Description

Thesis:

Soitec is a monopoly supplier of specialty engineered wafers consumed during semiconductor manufacturing. Their niche silicon-on-insulator (SOI) technology has gained technological relevance as performance requirements of chips has increased, most notably in smartphone RF applications where their RF-SOI solution has become the industry standard in switches. Further, we are emerging from a deep cyclical trough in smartphones and a covid inventory correction that is bottoming now. Going forward, Soitec content per phone continues increasing as the complexity of the front-end RF and Wi-Fi continues to go up and they ramp new POI filters. SOI FP is the only smartphone-levered stock which has not inflected with the cyclical trough + potential AI refresh cycle, and we believe a catch-up opportunity exists. Soitec is also expanding its TAM by adapting its core technology to create the best mousetrap for silicon carbide (SiC) wafers, which offers generational improvements in performance vs normal SiC that will inflect with 800V EVs in 2025/2026. We think the company can earn €8.00-8.50 in FY27. PT €175 over 1.75 years, 70% upside / 35% IRR.

Background:

Soitec is a monopoly supplier of specialty engineered wafers consumed during semiconductor manufacturing. Their core technology is called SmartCut used to create silicon-on-insulator (SOI) wafers, which acts like an atomic scalpel able to slice thin layers off traditional silicon wafers, insert an insulating layer, and combine it with a commodity bulk silicon wafer. This reduces electrical leakage / reduces noise, which enables 20-25% smaller die sizes, directly translating to lower cost and higher performance for chipmakers. Their different products (RF-SOI, FD-SOI, POI, etc) are all slight variations of the same technology used for different applications. Their direct customers are the foundries, but the use of SOI is dictated by the fabless chip designers. By end market, roughly 2/3 of revenue is from Mobile Communications, 20% Auto/Industrial, and the remainder in Edge and Cloud AI. They have 70% wafer share of SOI, and the remaining 30% license SOI technology from them and pay royalty per wafer.

Investment Thesis / Variant Perception:

Thesis point 1: Smartphone SOI content will continue to increase, with AI driving a potential refresh cycle and an indirect tailwind to RF content.

RF-SOI became the industry standard with 4G, now used in 100% of smartphones globally in switches. At higher frequencies and power, the insulating layer became necessary to reduce noise and keep die sizes under control. SOI content is found across the RF front end in smartphones today. The company has disclosed that average SOI content per phone has grown at +10% CAGR over the last few years with the 5G transition. Share is stable given performance advantages and high switching costs from having to re-qualify the entire supply chain as well as RF regulators around the world (ie FCC).

Mobile data traffic is expected to grow at a 20% CAGR through the end of the decade. To support this, the value of RF Front End chips goes up every year and with each generation. For Soitec, content will continue to increase with 5G-A and 6G as we will continue to have more complex RF band combinations, Wi-Fi 6E/7 adoption expands (adds new frequency), and potentially some more mmwave adoption (upside lever not in base case).

POI (Piezoelectric-on-Insulator) is a new product expected to drive 2/3 of mobile revenue growth in the mid-term ambition. There are currently 9 customers ramping their new POI filter product (with 10 more in qualification). QCOM is the lead customer, but China is the largest growth wedge given the growth in domestic chipmaking. In Wi-Fi, we are seeing RF-SOI starting to displace GaAS (gallium arsenide) given higher frequencies, which will increase the SOI content in phones significantly.

Note AAPL is a significant DD% indirect customer, so volume upside directly helps Soitec. As you upgrade the processor with AI functionality, the RF typically is also upgraded. For example Samsung content went up significantly y/y in Galaxy S24 as it rolled out FD-SOI for mmwave, and the AI functionality also drove volume upside. A high-end smartphone today has ~70 mm² of SOI content, and upgrading the Wi-Fi and POI filters alone can double this to 140mm. Mmwave would be upside to that.

Apple is forcing an upgrade cycle this year by only rolling out AI functionality on iPhone 15 Pro and newer (<10% of the install base). While the exact cadence of the AI penetration is unknown at this point, it is only going up which will drive higher device values including RF.

Thesis point 2: The smartphone cycle is at a cyclical trough.

The key debate and why the stock has been hurt has been overshipping to customers during covid with take-or-pay contracts. We think the trough is in 1HFY25 (in line with company guidance just reiterated), but more importantly normalize through-cycle growth rates in FY26-28 which drives our valuation. We back into this in two ways: 1) We use IDC end market unit growth rates and assume a deceleration in content vs the peak 5G years. 2) We break apart RF-SOI vs new POI product revenue and conservatively do not even have RF-SOI returning to FY22 levels.

It is important to note their products are consumables, not equipment, so it is only a matter of time before the inventory is worked through. End smartphone units have been positive for 3 Qs already (as per IDC), foundry customer inventories are coming down, and chipmaker revenues are also back to growth.

Thesis point 3: Their SmartSiC is the best solution for Silicon Carbide which is inflecting now.

As background, SiC is mainly used in EVs for the on-board inverter. The benefit from moving from legacy IGBT silicon 400V to SiC 800V is reducing charging time by 50% while extending the driving range by 5-10%, and costs about $250-300 more per vehicle. This becomes a no-brainer for the industry over the next few years as SiC supply becomes available, inflecting now with adoption from <10% today to 20-30% in the next 3 years and potentially 50%+ within 5 years.

VAR was very supportive that Soitec’s new SmartSiC product is the best solution for SiC. There are significant performance and cost/yield benefits vs monocrystal SiC, as electrical resistance goes down meaning you can shrink the size of the die, which reduces costs and yield also go up by a few % (compounding benefit). It also diversifies SiC supply, which is very critical for winning designs with Auto OEMs. Importantly, because Soitec uses monoSiC wafers as the donor wafer, if the price of SiC goes down (near term narrative around China) they are the only supplier where the COGS actually goes down. Soitec is pricing these at a slight premium to monoSiC which still yields the customer net benefit.

The company is expected to volume ramp the product with lead customer STM as well as a smaller second customer in late FY25/FY26, with qualifications being completed over the next year. To be clear, the technology is already proven out and they are already producing units from the factory. Expectations have already been reset as the original revenue goal of $200M was pushed out past FY26 given cyclical headwinds impacting STM’s lead customer TSLA.

Soitec has a long history of producing high-quality wafers with high 90s yields, and SmartSiC is only a slight adaptation of their core technology. We assume they can get to those yields by FY28 as the 500k wafer fab has already been built, which would generate ~$300M of revenue.

Valuation

We believe skew is very positive as the company has already reset FY25 guide including a steep correction in 1H, and removed the date on its medium term financial targets ($2B revenue/40% EBITDA margin). Management has low credibility and buyside does not believe the targets are achievable. We are less pessimistic and are getting ~20% EPS CAGR over the next several years despite being below sell side. We think the company can reasonably earn €8.00-8.50 in FY27. Assuming 21x P/E to near trough levels (historical range 20-40x), we get PT €175 over 1.75 years, 70% upside/35% IRR.

Importantly, every other smartphone-levered name has inflected on cyclical recovery + potential AI refresh cycle. We believe the opportunity still exists here from being an underfollowed SMID cap European stock further up the supply chain, with low management credibility. This is changing as for the first time in ~18 months, numbers did not get revised lower.  

Catalyst

Cyclical recovery / earnings upgrades