There has been a lot of recent discussion about the potential for someone to acquire Intel, the perennial sticking point is what to do with the company’s fabs. With press reports that Broadcom, Qualcomm and potentially others are looking to buy Intel, it seems that most suitors only want the Product side of Intel, not the Foundry side with its seemingly bottomless pit of cash needed to keep up with Moore’s Law.
In the middle of all this we realized something – the Street currently assigns significant negative value to Intel Foundry. What if someone just acquired Intel and shut down the fabs?
Intel currently trades for 2.00x Enterprise Value (EV) to 2025 Revenue, a rough average of its peers is closer to 6.00x. By that metric Intel is worth $315 billion, as compared to the $115 billion it currently trades for. Since Intel Foundry has no revenue, Intel Product should be worth that $315 billion, meaning Intel Foundry is currently valued at negative ($200 billion).
So, shut down the fabs and Intel’s value triples…
We know that there are people, some significant shareholders, who think that is the right course. Shut down the fabs, move Product to TSMC, and move on.
Of course it is not that simple. For starters, the fabs are worth something. Intel has $103 billion of PP&E on its books, and while the actual commercial value is much lower, there is still something there. The second problem is less tangible – “no one” wants to see the US lose its only company capable of producing leading edge semis. The problem with that is defining who exactly is opposed and what are they willing to do about it.
We argued that the big semis designers (including the hyperscalers) really need an alternative to TSMC, it’s just that none of them are willing to do much about it. Every chip design company has one strategy person screaming about this, but that person is always outvoted by 1,000 purchasing, operations and finance people who make the day to day decisions. The US government definitely wants Intel to survive too, but they cannot even pass an annual budget, so not much hope from that quarter.
The latest rumored suitor for Intel is Qualcomm. Neither Intel’s nor Qualcomm’s stock moved much on this latest news, probably because most investors see any such deal as highly unlikely. Just for fun, we ran the numbers through our Merger model. If Qualcomm announced it intended to acquire Intel, the math says Qualcomm’s stock should fall by about 25% (assuming no synergies, message us if you would like the model).
A more likely contender here is Broadcom. Qualcomm lacks the experience and probably capability to handle a massive post-merger integration, but those are Broadcom’s core competency. Under this scenario, Broadcom would buy Intel, take the stock price hit, then either shut down or divest Foundry. And while they were at it, they would probably reduce Intel Product’s headcount by 50%. The unsettling thing about this thought experiment is that it looks both achievable and very appealing mathematically. It would likely be a huge win for Broadcom.
That being said, these scenarios are premised on someone buying Intel Foundry. We suspect Broadcom would be perfectly willing to shut the whole thing down, but we know they would rather find a buyer. The obvious choice is Samsung. We have mused before that the end game of all of this is for Intel to buy Samsung Foundry, or the other way around. Politically, this is probably impossible, but we are not making any predictions about that in an election year.
Another possibility is for Intel to sell Product to someone, and use those proceeds to fund Foundry. Unfortunately the stock market makes that unlikely. Would anyone buy Intel Product for more than the current market cap of Intel? How would they justify that to their own Board and shareholders? So even if Intel sold Product for $120 billion (it’s current market cap), we do not think that would be enough to right the ship at Foundry. Any sale would have to leave production agreements in place that kept Foundry’s fab full of Product chips, but any suitor would be in a hurry to move everything to TSMC to get competitive again. See AMD and Global Foundries for how long and painful that can be.
In reality, separating the two companies would take years to pull off. Although Broadcom could probably do it under a year, everyone else, especially Intel’s current Board, would want to spend hundreds of millions of dollars and centuries of auditor and lawyer hours to do it.
The irony in all this is that everyone implicitly agrees that Intel is worth more than its current share price indicates. The trick is finding a way to unlock that. This seems to be a case study in the difficult trade-offs between short-term imperatives and long-term valuation.
Or they could just shut down the fabs tomorrow and see what happens.
Leave a Reply