Axcelis Stock: A Cheap Beneficiary Of Electrification (NASDAQ:ACLS)

Editor’s note: Seeking Alpha is proud to welcome Dynamis Cap as a new contributing analyst. You can become one too! Share your best investment idea by submitting your article for review to our editors. Get published, earn money, and unlock exclusive SA Premium access. Click here to find out more »

Investment Thesis

Axcelis (NASDAQ:ACLS) is a leading provider of Ion Implant tools, the Purion line, which semiconductor manufacturers leverage for preparing wafers and altering their material properties. Poor earnings revision trends, concerns over China exposure, and fears of Silicon Carbide supply glut as Electric Vehicles Sales slowdown caused multiples to shrink since Aug-23, in contrast to the rest of the sector. At <15x LTM Earnings, a price implying no FCF growth in the next 10y, and an analyst day in July, I think the upside trumps the downside at these levels. I am buying a small position and look to increase as fundamentals turn

Company Overview

History

Since its beginning as Nova Systems in 1978, Axcelis has focused on producing implantation tools for the semiconductor industry. Today, the Purion line continues on that heritage by focusing on ion implantation, a step altering the electrical properties of semiconductor materials in the manufacturing process. The company is the 9th largest semiconductor equipment vendor based in the US by market capitalization and the only specialist in ion implanters.

What Do They Do?

Axcelis designs, manufactures, and sells ion implantation systems. Semiconductor manufacturers, such as TSMC, leverage these tools to modify the electrical properties of silicon wafers, a process called doping. By introducing dopant ions into the surface of a semiconductor wafer, one can modify its conductivity and other electrical characteristics, which helps in transistor formation. But the tools might also be leveraged to modify the inherent characteristics of the material, such as when semiconductor device manufacturers employ radiation hardening for devices used in aerospace, nuclear, or high-energy physics applications.

Customers

Axcelis customers are semiconductor chip manufacturers, with various business models. Some own design and manufacture the devices, some only act as foundries and manufacture chips for others, some do both. Customer concentration has reduced in the last three years; the top ten largest customers accounted for 52% of sales in FY23 compared to 70% in FY21, and Axcelis had no >10% of customers in FY23. Currently, the company has most exposure to “Power Devices”, a family of semiconductors used in high-power and high voltage applications, across industries ranging from consumer electronics to automotive. Axcelis refers specifically to silicon carbide (SiC) and silicon IGBT, which together represented c.60% of System shipments in 2023. Electrification provides a tailwind for these devices. The largest producers of “Power Devices” are European (Infineon (OTCQX:IFNNY), STMicro (STM)) and Japanese (Fuji, Mitsubishi, Renesas). Axcelis has also benefited from Chinese customers’ demand, both in mature nodes, with SMIC, and “Power Devices”.

46% of FY23 Rev Came From China. Axcelis has one of the highest exposures to China among US semiconductor equipment manufacturers, compared to 26-27% for Applied Materials, Lam Research and KLA. So far, the U.S. export control regulatory framework focused on advanced logic, DRAM, and NAND. They had little impact on Axcelis ability to ship systems used for mature nodes and power devices. In my opinion, U.S. export controls represent a significant risk going forward, should they focus on mature nodes and power devices, given increasing tensions between the U.S. and China.

70% Exposed to Mature Nodes. The company derives 95%+ of sales from Ion Implanters. In the F4Q23 earnings presentation, Axcelis quotes a $2.4bn market, representing less than 3% of total Wafer Fabrication Equipment (WFE) spend. Since 2012, Ion Implanters grew more slowly than WFE; they represented 3.2% of WFE in 2012 compared to 2.7% in 2023. Ion Implanter demand also shows cyclicality, as highlighted in Fig. 1. In the last two years, we estimate that Axcelis systems captured 36% of the market. In terms of end-markets, the Mature Process Technology segment accounts for 70% of the TAM (slide 10 F4Q23 presentation). In my view, Axcelis is strongest in Power Devices, with 75% share, and weakest in Advanced Logic with 6%. I believe that in “Advanced Logic” Applied Materials leverages its strength in more important process steps to “bundle” their Ion implanters.

Fig. 1: Ion Implanter Demand

Source: Company Disclosures

Fig. 2: Market Share by End-Market

Source: Company Disclosures

Financials

20-25% FCF Margin. In the last five years, the company has grown Sales at a 27% CAGR, increased gross margins from 40.1% to 43.5%, reduced Opex/Sales from 25% to 20%, and delivered an average FCF/Sales of 12%. The balance sheet is solid, with almost no debt, and $506m of cash and short-term investment representing 15% of the market capitalization. The company has provided the following guidance for 2025: 1) Sales of $1.3bn (7% CAGR), 2) >45% Gross Margin, 3) 19% Opex, 4) 26% Operating Margin, 5) 25% FCF. In 2023, cash conversion was poor relative to history, mainly due to working capital investments, and there was a large gap between OP and FCF. Fig. 3 compares Axcelis current and targeted financials relative to peers; they deliver average OPM with lower Gross Margin and lower Opex intensity. The lower opex intensity is not related to scale; they are small relative to others. Instead, they are investing less, which might be a factor of underinvestment, or of the competitive landscape they face. On FCF conversion, the company has achieved a level in-line with what they target in four of the last five years, and 2023 might be viewed as an aberration. Overall, I think 20-25% steady state FCF margin to be more likely than the 25% 2025 target.

