Kuzmik_A/iStock via Getty Images
onsemi (NASDAQ:ON) (formerly “ON Semiconductor”) is rising quickly up my list of go-to examples when I say that successful turnarounds can go much further than the market initially thinks. CEO Hassane El-Khoury has quickly established credibility for the company’s margin-improvement efforts, including exiting less-efficient fabs and walking away from low-margin business, and it lends more credibility to a long-term growth story built on rapidly-growing demand for advanced power and sensing/imaging chips across a range of auto, industrial, and compute markets.
I’ve liked onsemi for a while, and I still do even after a 40% run since my last update. Along with Infineon (OTCQX:IFNNY) and STMicro (STM), I expect onsemi to ride strong demand in power and sensing to high single-digit long-term revenue growth, while company-specific drivers lead to substantially higher margins and free cash flow generation. With updated expectations, I think these shares have near-term upside well into the $70’s, though I do see more short-term cyclical risk here.
The Business Is Still Cyclical … Will 2023 Be A Pothole, A Sinkhole, Or Nothing?
Lead-times are still high across the industry (45 weeks for onsemi last quarter), and companies continue to boost their capacity addition plans. Inventories are also starting to slowly creep up though, and I think we’ll see supply catching up in 2023.
What happens then is where things get interesting. Given how desperate companies are for chips, I believe there is double-ordering now despite what many semiconductor management teams are saying and steps meant to mitigate it (including cancellation penalties and long-term supply agreements). That will stop when inventories are rebuilt and those lead-time numbers will shrink quickly.
This is nothing new, but these cyclical downturns usually lead to a few quarters of weaker top-line numbers and meaningful pressure on margins as pricing power goes away and companies see reduced overhead absorption.
Specific to onsemi, though, there are ongoing rumors that they’ve been one of the more aggressive companies on price. While management pointed out that almost all of the sequential growth in the fourth quarter was from volume, not price, a year-over-year comparison wasn’t provided and supposedly pricing has been particularly robust (or aggressive, depending on what side of the fence you stand) in areas like image sensors.
This all could be a setup for a rougher cyclical correction at onsemi than in years past. That said, there’s legitimate growth in underlying demand for advanced power and sensing products in industrial, auto, and power markets, and the company’s margin-improvement efforts have seen it exit some businesses that have historically been more cyclical. All told, I’d call this a “yellow flag” issue – it’s absolutely worth monitoring (and a lot of management’s credibility will be riding on it), but I think onsemi can navigate shrinking lead-times without too much disruption given how capacity is evolving in higher-end power and imaging chips.
Powering Up And Powering Through
The power business (MOSFETs, IGBTs, and so on) remains core to onsemi’s growth opportunities. The opportunity in EVs is likely pretty well understood by most investors, as electric vehicles will contain considerably more power management chips to handle power flows from the charger to the battery and from the battery to the motors and other onboard systems (AC/DC inverters, DC-DC converters, onboard chargers and so on). In pure battery EVs, power chip content could easily surpass $450 per unit, and there will likewise be extensive demand for power management in the charging stations that will be required to support electrification of autos.
What may be less well-appreciated are the opportunities in end-markets like industrial and computing. The rapid adoption of automation across a range of industries, including warehouses, is driving more complex and demanding electrification needs, which in turn drives demand for power management semiconductors. Moreover, as more grids turn to renewable energy sources, and as more industrial companies look to micro-grids to meet their power and ESG needs, there will be additional demand. In computing, there are not only server power management needs to consider (centered around CPUs and GPUs), but increasingly complex power needs for data centers as well.
Imaging Will Play Its Part Too
As with the power semiconductor opportunity, I feel like the imaging sensor opportunity for onsemi in autos is pretty well understood, as imaging sensors will be important enabling technology for driver assistance and autonomy modules, as well as in-car human/machine interfaces (touch/gesture-based control screens and so on). While Level 4 automation in autos is a long way off (and may never really be practical), advanced Level 2 systems still require over $100 in incremental content.
Like with power, there are meaningful opportunities in imaging in the industrial market. In addition to machine vision applications (often used in quality control or to direct other automated systems), sensor content is rapidly increasing in robots as developers look to design systems that can be safely deployed alongside human workers. To this end, I’d note that sales of imaging chips to industrial customers rose 43% in the fourth quarter – outgrowing the auto imaging business for the quarter.
Margin Improvements Are Ahead Of Schedule
There’s little question in my mind that this extreme supply crunch has accelerated onsemi’s margin improvement plans. With strong demand for more advanced power and sensing products, the company has had the luxury of being more aggressive on turning away lower-margin business, including $170M of business turndowns in FY’21 with gross margin in the neighborhood of 20%.
Management has also been moving forward on its fab closure/sale plans, recently announcing the sale of a facility in Belgium as well as the sale of a facility in Maine to Diodes (DIOD). Once complete, these closures/exits could add close to 200bp to gross margin on top of product pruning (around 100bp), and the latter may have upside as well, as I believe onsemi has over $600M of sub-40% gross margin business that could be exited over time. (as an aside, companies like Diodes may be worth a look as a potential beneficiary).
Then there’s still the move to 300mm wafers to consider. That should add another 100bp of margin leverage, supporting management’s guide to 48% to 50% gross margin in 2025. Provided that EV launch plans stay on target and the company can effectively ramp its 300mm capacity and silicon carbide efforts over the next three years, I think there’s upside to that 50% number.
The Outlook
In my last article on onsemi, I wrote that I saw long-term revenue potential in excess of $10B as part of my base-case. That “long-term” seems to be getting pulled forward, and I think the company will hit it three years earlier than my prior projection. I do see some risks to sell-side estimates in 2023, and maybe into 2024, but I expect roughly 10% annualized growth from FY’21 to FY’26 with long-term growth in the 8% to 9% range (similar to my expectations for Infineon and STMicro).
On the margin side, I think onsemi will get to 20%-plus adjusted FCF margin much sooner than before, as the outlook for gross margin and operating margin over the next few years is so much stronger now. Longer term, I’m expecting mid-20%’s adjusted FCF margins. In absolute terms, my out-year margin assumptions haven’t changed that much (200bp is significant, but not thesis-changing), but the ramp is a lot faster and steeper than I previously expected. In any case, I’m still looking for long-term FCF growth in the low-to-mid teens.
The Bottom Line
Discounting the cash flows back and using my new near-term margin assumptions, both discounted cash flow and margin-driven EV/revenue and EV/EBITDA give me a near-term fair value in the mid-$70’s, with longer-term total return potential still in the double-digits.
In a market where I haven’t found a huge number of bargains in the semi space despite some pullbacks, I’m surprised at the value I see in names like STMicro and onsemi. I do consider the risk that I’m too bullish on opportunities like auto and industrial electrification, not to mention the risk of a sharper cyclical correction whenever lead-times normalize, but given the long-term opportunities in advanced power management and sensing, I think onsemi is a name still worth considering.
Be the first to comment