Business

Know the Business — Wolfspeed (WOLF)

Wolfspeed is a vertically integrated silicon-carbide (SiC) materials and device manufacturer with one giant 200mm fab in upstate New York, one substrate franchise that has been the industry standard for a decade, and a brand-new post-emergence balance sheet. It is a high-fixed-cost capacity ramp — a recovery option whose payoff depends on Mohawk Valley utilization climbing, substrate ASPs holding, and design-wins translating to ship volume. The market is currently underwriting the recovery (EV roughly 5x revenue versus 1.7–3.7x for profitable peers); the bear case is that the SiC overcapacity worked into the system in 2022–24 doesn't drain fast enough.

1. How This Business Actually Works

Wolfspeed sells two product families with very different economics — SiC bare/epi wafers (Materials) and SiC MOSFETs, Schottky diodes, and power modules (Power Products). Materials is a process-IP business: pull silicon-carbide boules out of crystal-growth furnaces, slice them into wafers, sell them to peers (Renesas, Bosch, others) under multi-year contracts. Power Products is a fab business: take those same wafers, run them through 200mm device fabrication at Mohawk Valley, and ship MOSFETs to automakers, industrial customers, and a fast-growing AI-data-center pocket. The two halves were roughly 50/50 of revenue in FY2025 ($344M Materials, $414M Power).

The engine is simple to describe and brutal to operate. A 200mm SiC device fab is a $2–3B asset that depreciates whether you run it at 20% or 90% of capacity. Fixed costs (depreciation, technicians, utilities, IATF 16949 compliance) are roughly 60–70% of unit cost at scale; the variable piece (wafer cost, gas, photoresist) is the rest. Below ~60% utilization, depreciation alone consumes the gross margin. That is exactly what happened in FY2024–FY2025: Wolfspeed booked $105M of underutilization costs in FY25 and $124M in FY24, dragging consolidated gross margin to -16% in FY25 and Mohawk-Valley-specific gross margin deep into negative territory.

The reversal is visible in the Q3 FY26 print: GAAP gross margin recovered to (27)% from (46)% the prior quarter as the company shut the legacy 150mm Durham fab and consolidated production at Mohawk Valley. The math from here is straightforward: every 5–10 percentage points of Mohawk Valley utilization is worth roughly 10–15 points of consolidated gross margin once fresh-start depreciation step-downs flow through inventory (management has guided to a $30M/quarter D&A reduction).

Wolfspeed sits across the first four stages. That vertical integration is the bull case when SiC adoption accelerates — it captures more value per device than a fabless GaN player or a pure substrate house. It is the bear case in a trough, because there is no profitable analog or microcontroller franchise to absorb the underutilization. Two customers are 37% of revenue (Note 15, FY2025 10-K) — that concentration is the second-largest risk after utilization.

2. The Playing Field

Wolfspeed competes head-on with companies 4–20x its size, all of whom run profitable broad-line semi portfolios alongside their SiC investments. That asymmetry is the central competitive fact of this business.

Profitable broad-line peers cluster at 29–39% gross margin and 1.7–3.7x EV/Sales. WOLF sits alone at -16% margin but 5x sales — pricing a margin trajectory, not a margin reality. NVTS at 31x is a different kind of bet (fabless GaN, much smaller, R&D-funded story). The good version of this business looks like onsemi: vertically integrated into substrate (the GTAT acquisition), 33% gross margin, 3.7x sales / 29x EBITDA. The path from WOLF to ON is not innovation — Wolfspeed already has the 200mm technology lead — it is volume. One more EV traction-inverter design-win at Tesla or BYD scale and the utilization curve would do the rest of the work.

3. Is This Business Cyclical?

Yes — extremely cyclical, and Wolfspeed has the highest beta to the SiC cycle of any listed name. The cycle hits in three places before it hits earnings: wafer ASPs (substrate prices), fab utilization (the depreciation absorber), and design-win timing (when revenue recognized lands relative to capex sunk three years earlier).

