs involved in the contracts.

Intel and Pat Gelsinger had stated that in the IDM 2.0 and Smart Capital strategies, Intel Product division is free to choose between internal fab and external foundries in order to compete. In reality, it's gone.

no they won't. At least not until they are manufacturing ready

You build some fab shells to collect US and Euro Chips Act money, while stealthily going fab lite. It has to be stealthy, so stealthy. The handouts and private financing wouldn't be there if you were not planning to operate the shells. It wouldn't be the first time an IDM built a shell in Europe and never operated there. But in the US, this is new.

this idea is so silly. You spend more money to get less money. The silliest I've ever heard

 

[hist78]

Thanks for the additional information.

I meant to say that it's unlikely that Intel has to bear an annual payment yield of over 10% on this deal. As you posted originally, you assumed 15% yield on the debt obligation. It's highly unlikely in this particular deal.


You used private equity industry's average return and Apollo's return to assume the yield required on the deal. If this deal is an equity investment, the return may come mostly from a future exit.

I'm not sure your "equity investment" meant?

Regardless Intel "pays" less or a lot for several years then whatever amount at the Private Equity's exit , they are still Intel shareholders' money.

Apollo Assets Management is one of the private equity companies who competes against each other on the investment performance and founding sources. Apollo's performance and expected return, as shown in my earlier post, can't be out of industry's norm and Apollo's own historical performance.

From Micron and Texas Instruments recent corporate bonds information, they are issued around 5% rate. Why Intel didn't use the cheaper ways, such as corporate bonds, to raise cash? The reason may not be pretty.

 

[hist78]

The returns here doesn't matter, as long as it's positive. Intel will make back on Intel product. And that's the thesis until 2030

Profit margin, especially a decent one, is very critical to the semiconductor and foundry business. From raising more capital, merge and acquisition, R&D, buying expensive EUV tools, building new fabs, competing against competitors, attracting and retaining talents, to weathering the industry downturn, a fat profit margin is a must .

 

[tooLongInEDA]

Profit margin, especially a decent one, is very critical to the semiconductor and foundry business. From raising more capital, merge and acquisition, R&D, buying expensive EUV tools, building new fabs, competing against competitors, attracting and retaining talents, to weathering the industry downturn, a fat profit margin is a must .

Quite. "Returns don't matter" !!!

More seriously, I was reflecting the other day that Intel is in danger of changing from a company that did one thing exceptionally well (80%+ share in high end CPUs) to one where they compete in many markets, but are no longer #1 in any of them. Currently << #2 in GPUs, foundry, AI, bailed on memories (Optane, NAND), selling off #2 position in FPGA (Altera), losing share in data centre, up against resurgent AMD in PCs. It's not impossible they lose their #1 position in CPUs from here (or that business shrinks further relative to alternative CPU architectures).

Intel is still paying dividends. And (I think) only stopped share buybacks (one report suggests this totalled $152bn over 35 years) in 2022. And yet they "need" government subsidies and these complex PE deals.

MKW commented the other day that Bob Swan started the TSMC outsourcing program and Pat Gelsinger reversed it. I'm starting to get the sense that Bob Swan was correct and that Intel should have gone fabless. And that US taxpayers are on a hiding to nothing with their CHIPS program funding here.

 

[siliconbruh999]

Quite. "Returns don't matter" !!!

More seriously, I was reflecting the other day that Intel is in danger of changing from a company that did one thing exceptionally well (80%+ share in high end CPUs) to one where they compete in many markets, but are no longer #1 in any of them. Currently << #2 in GPUs, foundry, AI, bailed on memories (Optane, NAND), selling off #2 position in FPGA (Altera), losing share in data centre, up against resurgent AMD in PCs. It's not impossible they lose their #1 position in CPUs from here (or that business shrinks further relative to alternative CPU architectures).

Intel is still paying dividends. And (I think) only stopped share buybacks (one report suggests this totalled $152bn over 35 years) in 2022. And yet they "need" government subsidies and these complex PE deals.

MKW commented the other day that Bob Swan started the TSMC outsourcing program and Pat Gelsinger reversed it. I'm starting to get the sense that Bob Swan was correct and that Intel should have gone fabless. And that US taxpayers are on a hiding to nothing with their CHIPS program funding here.

I think that was the reason he was outed as well intel has been a manufacturing company it's in their DNA they need to adapt with time though which they failed and their biggest problem waiting for 18A and LNL ARL to see what they have in store and it will define their position

 

[bilau]

Quite. "Returns don't matter" !!!

More seriously, I was reflecting the other day that Intel is in danger of changing from a company that did one thing exceptionally well (80%+ share in high end CPUs) to one where they compete in many markets, but are no longer #1 in any of them. Currently << #2 in GPUs, foundry, AI, bailed on memories (Optane, NAND), selling off #2 position in FPGA (Altera), losing share in data centre, up against resurgent AMD in PCs. It's not impossible they lose their #1 position in CPUs from here (or that business shrinks further relative to alternative CPU architectures).

