Intel is obviously in a bad place right now, with the stock down 30% since its earnings call. The panic has clearly set in. We have been fielding all sorts of questions which mostly boil down to “Will Intel survive?”.
As we said last week, we think it is very unlikely that Intel goes out of business, but we understand the concern. We have been very cautious about Intel for years. Businesses with heavy fixed costs can collapse surprisingly quickly. Nokia’s handset business comes to mind, and Intel’s fixed costs are much higher than Nokia’s were at the time. That being said, Intel is in a very different position.
For one thing, the market for Intel’s core CPU business is not disappearing. It is shrinking, and Intel is losing share, but the world will still need a lot of CPUs far into our brilliant AI future. Second, Intel’s manufacturing operations are worth something. Even if they cannot reach 5 nodes in 4 years and fail to get to 14A, that capacity is still valuable. And if 18A and 14A do work, then those operations are worth a lot more. A very crude comparison – Global Foundries is stuck at 16nm and older technologies. They did $8 billion in revenue last year and currently have a $24 billion market cap. Intel Foundry did around $17 billion in revenue last year (yes, all with their captive internal customer), and Intel’s current market cap is $85 billion. Of course we are getting into dangerous sum-of-the-parts valuation territory, our point is just that there is still a lot of value in Intel.
The question for Intel really boils down to two things – can they get their manufacturing process working, economically, at scale? and who will pay for all of that? Even with last week’s cost cutting, the company is still spending around $24 billion over the next years in capex, net of any subsidies. With the scale of their current earnings shortfall, they can not fund all that investment with internal cashflows, at least not comfortably. At this stage, any company in this situation would likely be in play with buyers circling. But we think all the obvious players took a look a year ago and passed when they realized the size of the funding hole the company faced.
A few months back we semi-jokingly suggested that Broadcom could buy Intel. As a potential customer for Intel Foundry, they have an inside view on the health of Intel’s manufacturing process and will know Intel’s true value long before the public markets. We received a lot of feedback on that post, much along the lines of “the regulators will never let that happen”. And while that is certainly true, we think there is another similar, but much more practical option available to Intel.
They need to sell a chunk of the company to a consortium of outsiders, and those outsiders would have to be a host of major chip companies – all potential Intel customers. Below a certain threshold – say 20% – the anti-trust scrutiny would be fairly minor. This would have a number of benefits. First and foremost – the company could raise the cash it needs to fund the upgrade of its manufacturing operations. Second, it would present a very strong validation of the company’s prospects, especially if some of those investors could simultaneously announce they would be customers. Third, if one of those companies ended up with a board seat it would essentially double the number of non-executive board members with semiconductor expertise (which is a post for another day). There is also precedent for this. Fifteen years ago, a consortium of customers invested in ASML to fund their completion of the development of EUV systems. That turned out pretty well for everyone, although we should note the irony that Intel led that consortium but ultimately decided to delay their own purchase of EUV. Current shareholders will not like the dilution that such a deal would require, but will probably be assuaged by the bump Intel’s stuck would get by the announcement.
We hope that Intel management is currently contemplating a transaction like this. It would solve a lot of problems for them.
Leave a Reply