Fig. 3: Semiconductor Equipment Economics

Source: Companies 10Ks

Investment Case: Low Price for Moat with Structural Tailwind

The SiC Tailwind… Axcelis SiC exposure was 35%, growing more than 2x in 2023. Long-term, McKinsey expects 31% CAGR to ’25, with most of the demand driven by EV adoption and higher content per EV. Dominant players’ capacity expansion plans paint a similar picture. STMicroelectronics (STM) and Infineon, with a combined 50% share of the SiC market in 2022 (Source: Infineon, slide 36), dominate. Infineon targets a 15x capacity expansion to the end of the decade with its Kulim factory, and 47% CAGR in Sales to 2030. Management committed to a 50% increase in SiC sales in FY24 during the F1Q24 Earnings Call. STM disclosed $1.14bn revenue in SiC for 2023, growing 60%, with a target to get to $2bn by 2025. At the 2022 CMD, they highlighted plans to 2x their 200mm SiC capacity between 2022 and 2025 (slide 8) and build a new SiC integrated plant in 2023. The third-largest player, Wolfspeed, targets capacity supporting 5x revenue growth between 2022 and 2027 (Source: CMD2022, slide 16), and Axcelis is shipping Purion to them. Looking ahead, I think the tailwinds for that market appear to be strong, and the short-term US EV demand risk lull does not change that.

…And the attached risks. There are, in our view, three key risks to the SiC tailwind for tool demand. First, we think EVs drive 60-80% of SiC demand to the end of the decade; any change in EV volumes or SiC content per EV poses a risk to forecasts. Second, we reckon China will drive 35-45% of overall EV SiC demand. Focusing on technological sovereignty means Chinese manufacturers, who may not want to rely on Western toolmakers, will account for an increasing portion of SiC devices. Finally, manufacturers are transitioning to large wafer sizes. If history is any guide, when that transition happened in logic this should result in lower tool spend per wafer, as it did in the logic space. Taking these risks into account, we see Axcelis shipments to SiC providers growing at 15-25% per year over the next five years.

Prospects in Other Markets Remain A Question Mark. The company shared a guidance for 50% of System Shipped Sales to come from SiC in 2024, alongside information for the other end-markets. Management does not provide System shipped sales directly, but I derive them from disclosures of “CS&I” Revenues, which is Services + Spares, and Product/Services in the P&L. Systems Revenues are Product minus Spares, and we derive Spares from subtracting Services in CS&I. I estimate $884m of Systems Revenues in FY23, or 78% of Sales. From there, we build a weighted growth rate for Systems, based on exposure in 2024, and forward growth expectations, which give us a rough idea of the growth potential for the company, which we can use in a DCF: 8% 5Y CAGR. The issue of the terminal value of the non-SiC verticals remains a valid concern, in our view, and as suggested by the company’s capital allocation guidance, where M&A to grow beyond the $1.3bn implant business model features.

Fig. 4: System growth Potential by End-Market

Source: Author estimates

Large Valuation Disconnect To SOXX. By now, concerns over EVs are all over the popular press, and Axcelis shares multiple reflect these concerns. Using P/E (LTM), a metric not relying on sell-side estimates, Axcelis shares sit at 14.2x, down from 35x in July of 2023, and nearly 1SD below the 5Y average. In contrast, the SOXX – a proxy for semiconductors – sits at a higher level than in early 2024 and at a 5-year high.

Fig. 5: Largest Multiple Disconnect in the last 5 years

Source: Koyfin

Buy-side Expectations Are Low: 0% 10Y FCF CAGR. I use a reverse DCF to further delve into expectations. The inputs are simple:

1. WACC
2. Perpetuity Growth Rate (after Y10)
3. Next Twelve Months Forward FCF
4. Current Price.

I solve for the implied FCF growth rate between Y1-10. The purpose of the exercise is to get a rough idea rather than a precise number. Changes to inputs such as the Perpetuity Growth Rate, risk-free rate or equity risk premium would modify the implied FCF growth rate. Today’s price suggests an implied FCF growth rate of 0-1% in the next 10y. It only takes 4.5% FCF growth to deliver 24% upside. Expectations are very low here, giving us a decent margin of safety.

