Global Foundries reported earnings yesterday, and the Street was not happy with the results. The company disappointed with both its revenue and profitability outlook for the next quarter, and the stock is down about 9%. At one level, these results reflect the same trends every other semi company is seeing right now – weak consumer demand and high inventories, contrasting with continued strength in automotive and industrial. And like everyone else, they think things will pick up in the second half of the year. At another level, the constraints of the company’s model may be starting to show. Like everyone else in the sector, the company enjoyed several years of excess demand and constrained supply. They are now paying the price for that as the gravity of cycles reasserts itself. Other companies can point to strong secular trends supporting growth once the current doldrums are over, but what is GloFo’s next act?
Our view is that the company’s future rests in its ability to convince mixed signal analog companies to forego capacity expansion and outsource to GF instead. This is just the standard fabless model seen with most digital logic companies, but the analog business is different. Long product cycles and lower, unpredictable volumes have meant the analog companies feel they can do better with internal capacity. From a straight dollar perspective, fabless makes a lot of sense, but it is that unpredictability that matters most. If a company is running a fab and a large customer pushes out an order by a quarter or two, gross margins fall because the fab is under-utilized. If they are relying on someone else’s fabs, then they have to pay actual cash. Gross profit dollars are better than gross margin percentage points.
A big part of the GF story since IPO has been the many long term purchase agreements (LTAs) that it got customers to sign during the worst of the supply crunch. Those are starting to time out, with prices looking to, at best, stabilize in the second half of the year. Meanwhile, the large analog companies are all decreasing their capex next year, and China analog foundries are all cutting prices sharply. Perversely, this puts GF in a difficult position. It needs to get everything it can from those LTAs, but it also needs to attract new customers and new business. The company cannot continue to push for long-term commitments while also trying to get new orders in the door. Any analog company would be wary of signing off on a large deal with GF if they are worried that when times are tight they will have to pay for the privilege of doing so.
And then there is TSMC. On their earnings call it emerged that they have a lot of under-utilized capacity at 7nm, as a result of weak demand for mobile phones. While TSMC appears to be holding the line on pricing for its advanced processes, we believe they are being much more flexible with 7nm pricing, encouraging 14nm customers to upgrade to 7nm. This could be a big problem for GF which currently does a lot of business at 14nm and cannot produce at 7nm. If TSMC cared to do so, they could wreak havoc on GF. They probably have other priorities, but the risk remains and grows by the day.
Ultimately, the question for GF is can they out-operate the internal fab capacity at companies like NXP, On and Microchip. The answer to this is uncertain. GF did cut costs this year, but they will need to be incredibly cost-disciplined going forward. Compare the latest earnings releases from GF and Texas Instruments. GF’s release boasts a beautiful design with a handy infographic. Texas Instruments’ release is about as bare bones as they come. Obviously, this is a trivial anecdote, but our sense is that GF approaches cost discipline in a very different manner than TI.
Which brings us to another matter that has sparked some concern with the Street – a CFO transition. GF’s outgoing CFO took the company to IPO and managed through the pandemic years. Their new CFO comes from AWS and Snap. There are many who want to grossly stereotype and question the different cost structures of employee-centric Internet companies with decade-long timeframes and capex heavy businesses like GF.
Lastly, we are left with the question of GF’s controlling shareholder Mubadala, the sovereign wealth fund of the United Arab Emirates. They paid $23 billion for GF back in 2008. The company is worth $30 billion today, a CAGR of 1.8%. That is a better return than they could have gotten from sticking the money in a bank for that time period, but still not exciting. With interest rates now rising, how will they view their investment? Do they look to exit and call it a decade? Or do they become more active and get more hands on in day-to-day matters? Neither option looks terribly appealing to outside investors.
It is still early days. GF has a clear set of priorities and things it needs to do. They are starting from a position. Can they move quickly and efficiently enough to build on that analog future? There is time, but the clock is definitely ticking.