There’s no point in sugar-coating it – if you like “gee-whiz” cutting-edge technology, Diodes (NASDAQ:DIOD) is not going to be the stock for you. Diodes focuses on lower-end, low-ASP/low-margin, mass-market analog and discrete semiconductors that fulfill a range of functions (power management, switching, timing, et al) in a broad range of end-markets like autos, communications, computing, consumer electronics, and industrial. It’s a market that, broadly speaking, companies like Analog Devices (ADI), ON Semi (ON), STMicroelectronics (STM) and Texas Instruments (TXN) have been scaling out of in recent years.
And yet, the products are still needed and there’s still money to be made here. Diodes has shown some agility when it comes to acquiring other companies (including the well-timed acquisition of Lite-On, which added much-needed capacity at the right time), and I like the fact that most of the company’s products have multiple use-cases across end-markets.
There are drawbacks here. Diodes is never going to earn the sort of margins that more specialized chip companies can, and the long-term track record for adjusted free cash flow margin (high single-digits) is not exciting. Still, there has been a fair bit of growth here over the years, and if Diodes can continue to grow at a mid-single-digit pace (a little bit above my expectation of underlying semiconductor volume growth), the shares have some appeal here.
Increased Pressures Evident In Fourth Quarter Results
Diodes isn’t immune to the challenges facing the semiconductor industry, and with about 60% of the company’s revenue coming from more pressured markets like consumer, computing, and communications, that end-market stress is showing in the financials.
Fourth quarter revenue rose 3% year over year and declined 5% sequentially, coming in just ahead of sell-side expectations. Gross margin declined 20bp qoq on a non-GAAP basis (to 41.6%), beating by 60bp, while operating income declined 14% qoq, beating by 10%, with margin down 220bp to 20.2% (a 170bp beat).
With inventory corrections underway at many customers, company inventory increased about 4 days from the September quarter to 117 days.
As Semis Go, So Goes Diodes
The good and bad news of the Diodes model is that they don’t drive the bus; virtually all of the company’s products are ancillary to other chips, and so they “ride along” with the underlying demand across their addressed markets.
Said differently, wherever there’s a SiC IGBT, a high-end GPU, or a high-end switching processor, there will be Diode MOSFETs, diodes, rectifiers, timing components, and switches around it. Likewise, NVIDIA (NVDA) engineers are not burning the midnight oil to figure out how to develop a better version of the products that Diodes manufacturers. I don’t mean this as an insult toward Diodes or its products; they’re very necessary, but they’re not high-value components that attract a lot of competitive R&D.
I see a lot of leverage for Diodes tied to the increased chip content in autos as well as many industrial markets.
Auto electrification will drive meaningful demand growth for a wide range of Diodes products, and electrification doesn’t just mean the replacement of a gas-burning engine with a battery. Many functions that used to be handled manually (like moving your seat or rolling down a window) or through hydraulics are now electrified with motors doing the work – requiring motor control chips, power management chips, and so on. Likewise, increasingly sophisticated ADAS systems, increased car connectivity, and increased use of LED lighting drive more demand for Diodes.
All of this means a roughly 6-fold increase in addressable content per car relative to 2013 levels. Take a single example provided by the company – a USB charging feature in a car drives more than $8 in content for Diodes between power management, MOSFETs, bipolar junction transistors, diodes, rectifiers, timing, and connectivity products. Likewise, LED lighting is a roughly $14/car opportunity, and not that many cars have that yet.
Broadly similar themes are evident in end-markets like industrial, compute, and consumer. As factory automation advances, there will be increased demand for motor control subsystems and advanced power supplies, let alone future demand for industrial IoT packages for monitoring and machine control. In compute, timing and connectivity are still essential functions, and this is a market where many of Diodes competitors have exited, and consumer IoT (smart home, wearables, etc.) likewise offer growth in power management, timing, and other components.
The Outlook
I don’t know whether this rises to the level of irony or not, but I would note that Diodes is also following some of its former rivals in reprioritizing its business mix – diversifying more towards higher-value analog components like linear redrivers, packet switches, low-jitter timing products, and silicon carbide power components (the company recently announced the launch of a SiC Schotky barrier diode), and deprioritizing some lower-margin discrete components.
With those efforts, I do see some opportunity for Diodes to grow a little faster than the underlying volume growth of the semiconductor industry, and Diodes has used M&A to add to its product portfolio (like the Pericom deal in 2015). I’m comfortable with a long-term growth rate in the 5% to 6% range, and I do expect to see further select M&A here and there along the way.
On margins, the company has actually been over-earning to its target, with component shortages in 2022 supporting robust pricing. I think it will be tough to get much above the low-40%’s for gross margin, and I likewise don’t see the company moving much above the low-20%’s for operating margin. Even so, that’s a marked improvement from the mid-teens (and below) of past years, and I think Diodes is on a path where mid-teens adjusted FCF margins are attainable and sustainable, though I don’t expect a lot of FCF outgrowth relative to sales.
Between discounted cash flow and margin-driven EV/revenue and EV/EBITDA, I believe Diodes is undervalued today. Discounted FCF supports a potential annualized long-term total return of more than 10%, while I can easily get to $110 on my EV/revenue approach. In fact, using my 2023 operating margin estimate (19%), which should be the low for the cycle, and a cycle-low EV/revenue of 2.85x (semiconductor multiples typically correlate well to margins) gets me to over $120 – I actually used a 10% haircut to the multiple (2.6x) to drive that $110-plus number.
The Bottom Line
What Diodes does isn’t sexy, but I don’t need sexy from my investments. Cyclicality will always be an issue here, as will comparatively low margins, but Diodes has shown it can succeed operating in businesses where other companies no longer want to compete but where there is still meaningful volume demand. I do expect another soft quarter for March and there could be some risk to June, but even after a substantial rebound from its lows, this is a name worth further due diligence.
Be the first to comment