Intel is still paying dividends. And (I think) only stopped share buybacks (one report suggests this totalled $152bn over 35 years) in 2022. And yet they "need" government subsidies and these complex PE deals.

MKW commented the other day that Bob Swan started the TSMC outsourcing program and Pat Gelsinger reversed it. I'm starting to get the sense that Bob Swan was correct and that Intel should have gone fabless. And that US taxpayers are on a hiding to nothing with their CHIPS program funding here.

delong.height did not say "Returns don't matter." He said: "The returns here don't matter, as long as it's positive. Intel will make back on Intel product. And that's the thesis until 2030"


Having to do these PE deals is obviously far from ideal, and it is exactly as what you had said -- they now have to compete without the WinTel hegemony. But to go fabless three years ago would have meant selling off (realistically still) top -tier foundry assets at fire sale prices (because the foundry is very Intel specific and anyone else taking over will need to pour in many billions.)

This separate accounting thing obviously opens up the possibility of a spin-off, but I tend to believe that isn't PG's preference. I think they do this separate accounting (very) primarily just to highlight out the cost centers so that they can fix them.

Intel has now cleaned up the 10nm mess (Alder Lake), and has shown me they can walk again with MTL (I4)and SF(I3). But these are all hors d'oeuvres, the main course, A20 and then A18, will be served soon enough and until then, I am willing to cut PG some slack.

 

[MKWVentures]

Intel has now cleaned up the 10nm mess (Alder Lake), and has shown me they can walk again with MTL (I4)and SF(I3). But these are all hors d'oeuvres, the main course, A20 and then A18, will be served soon enough and until then, I am willing to cut PG some slack.

This SCIP is needed for Intel to grow. they board will not allow the negative cash flow to continue. Its a good deal for Intel.... A company that is struggling but has plans and wants to grow big. this is how it is done.

Will we still cut Intel slack if 20A and 18A don't ramp into a production factory by end of 2025? if they are running only 12K wafers per month in dec 2025? what if 5 of 6 announced factories are delayed in 2025.

 

[jiakun0618]

I'm not sure your "equity investment" meant?

Regardless Intel "pays" less or a lot for several years then whatever amount at the Private Equity's exit , they are still Intel shareholders' money.

Apollo Assets Management is one of the private equity companies who competes against each other on the investment performance and founding sources. Apollo's performance and expected return, as shown in my earlier post, can't be out of industry's norm and Apollo's own historical performance.

From Micron and Texas Instruments recent corporate bonds information, they are issued around 5% rate. Why Intel didn't use the cheaper ways, such as corporate bonds, to raise cash? The reason may not be pretty.

Equity and debt are treated differently in book and have different natures.

Refer to the Fitch report for some clue:
https://www.fitchratings.com/resear...ok-unaffected-by-scip2-transaction-04-06-2024

Fitch’s equity-like view on SCIP2 mainly reflects the variability of investment returns for Apollo, which is subject to the JV’s ability to meet Intel’s demand for leading-edge wafers, market demand for Intel’s products and Intel’s success attracting third-party foundry customers. Fitch treats the first SCIP transaction as debt-like, given Fitch’s view that wafer purchases and returns to the minority investor are largely fixed. Fitch adjusts Intel’s debt quantum by net flows from the SCIP1 minority investor.

The JV will have rights to manufacture wafers at Fab 34 using Intel 4 and 3 process technologies to support demand for Intel’s products and provide capacity for foundry customers. Under the agreement, Intel is required to complete construction on the facility, although high volume manufacturing in Fab 34 already began in September 2023. The JV will sell wafers to Intel on a cost-plus-margin basis and the agreement requires Intel to meet volume commitments for its demand following the facility’s substantial completion.

Intel has agreed to preferential loading of Fab 34 relative to other Intel fabs running the same process and tooling and payment of liquidated damages to Apollo in various circumstances if Intel fails to meet minimum volume commitments, which Fitch views as typical of wafer supply agreements. In addition, under various ancillary Intel and Apollo agreements, Intel or Intel Ireland may be required to pay liquidated damages or termination payments under various scenarios, including under-performance or termination, an amount up to 140% of Apollo’s purchase price. Intel will provide a parent guarantee to the obligations of Intel Ireland, which owns the Fab 34 assets.

Even the deal is structured more as an equity investment, there are embedded terms for a debt-like guarantee. So it's smart financial engineering to use third party investments while avoiding further lowering credit rating.

Intel's decision to avoid more debt is simply because they cannot afford more. In the past 8 quarters, they were losing money for 4 of them. Debt leverage ratio, as measured as gross debt/EBIDTA, was over 3.7x at the end of 2023 (1.8x on net debt). They have over $50 billion gross debt. Additional debts will push credit ratings lower, making them more expensive.

Failing to turn around in the second half of this year will push the company down further. Based on the current competition in the client and data center markets, it's going to be an uphill battle for Intel. All kinds of scenarios are possible for the company.