Fig. 6: Reverse DCF Implies No FCF Growth Baked In

Source: Author Calculations

DCF and Relative Valuation Suggest >20% Upside. Running a DCF forecasts of 8% Sales Growth over the next 5Y, converging to 3.3% perpetual growth, with 11% WACC and OPM peaking at 28%, we get to $126 a share, or 20% upside. Finally, simply thinking in terms of relative valuation, and assuming no further downside to EPS, and a multiple back to the 5y average, gives us 50-100% upside from here; that is a possible scenario one year from now.

Too much risk priced in. Given the setup, with a low valuation relative to history and sector, upside on a very conservative DCF, I reckon there is an opportunity to generate high returns with a 1-3y horizon. Let’s explore the key concerns in more details: 1) Terminal Value, 2) China Reliance, 3) EV demand.

Possible Acquisition Target. It seems unlikely that management would sit idly at this valuation. The asset relies on has IP, an installed base from which it generates sales, and a key position in the market. I believe Axcelis will find a home among a larger public peer, or even privately.

Non-China Doing Well on a Standalone Basis. Management guided China to account for 40-60% of 2024 sales, from 46% in FY23. Based on geographical disclosures in the 10Ks, we can split Sales between China and non-China to study the dynamics separately. China sales achieved 36% 5Y CAGR, non-China 20%. In 2024, based on flat sales and mix guidance, non-China sales might be down YoY. That seems surprising given the capacity increases we mentioned above; it could be a source of upside, in my view. Looking ahead, that revenue stream is still likely to increase at a decent clip; the 2018 partnership with Screen in Japan will help Axcelis in a large underpenetrated market. Stripping out China entirely, an extreme case, the company is priced at 5-6x Sales, which is lower than most other front-end semi-capital tool equipment vendors.

EV Demand Risk Are Short-Term And US-focused. China represents the largest EV market in the world, accounting for 59 percent of 2023 units, according to BNEF. BNEF expects growth to slow to 21% YoY in 2024 from 35% in the previous year. EVs represented 34% of new-auto units sold in 2023, compared to 4% in 2018. Volumes can 2-3x from here, but we are entering a slower penetration phase. The government recently brought forward NEV target to 45% of auto sales by 2027, compared with 40% by 2030 set in 2020; we expect them to beat that target easily. In other regions, the transition has been slower and should pick up pace as manufacturers offer models that are more attractive to the consumer than the ICE equivalent (already the case in China).

Fig. 7: EV Adoption By Region

Source: BNEF, Nikkei, Rystar

Fig. 8: Penetration vs Volume by Regions

Source: BNEF, Nikkei, Rystar

CS&I Business underpins downside. A final way to look at the asset is to consider the value of its recurring business of CS&I, based on Spares and Services. Spares represent the main part of that revenue stream, and we estimate $222m in Sales in FY23. These depend on installed base and fab utilisation (economics activity). Private Equity firms would typically look to acquire such a revenue stream for 5x Sales, or $1bn. The implied value priced by investors for the rest of the company, including the R&D assets, and the System sales, is $2bn, since Axcelis has a $3bn EV, or 2.3x, unreasonably low in our view.

Catalysts. It is key to understand how and when expectations will shift. The company reports earnings on May 2nd, and there will be an investor date on July 7th. I believe management will provide more clarity on the future of Axcelis, both of Ion Implanters and new TAM, which will crystallise the terminal vale and reduce risk perception among investors. As a result, I believe the idea to be actionable in the next nine months.

Risks

Reliance On China In The Long Term. Since 2019, Axcelis reliance on Chinese semiconductor manufacturer demand has increased. In 2023, 46% of revenue originated from Chinese customers, compared to 31% five years earlier. A shift in the US Export restriction framework from advanced nodes to power devices would impede Axcelis ability to ship to Chinese customers. The upcoming US Presidential election could cause further uncertainty around US Export policy, and impact the share price.

The Next Act After Silicon Carbide. Management has yet to detail their vision of demand drivers beyond Silicon Carbide. They hinted at M&A, which highlights the risk that they see a limit to the Ion Implanter use case, with low growth or declining prospects.

EV Demand. As noted earlier, we think EVs drive 60-80% of SiC demand to the end of the decade. New designs or technologies could reduce EV volumes or SiC content per EV, posing risks to our forecasts. Furthermore, with China driving 35-45% of overall EV SiC demand, and adoption high already, demand could slow down faster than expected.

Conclusion

Axcelis shares have performed poorly since the start of the year, with valuation compressing and negative 2024 Sales and EPS revisions. Risks that negative earnings revisions continue in 1H24 are high, but the long-term tailwinds are strong. In our view, the upside risk far outweighs the downside risk, and we would buy the shares. I see $120 of intrinsic value, and 50-100% upside if valuation comes back to the 5y average.