The cycle took gross margin from 32% in FY23 to -16% in FY25. That move was not demand collapse — revenue is roughly flat at $758M — it was the depreciation step-up from the Mohawk Valley ramp colliding with weaker-than-modeled EV unit growth.

The cycle is asymmetric: gross margin can improve faster on the way up than it deteriorated on the way down, because utilization moves on top of a now-much-smaller depreciation base (the FY25 PP&E write-down took net PP&E from $3.92B to $717M under fresh-start accounting). The pure-play hits the trough harder than diversified peers, because IFX/STM/ON spread underutilization across analog and MCU lines that Wolfspeed lacks.

4. The Metrics That Actually Matter

For a fixed-cost, single-fab business in a capacity-led trough, three operating metrics drive the stock and one balance-sheet metric drives survival.

The composite picture: balance sheet is a 4 (fixed), commercial pipeline is a 4 (still winning despite distress), operations are a 1–2 (the core unfinished work). The whole thesis lives in the gap between commercial-pipeline strength and operational performance closing.

5. What Is This Business Worth?

This is not a SOTP business. The two segments (Materials and Power) share fab infrastructure, capital, customers, and tax credits; valuing them separately would obscure more than it reveals. The right lens is operating leverage on a single-fab story, with three value drivers.

The valuation lens that fits is EV / forward revenue at modeled steady-state utilization, calibrated against onsemi at 3.7x sales and 33% gross margin. At today's $65 share price and ~45M shares, market cap is roughly $3.0B and EV roughly $3.5B. On trailing revenue of ~$0.71B, that is ~5x. On a modeled FY28 revenue of $1.3–1.5B (consensus rough range), it compresses to 2.3–2.7x — peer-like, but only if the volume actually arrives.

The shape of the scenarios is the lesson: the spread between bear and bull is roughly 3.5x of equity value, and which one prints depends on operational facts visible in quarterly Mohawk Valley revenue prints over the next four quarters. This is a name to underwrite by watching utilization, not by modeling P/E.

6. What I'd Tell a Young Analyst

Stop reading the consolidated P&L; read Mohawk Valley revenue. That single line — disclosed quarterly — is the most informative number in any Wolfspeed filing. Everything else (gross margin, operating loss, EBITDA, EPS) is a derivative of how fast that fab fills. Build your model on Mohawk Valley revenue trajectory and assume the rest follows mechanically.

Treat this as a recovery option, not a value stock. Pre-emergence, equity was nearly wiped out; post-emergence, the new equity holders bought a call option on SiC adoption with a strike price somewhere around 50% fab utilization. Size accordingly — it has been volatile (90-day return was up nearly 200% by mid-May 2026), and the right framing is risk-reward, not intrinsic value.

The three things that would change the thesis: (1) a Tier-1 EV OEM moving a major platform away from Wolfspeed to a Chinese SiC supplier; (2) substrate ASPs dropping another 30%+ as Chinese 8-inch capacity ramps; (3) gross margin not turning positive by H2 FY2027 even with the Durham 150mm fab closed and the depreciation step-down flowing through inventory. Any one of these reverses the recovery narrative.

The three things that would confirm it: (1) Mohawk Valley quarterly revenue above $150M with the underutilization-cost line shrinking; (2) AI data-center revenue growing from low single-digit % of total to 10%+ (this changes both the customer-concentration story and the multiple narrative — hyperscaler revenue trades at a different multiple than automotive); (3) reinstatement of any portion of the $750M preliminary CHIPS Act award to the post-emergence entity.

What the market is most likely getting wrong. Bulls underestimate how long ASP erosion at the substrate layer can compress segment-level returns even as device-layer utilization improves. Bears underestimate the operating leverage embedded in the now-much-smaller fresh-start depreciation base — the same revenue dollar carries dramatically more gross margin under the Successor accounting. Both effects are real; which one dominates is the trade.