Can the moves that activist investor Dan Loeb wants Intel (INTC) to explore unlock significant value for the chip giant? It's complicated.
Loeb's hedge fund, Third Point LLC has reportedly taken a $1 billion stake in Intel and wants the company, whose stock was down 21% on the year before news of Loeb's move sparked a 4.9% Tuesday gain, to consider a slew of strategic changes.
Among those changes: Spinning off Intel's manufacturing operations (possibly through a JV), divesting "failed acquisitions" and taking steps to halt the loss of technical talent that has happened in recent years.
Though I think Loeb's actions are on the whole good news for Intel investors, I also think there's no easy fix for the financial Catch-22 that Intel's manufacturing stumbles -- both related to delays for its next-gen, 7nm, manufacturing process node and the challenges it's facing as it develops CPUs relying on its current 10nm node -- have placed it in.
Outsourcing more CPU production to foundries (specifically, TSMC (TSM) and/or Samsung) would reduce some of the competitive pressures Intel is facing as it deals with stiff competition from both AMD (AMD) and ARM CPU developers. Indeed, one of my 2021 tech predictions (No. 3) was that Intel would outline plans to have foundries handle some of its high-end CPU production in 2022 and 2023, as it contends with the fallout from its 7nm delay (the company has said it would share more about its outsourcing plans during its January earnings call).
But there is a large margin hit involved with outsourcing to foundries. While I think it's in Intel's interests to absorb that margin hit for certain products over the next 2-to-3 years (and perhaps longer) rather than put itself at a major competitive disadvantage, Intel's traditional 60%-plus gross margin would see a large haircut if it fully outsourced to foundries -- and that's assuming foundries can ramp up quickly enough to fully handle Intel's considerable needs.
What if Intel spun off its manufacturing operations into a separate business, and then subsequently relied on both the spun-off business and foundries for manufacturing? In that case, the margin hit would still be there, but in theory Intel could reap financial value from the equity it would have in the spun-off business, which could service both Intel and other chip developers in the coming years.
The trouble with this strategy is that it's far from clear that a spun-off Intel manufacturing unit would have a lot of success landing other clients, given that TSMC and to an extent Samsung have opened up process technology leads and Intel hasn't historically been seen as a manufacturing cost leader. One could also note here that Intel's last attempt to build a foundry business ended up falling flat.
Another option (one that Loeb seems to want Intel to consider) is for Intel to spin off its manufacturing ops into a JV that would be jointly owned with a foundry partner. In such an arrangement, the foundry partner (likely TSMC or Samsung) would make a capital investment and supply technology know-how to the JV, while Intel would supply its fabs and make purchase commitments.
However, the financial hit from such a move would be larger than the one from a standard spinoff, since the JV partner would presumably be given a large equity stake in the manufacturing business. Also, while it might be possible for Intel to strike a JV deal with Samsung, a deal with TSMC would probably be harder, both due to antitrust scrutiny (TSMC has a 50%-plus foundry share) and because TSMC might be reluctant to transfer technology to a JV that could then compete against it for business.
As for Loeb's call for Intel to divest "failed acquisitions," it's not clear which deals he's talking about. But in the event that he's talking about Intel's $16.7 billion, 2015 acquisition of FPGA developer Altera (a business that has admittedly been struggling lately), the case for unloading Intel's FPGA business is questionable given that AMD is set to buy the business' archrival (Xilinx (XLNX) ), with the goal of offering more comprehensive computing platforms in the data center and elsewhere.
And if Loeb is talking about Intel's $15.3 billion, 2017 purchase of Mobileye, the case for calling the acquisition a "failed" one is questionable, given that Mobileye (the #1 supplier of ADAS vision processors) was seeing healthy double-digit growth before COVID hit and returned to growth in Q3.
One area where I do strongly agree with Loeb: Intel needs to take steps to reduce brain drain, improve morale and otherwise make itself more appealing to A-level engineering talent. And on that note, it's worth wondering if (following the manufacturing and R&D leadership changes announced in July) additional management changes are needed, including possibly at the highest level.
Bob Swan is doubtlessly a capable exec in many respects. But unlike Lisa Su or Jensen Huang, he doesn't have an engineering background (he was Intel's CFO, and before that served as the CFO of several other companies, before being named Intel's CEO). And given where things stand today for Intel, one has to wonder if the 7nm setback would have happened, or if its recent engineering brain drain would have been as severe, under the oversight of a more technically-minded CEO.
Regardless of where one stands on that particular question, or on the question of what manufacturing strategy Intel should adopt, having an activist such as Loeb push Intel's leadership to make some tough decisions at this difficult juncture feels like something that will probably do more good than harm. And with Intel's stock having been beaten down to a rock-bottom valuation, it's hard to blame investors for bidding Intel's shares higher on news of Loeb's actions.
Just don't count on a silver bullet arriving that will quickly solve Intel's manufacturing and competitive challenges without incurring a meaningful financial hit.