Intel (INTC) and Brookfield Infrastructure Partners (BIP)

Last Monday INTC announced a deal with BIP, a publicly-traded infrastructure limited partnership. The partnership’s purpose is to build and own two semiconductor fabs that INTC is in the process of constructing in Arizona.

Terms: INTC will contribute its semiconductor fabrication expertise plus everything it has done so far on the project, which it values at $14 billion+ and will own 51%; BIP brings $14 billion+ in cash to the table for its 49%.

The glitzy slides of the investor presentation explaining all this, as well as the turgid prose of the agreement itself, can both be found on the SEC Edgar site. The bottom line as I see it, though, is that instead of getting a bank loan at, say, 5%, or using its A+ credit rating to issue bonds, INTC is ceding almost half the project to BIF for the financing it needs to complete it.

What does this mean?

(highly subjective) background

In the 1990s, the semiconductor business was transformed by the emergence of ARM Holdings and Taiwan Semiconductor Manufacturing Corp. ARM makes semiconductor design tools that allow small groups of semiconductor design engineers to create chips without having to work for the industry behemoths, like INTC or Samsung. TSMC makes these designs into functioning chips in its fabs. This industry development meant that, for the first time, INTC had serious competition for its we-do-everything model.

Though INTC chips are bigger and throw off more heat than ARM/TSMC products, the company kept itself competitive by the ability of its fabs to stay a couple of years ahead of TSMC in the race to “shrink” chips, that is, to etch chip structure on ever smaller pieces of silicon.

Three or four years ago, that changed, however. TSMC pulled even with INTC in fab sophistication …and then moved clearly ahead. The stock market picked this up right away. From the beginning of 2019 until last Friday, INTC shares have lost about 30% vs. a gain of 411% by long-time rival AMD (a user of TSMC fabs).

I find INTC’s response to its less favorable positioning to be curious. Yes, there is a several billion dollar per year increase in capex. But that’s dwarfed by $40 billion in stock buybacks during 2018-20. And there’s the steady outflow of $5 billion+ a year in cash from the dividend INTC chooses to pay.

back to the Brookfield project

I don’t know enough any more about the ins and outs of INTC to be more specific, but generally speaking I see three possibilities–not mutually exclusive–for INTC’s decision to find outside capital to complete the Arizona project rather than use its own money:

–the project itself is a clunker, at the very least in the sense that it will offer lower returns than anticipated. So it would make sense to turn to someone like Brookfield, which does things like building toll roads for cash-strapped municipalities, to get the capital to complete it

–INTC finally realizes how far behind ARM/TSMC it has fallen over a decade of complacency (apparently Only the Paranoid Survive is no longer required reading for board members) and now has a sense of urgency in trying to catch up. Not a great sense of urgency–it is still paying out about $6 billion in annual dividends, dulling the positive impact of Brookfield on cash flow–but still more than two years ago

–a variation on the first. It may be that INTC no longer has the easy access to inexpensive borrowing that its pristine credit rating would suggest. It may also be that directors would balk at a stock offering today, thinking (incorrectly, in my view) that this would make their buyback decisions of past year look even worse

the stock

To me, INTC feels like DIS as it stagnated under Eisner in his last decade. For DIS, of course, what followed was the spectacular rebound under Iger. Whether Ishrak is the new Iger or not, and therefore whether INTC is a classic value stock today, I don’t know. But the BIP deal suggests it’s more thinkable than before.

One response

  1. Intel has issued convertible securities in the past, and the convertible market would love to have a new issue with Intel’s credit rating. Moreover, a convertible could be structured to minimize apparent dilution, and would involve a strike price considerably higher than the market price of INTC. Of course, the proceeds of a convertible would fall far short of $14 billion.

Leave a Reply

%d bloggers like this: