Wolfspeed
This is the archived full research. Read the story
Wolfspeed grows silicon carbide crystal and makes the power chips that electric cars, solar inverters and data-centre power supplies use to waste less energy.
The scores
Growth1/4 ↓
Revenue fell 12.2% in fiscal 2026 and 24.1% in Q4; Power's gain leaned on closure buys while wafers fell 39.3%.
Business quality1/4 ↓
Its scale is idle, gross margin is negative, Chinese wafers undercut it and its large buyers grow their own crystal.
Financial position2/4 →
About nine quarters of cash above the covenant floor at Q4's burn; breakeven needs sales up a fifth or new money.
Execution record2/4 ↑
The old team took the company into Chapter 11 and faces a fab-disclosure suit; the new team's three guides hit, but set low.
Per share capture2/4 ↓
Notes and a warrant would add 55.9M shares to 55.6M; not lower because conversion also retires $839.7M of debt.
Potential2/4 →
Real markets in cars and data centres, but nothing signed with volume and a downturn to 2027 or 2028 first.
Risk profile3/4 →
Two unnamed customers are 38% of sales and Renesas holds 34.2%; serious, but no single ruling or contract decides the case.
Valuation2/4 ↓
Counting notes as shares, buyers pay a rival's price per sales dollar, assuming sales up about a quarter and margins never earned.
Evidence quality3/4 →
Debt, shares and sales are filed and the wafer market is measured outside; data-centre revenue and recent wafer shares are unknown.
If it works 2.0× · if it fails you keep about 0%.
The Verdict
We rate Wolfspeed Unattractive. It grows silicon carbide crystal and makes the power chips that electric cars and data centres use to waste less energy. Bankruptcy fixed the debt, not the business, which is weak: sales are falling, every sale loses money, cheaper Chinese wafers and customers making their own take ground, and cash above the lenders' floor runs short. Counting the lenders' notes as shares, the price matches profitable rivals. As a bet it is fragile: roughly double if it recovers, wiped out in a second restructuring if not.
The long caseElectric cars and AI data centres are moving to higher voltages, where silicon carbide chips work better than silicon; Wolfspeed must win that business back from bigger rivals.
How we weighed it
The price against our value
Cuts both ways
No fair value is drawn: the forecasts show no lasting profit after stock pay, so the bet prices it.
The business
Against it
Weak
A weak business: sales fall and it loses money on each sale.
Growth against what the price assumes
Against it
Wafer sales fell below our line and data-centre sales have no figure.
What could go wrong
Against it
Cash runs short, lenders hold half the upside and a large holder may exit.
Against competitors
Against it
Priced inside the range of profitable rivals despite its losses.
How its group is behaving
Cuts both ways
AI chip names move together on spending news.
The latest news and report
Cuts both ways
The last report narrowed the loss per sale while wafer sales fell.
The evidence under this - 8 claims
- Treating the notes as debt instead gives about $2.41B, 3.6 times sales, below both rivals; the vendor prints 4.5 times on its own share count.
- The court's own emergence value was $2.6B (C36), close to the debt-basis figure.
- 2026-11-04editedre-struck at the 2026-11-04 close on the Q1 fiscal 2027 balance sheet
- 2026-11-11editedre-struck at the 2026-11-10 close, up 15% since 4 November, on the 10-Q's share count and debt
Yole's later read is an overcapacity downturn lasting to 2027 or 2028, with factories at about 50% and 70% utilisation in 2025 (C23).
The company
Power electronics switch electricity on and off thousands of times a second to change its voltage. Chips made of silicon carbide do this faster and waste less energy as heat than ordinary silicon, which matters in an 800-volt electric car or a data-centre power supply.
Wolfspeed sells two things: the raw wafers other chipmakers build on, and its own finished chips. Its wafer buyers now also buy cheaper Chinese wafers; its chip buyers include carmakers and power-supply makers such as LITEON.
The evidence under this - 6 claims
A TrendForce piece a month earlier put international 6-inch prices at $750 to $800, so the sub-$500 figure is a Chinese quote, not necessarily Wolfspeed's price.
How it makes money
A dollar arrives when a carmaker, a distributor or a chipmaker orders chips or wafers, mostly by purchase order with no long-term promise. In fiscal 2026 chips brought $456.5M and wafers $208.6M.
Making the crystal and running the fab cost mostly the same whether the plants are full or not. At today's sales the plants sit partly idle, so each sale loses money. Management says gross profit turns positive near $800M of yearly sales.
The evidence under this - 4 claims
- At $807.2M of revenue in fiscal 2024 the vendor's gross margin was 9.6%, and at $757.6M in fiscal 2025 minus 16.1%: near $800M the company earned little gross profit, not the 34% to 37% rivals earn (C38).
- Substrate prices fell about 30% in 2024 (C14), so the same wafer volume brings less revenue and the breakeven in dollars can drift up.
- 2026-11-05editedQ1 fiscal 2027 call read: the chief financial officer says breakeven is now a little under $200M a quarter.
Where the money comes from, stream by stream
| Stream | Share of sales (FY2026) | How it earns it |
|---|---|---|
| Power | 68.6% | Its own silicon carbide diodes, MOSFETs and modules, sold to carmakers, industrial and data-centre power makers, about 45% of all revenue through distributors. |
| Materials | 31.4% | Bare and epitaxial silicon carbide wafers sold to other chipmakers, who turn them into their own power devices. |
Who runs it
Robert Feurle took over in May 2025, steered the company through bankruptcy and has met each sales forecast since, all set low. Insiders own under 1%, because the old shares and awards were wiped out; the lender-turned-holder Renesas owns about a third. Last year's bonus paid half its target, all of it on a factory-output measure the company keeps secret, while sales, cash flow and profit goals all missed. Share awards ride on the stock against the market, sales and cash.
The evidence under this - 4 claims
Ownership is low because every pre-bankruptcy award was cancelled and the new grants only began vesting on October 1, 2026; the proxy sets ownership guidelines of six times salary for the CEO.
- The target and result for yielded amps per wafer are withheld as competitively harmful, so the only paid measure cannot be checked from outside.
- The annual long-term mix is about 60% PSUs and 40% RSUs for certain executives, and one-time post-emergence RSUs vest on time alone.
- 2026-11-04editedQ1 fiscal 2027 landed inside the guide; Q2 target from the Q1 release
Why it wins, and who could beat it
Why customers come back
Wolfspeed's edge was size: the most crystal-growing capacity and the most experience, which made it the biggest wafer seller, with 33.7% of the market in 2024 by TrendForce's count.
That edge is wearing away. Two Chinese makers already held about 17% each, a Chinese wafer cost under $500 in 2024 after a 30% price fall, and Wolfspeed's wafer sales fell 39.3% in fiscal 2026.
The evidence under this - 5 claims
A TrendForce piece a month earlier put international 6-inch prices at $750 to $800, so the sub-$500 figure is a Chinese quote, not necessarily Wolfspeed's price.
The threat
The attack comes from three sides. SICC in China sells wafers from a plant running nearly full. onsemi and STMicroelectronics, once wafer buyers, now grow their own. And in data centres NVIDIA lists fourteen chip suppliers for its 800-volt racks, Infineon and onsemi among them, but not Wolfspeed.
The evidence under this - 3 claims
The rivals: ON, STM, NVTS, COHR, and 3 more that are private, in-house or listed abroad.
Who else does this, rival by rival
| Ticker | Rival | Where it competes | How far along | Priced against it |
|---|---|---|---|---|
| SICC (Shandong Tianyue) conductive SiC wafersPriced against it: Listed abroad | Materials Products | On sale | Listed abroad | |
| Private | TanKeBlue SiC wafersPriced against it: Private | Materials Products | On sale | Private |
| ON | onsemi's own crystal and EliteSiC MOSFETsPriced against it: Yes | Materials Products | On sale | Yes |
| STM | STMicroelectronics SiC MOSFETs and in-house wafersPriced against it: Yes | Power Products | On sale | Yes |
| Infineon CoolSiC and CoolGaNPriced against it: Listed abroad | Power Products | On sale | Listed abroad | |
| NVTS | Navitas and Innoscience gallium nitridePriced against it: Yes | Power Products | On sale | Yes |
| COHR | CoherentPriced against it: One segment only | - | On sale | One segment only |
How big it could get
How big this could get
The market is real. Electric car sales grew 20% to more than 20 million in 2025, and Yole expects silicon carbide chips to be a $10.3B market by 2030.
But it is in a slump first: factories ran at about half capacity in 2025, and Yole expects the glut to last until 2027 or 2028. Wolfspeed's $665.1M of sales is already a sizeable slice of today's market.
The evidence under this - 5 claims
Yole's later read is an overcapacity downturn lasting to 2027 or 2028, with factories at about 50% and 70% utilisation in 2025 (C23).
Yole's own March 2025 forecast had a strong rebound in 2026, so its timing has already moved once.
Who has signed
Nothing with a volume. LITEON approved Wolfspeed's chips for its 800-volt power supplies, and GE Aerospace signed a non-binding memorandum on high-voltage chips. The two biggest customers, at 11% and 27% of sales, are unnamed and buy order by order.
The evidence under this - 3 claims
Where the growth is coming from, business by business
Where the growth is coming from
| Business | Share of revenue | Growth: last quarter | Last half-year | Last year | Share in 3 years at its latest-quarter rate |
|---|---|---|---|---|---|
| Power Products | 75% | -11.4% | - | - | 91% |
| Materials Products | 25% | -41.2% | - | - | 9% |
| Whole company | 100% | -21.2% | - | - | 100% |
On the filed split through 2026-09-27, if each business keeps its latest quarter's growth, the company grows about -19% next year and -15% in year 3, against -12.2% over the last twelve months. A scenario, not a forecast: no business holds its rate.
How we read it: The score believes the blend: Power's rise came partly from end-of-life buys before the Durham closure and has since fallen 10.4% on a year, so neither business is carrying the case.
The results
The numbers
Sales fell 18% over the twelve months to Sep 2026, slower than the −6.1% in the year to Jun 2025.
Operating margin is −80%, down from −41% in the year to Jun 2023.
Running costs take 50¢ of each sales dollar, down from 91¢ in the year to Jun 2021.
Yearly sales growth has ranged from −6.1% to +33% over the last 5 years, with sales falling in 1 of them.
It used $348M of cash over the last twelve months, so the $995M it holds lasts about 2.9 years at that pace.
It owes $686M more than the cash it holds, and its operations bring in no cash to repay it.
The share count fell 51% in 5.3 years, and pay in shares costs 6.0% of sales.
All the numbers
Filed figures from the company's 10-K and 10-Q, read 2026-11-10. Fiscal years end in June.
| Jun 2021 | Jun 2022 | Jun 2023 | Jun 2024 | Jun 2025 | 12 mo to Sep 2026 | |
|---|---|---|---|---|---|---|
| Size and growth | ||||||
| Sales | $526M | $572M | $758M | $807M | $758M | $623M |
| Sales growth | +12% | +8.8% | +33% | +6.4% | −6.1% | −18% |
| Profit | ||||||
| Gross margin | 31% | 36% | 32% | 9.6% | −16% | −30% |
| Operating margin | −60% | −36% | −41% | −55% | −175% | −80% |
| Net margin | −100% | −35% | −43% | −107% | −212% | −90% |
| What growth costs, per sales dollar | ||||||
| Research and development | 34% | 25% | 22% | 25% | 23% | 17% |
| Selling, general and admin | 35% | 32% | 28% | 31% | 25% | 23% |
| Pay in shares | 10% | 9.4% | 9.6% | 11% | 9.7% | 6.0% |
| Capital spending | 109% | 111% | 125% | 282% | 168% | 13% |
| Cash | ||||||
| Cash from operations | −$126M | −$154M | −$143M | −$726M | −$712M | −$265M |
| Free cash flow | −$696M | −$791M | −$1.09B | −$3.00B | −$1.98B | −$348M |
| Balance sheet | ||||||
| Cash and short-term investments | $1.15B | $1.20B | $2.95B | $2.17B | $955M | $995M |
| Debt | $824M | $1.02B | $3.03B | $6.16B | $6.54B | $1.68B |
| Net cash | $331M | $177M | −$70.7M | −$3.99B | −$5.58B | −$686M |
| Returns and shares | ||||||
| Shares (diluted) | 112M | 120M | 124M | 126M | 141M | 55.0M |
What growth costs
Wolfspeed grew by building factories, and their cost stays whether they are busy or not. As sales fell, that fixed cost turned every sale into a loss. Research and office costs have shrunk through cuts, not because sales outgrew them. Only a return to much higher sales would make growth profitable again.
The evidence under this - 3 claims
The fall since fiscal 2024 came from cuts and restructuring while sales shrank, not from sales outgrowing costs: research and development dollars fell from $201.9M to $111.7M (S01).
- At $807.2M of revenue in fiscal 2024 the vendor's gross margin was 9.6%, and at $757.6M in fiscal 2025 minus 16.1%: near $800M the company earned little gross profit, not the 34% to 37% rivals earn (C38).
- Substrate prices fell about 30% in 2024 (C14), so the same wafer volume brings less revenue and the breakeven in dollars can drift up.
- 2026-11-05editedQ1 fiscal 2027 call read: the chief financial officer says breakeven is now a little under $200M a quarter.
The last report (4 Nov)
Quarter ended September 2026, filed figures.
Sales of $155.0M were down 21% on a year and above the middle of the $140M to $160M forecast. Chip sales were $116.8M, up from $106.3M in June, but wafer sales fell to $38.2M from $43.3M. It still lost 22 cents of gross profit on each dollar, 13 cents on its adjusted basis, from 20.
It used $84M of cash in operations and guided the next quarter to $155M to $175M. It gave no dollar figure for data-centre sales. On the call it said the $1.5B government loan could close by year end after a limited share sale.
What it did to the thesisThe quarter narrowed the loss per sale and chip sales grew, but wafer sales fell below the line we set and data-centre sales still have no dollar figure. Neither side of our test won, so the Unattractive view stays.
The quarter in detail
Sales growth from the previous quarter, filed figures.
Earnings per share $-2.03 against $-1.99 expected.
The evidence under this - 7 claims
- The figure is unfiled and unsized; the 10-K names AI data centres only as an emerging application, and total Power revenue still fell 10.4% in Q4 (C03).
- NVIDIA's 800-volt partner lists name fourteen silicon suppliers and not Wolfspeed (C16).
- 2026-11-04editedtested by the Q1 fiscal 2027 release, which again gave no dollar figure
- 2026-11-04editedQ1 fiscal 2027 landed inside the guide; Q2 target from the Q1 release
- The prior-year quarter is the predecessor company before fresh-start accounting, so margins and costs are not like for like; revenue is.
- Materials fell below the $43.3M of Q4, operating cash outflow widened from $54M to $84M, and the release gives no dollar figure for AI data-centre revenue; the call was not yet available.
- At $807.2M of revenue in fiscal 2024 the vendor's gross margin was 9.6%, and at $757.6M in fiscal 2025 minus 16.1%: near $800M the company earned little gross profit, not the 34% to 37% rivals earn (C38).
- Substrate prices fell about 30% in 2024 (C14), so the same wafer volume brings less revenue and the breakeven in dollars can drift up.
- 2026-11-05editedQ1 fiscal 2027 call read: the chief financial officer says breakeven is now a little under $200M a quarter.
- 2026-11-05editedQ1 fiscal 2027 call read: government checks about two weeks from done, an equity raise still needed, closing aimed before year end.
What management did with the money
The cash came from the government, not customers: $733.2M of tax-credit and grant refunds in fiscal 2026. Management used much of it to repay the costliest debt, though $635.9M still costs about 16% a year.
Insiders in 2026Q3: 6 purchases and 4 sales, from their Form 4 filings.
The evidence under this - 3 claims
- 2026-11-11editedre-sourced to the Q1 fiscal 2027 10-Q debt table: $10.0M of the $12.23 notes converted
The price, and where we disagree
In short. Full. Counting the lenders' notes as shares, buyers pay about what the cheapest profitable rival costs, for a company that loses money on every sale.
The shares are 27% below the 1 October level this view was struck at, after nearly doubling in six months on a new higher-quality wafer and a rally in chip stocks.
This has happened before: the price ran up after the bankruptcy and again in May 2026, and both times fell back by more than half. Meanwhile 58.1M shares, including Renesas' 34.2% stake, were registered in September so holders can sell them.
The evidence under this - 5 claims
How we value it
No fair value: the analysts' forecasts show no lasting operating profit after stock pay, so a value built on earnings would rest on a guess; the bet prices it instead.
How we read the multiple
The lenders' notes turn into shares at prices far below today's. Counting them, buyers pay about 4.4 times a year's sales, inside the range between STMicroelectronics and onsemi when we last compared them.
For the price to look ordinary beside STMicroelectronics:
- Wolfspeed would need about $720M of yearly sales, about 16% more than it sold over the last four quarters.
- Rivals at that price also earn 34 to 37 cents of gross profit per dollar, which Wolfspeed never has.
- Management puts the gross-profit breakeven at about $800M of yearly sales.
Against its own history. Dearer than its own short history as a new company, while sales per share fell, so the rise is hope about the future, not a better business.
The evidence under this - 6 claims
- Treating the notes as debt instead gives about $2.41B, 3.6 times sales, below both rivals; the vendor prints 4.5 times on its own share count.
- The court's own emergence value was $2.6B (C36), close to the debt-basis figure.
- 2026-11-04editedre-struck at the 2026-11-04 close on the Q1 fiscal 2027 balance sheet
- 2026-11-11editedre-struck at the 2026-11-10 close, up 15% since 4 November, on the 10-Q's share count and debt
Two analysts expect $794.9M of revenue in fiscal 2028 (C43), near the breakeven, so the market may be paying for a recovery that consensus already sees.
- 2026-11-11editedre-struck on the 10 November company value (C37)
- At $807.2M of revenue in fiscal 2024 the vendor's gross margin was 9.6%, and at $757.6M in fiscal 2025 minus 16.1%: near $800M the company earned little gross profit, not the 34% to 37% rivals earn (C38).
- Substrate prices fell about 30% in 2024 (C14), so the same wafer volume brings less revenue and the breakeven in dollars can drift up.
- 2026-11-05editedQ1 fiscal 2027 call read: the chief financial officer says breakeven is now a little under $200M a quarter.
- The company may settle conversions in cash, and Renesas' notes stop being convertible in September 2027, so fewer shares may be issued.
- Holders have already converted $64.3M voluntarily, which lowered debt and interest.
- 2026-11-11editedre-struck on the 10-Q's share count and the notes left after the quarter's conversion
Priced against its rivals
The competitors its own annual report names, each confirmed against the rival's filing. Measured on company value for each $1 of sales, cheapest first:
| Ticker | Company | company value for each $1 of sales |
|---|---|---|
| STM | STMicroelectronics | 3.8× |
| WOLF | Wolfspeed | 4.5× |
| ON | ON Semiconductor | 5.5× |
| NVTS | Navitas Semiconductor | 75.8× |
The middle of the other 3 is 5.5. Wolfspeed Inc. is 2nd cheapest of 4.
Rivals' figures come from a data provider, so read the spread of the group rather than any single pair.
Priced inside the range of profitable rivals despite losing money on each sale, so the gap in quality is not reflected in the price.
Private or foreign rivals that cannot be priced: Infineon Technologies AG, Rohm Co. Ltd., Robert Bosch GmbH, Silan Microelectronics Co. Ltd., SICC Co., Ltd., TanKeBlue Semiconductor Co., Ltd.
The bet: what it pays if it works
Even if it works, the shares look expensive. The best case we can defend is sales half again as large as today's within about three years, helped by a growing market and the LITEON approval. At what profitable rivals cost, that roughly doubles the company's value, and the lenders' new shares take much of it.
If it fails, two more years of losses leave little cash against the debt, and shareholders could be wiped out in a second restructuring.
| Revenue in two to three years | Points of that market needed | At the stated multiple |
|---|---|---|
| $1.00B (the big win) | about 8 at 100% kept | 2.0× |
| $1.40B | about 12 at 100% kept | 2.8× |
To pay 2× (a Fair bet) it would need about $996.28M of revenue (about 8 points of that market); to pay 5× (a Cheap bet), about $2.49B of revenue (about 22 points of that market), at the same multiple.
The steps it has to take
- Done Emerge from Chapter 11 with debt cut
- Done All devices moved to the 200mm fab and the 150mm fab closed
- Done LITEON qualifies the MOSFETs for 800-volt racks
- Done Q1 fiscal 2027 revenue at or above $150M with adjusted gross margin above minus 19.9%
- Not started A filed 200mm wafer supply agreement with a named device maker due 2027-06
- Not started First-lien notes refinanced below their near-16% rate due 2027-06
- Not started Quarterly revenue of $200M with adjusted gross margin above zero due 2028-06
The detail behind the label
The evidence under this - 7 claims
Yole's later read is an overcapacity downturn lasting to 2027 or 2028, with factories at about 50% and 70% utilisation in 2025 (C23).
- The figure is unfiled and unsized; the 10-K names AI data centres only as an emerging application, and total Power revenue still fell 10.4% in Q4 (C03).
- NVIDIA's 800-volt partner lists name fourteen silicon suppliers and not Wolfspeed (C16).
- 2026-11-04editedtested by the Q1 fiscal 2027 release, which again gave no dollar figure
- 2026-11-04editedre-sourced on the Q1 fiscal 2027 release balance sheet
- 2026-11-11editedre-sourced to the Q1 fiscal 2027 10-Q debt table: $10.0M of the $12.23 notes converted
Conversions cut interest, asset sales and inventory can release more cash, and a refinancing of the near-16% first lien would cut burn; capex could also rise if demand returns.
- 2026-11-04editedre-struck on September 27 cash; the tax-credit receivables were not re-read in the release
- 2026-11-11editedcapital spending from the Q1 fiscal 2027 10-Q added to the quarterly burn
Where we disagree with the market
The restructured Wolfspeed is America's silicon carbide champion for electric power and AI, with its factories as a floor under the price.
The debt was fixed, the business was not. Wafers are losing to cheaper Chinese supply and buyers growing their own.
The market counts the factories and the AI headlines. The filings show wafer sales down by more than a third, data-centre sales too small to state, and lenders who will own about half the company once their notes turn into shares.
We are wrong if quarterly sales pass $200M with a positive gross margin, or a named chipmaker signs a filed wafer supply contract with a volume.
What analysts say, and the evidence
The evidence under this - 7 claims
- Treating the notes as debt instead gives about $2.41B, 3.6 times sales, below both rivals; the vendor prints 4.5 times on its own share count.
- The court's own emergence value was $2.6B (C36), close to the debt-basis figure.
- 2026-11-04editedre-struck at the 2026-11-04 close on the Q1 fiscal 2027 balance sheet
- 2026-11-11editedre-struck at the 2026-11-10 close, up 15% since 4 November, on the 10-Q's share count and debt
- The figure is unfiled and unsized; the 10-K names AI data centres only as an emerging application, and total Power revenue still fell 10.4% in Q4 (C03).
- NVIDIA's 800-volt partner lists name fourteen silicon suppliers and not Wolfspeed (C16).
- 2026-11-04editedtested by the Q1 fiscal 2027 release, which again gave no dollar figure
- The company may settle conversions in cash, and Renesas' notes stop being convertible in September 2027, so fewer shares may be issued.
- Holders have already converted $64.3M voluntarily, which lowered debt and interest.
- 2026-11-11editedre-struck on the 10-Q's share count and the notes left after the quarter's conversion
What could go wrong, and how we would know
The risks
- Cash runs short. At the latest quarter's pace the cash above the lenders' $350M floor lasts about two years unless sales rise.
- Half the upside to lenders. The notes and a warrant would add more shares than exist today, so each share gains far less than the company.
- A big seller waiting. Renesas owns 34.2% and its shares are now registered for sale; its board member stepped down.
- Unnamed, uncommitted buyers. Two customers make up 38% of sales and buy order by order, with no long-term promise.
- A price war. Chinese wafers fell about 30% in price in 2024 and Chinese makers hold about 40% of capacity.
The evidence under this - 6 claims
Conversions cut interest, asset sales and inventory can release more cash, and a refinancing of the near-16% first lien would cut burn; capex could also rise if demand returns.
- 2026-11-04editedre-struck on September 27 cash; the tax-credit receivables were not re-read in the release
- 2026-11-11editedcapital spending from the Q1 fiscal 2027 10-Q added to the quarterly burn
- The company may settle conversions in cash, and Renesas' notes stop being convertible in September 2027, so fewer shares may be issued.
- Holders have already converted $64.3M voluntarily, which lowered debt and interest.
- 2026-11-11editedre-struck on the 10-Q's share count and the notes left after the quarter's conversion
A TrendForce piece a month earlier put international 6-inch prices at $750 to $800, so the sub-$500 figure is a Chinese quote, not necessarily Wolfspeed's price.
What we still do not know
- How large data-centre sales are in dollars; the company gives only growth rates.
- How much of the wafer market Wolfspeed kept in 2025; the latest outside count is for 2024.
- How much of the large short position is hedging by noteholders rather than bets against the company.
- Whether the conditional $1.5B government loan is ever signed; it needs Congress, conversion of the notes and new shares.
The evidence under this - 5 claims
- The figure is unfiled and unsized; the 10-K names AI data centres only as an emerging application, and total Power revenue still fell 10.4% in Q4 (C03).
- NVIDIA's 800-volt partner lists name fourteen silicon suppliers and not Wolfspeed (C16).
- 2026-11-04editedtested by the Q1 fiscal 2027 release, which again gave no dollar figure
Nothing filed splits hedges from outright shorts, and short interest rose 5.9% in the last fortnight with no new notes issued.
- 2026-11-11editedthe notes now convert into 50.9M shares (C29)
- 2026-11-05editedQ1 fiscal 2027 call read: government checks about two weeks from done, an equity raise still needed, closing aimed before year end.
What would change our view
- Unattractive→NeutralSales cross the gross-profit breakeven
It needs all of:
- ○Quarterly sales reach $200M, up from $155M.checked at the next results, due in Feb 2027
- ○Gross profit turns positive, from a loss of about 22 cents per sales dollar.checked at the next results, due in Feb 2027
- Unattractive→NeutralThe price stops assuming the recovery
It needs:
- ○The company's value falls to 2.5 times its last year of sales, about $15.60 a share counting the lenders' notes as shares.about ≈$14.90 today, 51% below the last close
We would also be wrong if
- Wolfspeed files a wafer supply contract with a volume from a named chipmaker.
- Data-centre sales are disclosed and pass a quarter of the total.
- Wafer sales fall below $35M in a quarter, or cash below $700M before sales recover.
What is coming
- February 2027 Q2 fiscal 2027 results: Revenue against the $155M to $175M target, whether Materials steadies, non-GAAP gross margin against minus 13%, and whether AI data-centre revenue gets a dollar figure.
Every dated event ahead
- February 2027 Q2 fiscal 2027 results: Revenue against the $155M to $175M target, whether Materials steadies, non-GAAP gross margin against minus 13%, and whether AI data-centre revenue gets a dollar figure.
- May 2027 Q3 fiscal 2027 results: Whether Power grows without the end-of-life buys that flattered fiscal 2026, and whether Materials steadies against cheaper Chinese wafers.
Since we wrote this
- 11 NovThesis reviewStill Unattractive: shares up 15%, priced at 4.4 times salesThe shares rose without new news about the business. Counting lenders' notes as shares the price is 4.4 times sales, inside profitable rivals' range, for a company losing money on each sale.
- 5 NovEarningsCall: loan checks weeks from done, breakeven under $200MManagement aims to close the $1.5B government loan by year end after a limited share sale, but nothing is signed and no data-centre figure came. Unattractive holds.
- 4 NovEarningsMixed quarter: loss narrowed, wafer sales fell furtherSales of $155M and chip sales beat our lines, but wafers fell to $38.2M and no AI figure came; the Unattractive view holds.
- 10 OctBet moveBet re-struck: Priced for success to SpeculationRe-priced at the 9 October close; the rating moved from Expensive bet to Fair bet.
- 8 OctNewsA conditional $1.5B government loan is announcedIt would repay the 16% notes but is unsigned and needs Congress, conversion of the notes and new shares, so the view held.
- 5 OctThesis reviewThe bet now names what it ridesThe owner asked that every bet name the macro trend behind it, after BWXT's bet was missing its chief executive's case for nuclear over gas.
- 4 OctThesis reviewOwner's checklist: revenue streams, cost of growth, management, catalystsThe owner's checklist found the method silent on who runs the company, what growth costs, where the money comes from and what is dated ahead; this pass fills those four blocks from the proxy and filings.
- 2 OctThesis writtenUnattractive, medium convictionWritten a year after the bankruptcy, six weeks after Q4 fiscal 2026 results, with the shares nearly doubled in six months on a new wafer and a chip rally.
Also checked 3 times with nothing new to act on.
Sources 39
Sources - 39
Every claim above cites one of these, or states the reasoning that derived it. The quote is frozen at the date it was read: a link can rot, and an agent fetch returns a reading of a page rather than the page. Backs counts the whole claim ledger, so it can name a claim this page does not surface.
- S01primarycompany filing2026-08-20Item 1 (products, plants, competitors, customers), risk factors, MD&A by product line, liquidity, the debt note with conversion prices, fresh-start valuation, the AMIC and New York grants, customer concentration, litigationbacksC01C02C03C04C09C10C11C20C26C27C28C31C34C36C42C44C46C49C50C51C55
- S02primarycompany release2026-08-19Q4 revenue by product line, GAAP and non-GAAP gross margin, adjusted EBITDA, operating cash flow, the Q1 fiscal 2027 outlook, the balance sheet and cash flow statement
AI data center revenue more than doubled year-over-year in fiscal 2026 and increased approximately 20% sequentially in the fourth quarter, demonstrating the long-term potential of this moderate but growing opportunity.
backsC03C04C05C08C41 - S03primarytranscript2026-08-19Q4 fiscal 2026 earnings call transcriptdata/cache/transcripts/WOLF-2026Q4.json (FMP earning-call-transcript, quarter label 2026Q4, call held 2026-08-19)the $800M gross-margin breakeven, the first-lien rate near 16%, $32M of cash interest in Q4, AI data-centre growth, LITEON and MacMic wins, 200mm substrate samples; questions from Susquehanna, TD Cowen, William Blair and JPMorgan (a separate origin from the answers)
But again, in a ballpark, we would say on an $800 million annual run rate, that's probably the ballpark where a breakeven gross margin point lies right now.
backsC05C06C08C19C21C27C49 - S04primarycompany release2026-05-05Q3 revenue of about $150M, gross margin, the first-lien refinancing, the Q4 guide of $140M to $160M
Continued sequential quarterly growth in AI data center applications of approximately 30%, reflecting a moderate but expanding part of the Company's business with meaningful long-term potential.
backsC04C08C19C41 - S05primarycompany release2026-02-04Q2 revenue of about $168M, $48M of underutilization cost, the Q3 guide of $140M to $160M and its stated reasons, AI data-centre growth of 50% on the quarter
The decline is driven by accelerated fiscal first half customer purchases, as certain customers built up inventory by placing orders from the Durham fab prior to its planned closure at year-end, certain customers pursuing second sourcing of products during Wolfspeed’s bankruptcy process and weaker EV demand.
backsC04C07C08C11C41 - S06primarycompany release2025-10-29Q1 revenue of about $197M, $47M of underutilization, $97M from the Mohawk Valley fab, the Q2 guide
The Company expects to generate revenue between $150 million and $190 million for its fiscal second quarter, driven primarily by accelerated customer purchases in our first fiscal quarter
backsC07C11C41 - S08primarycompany filing2026-09-1758,148,889 shares registered for resale by selling stockholders, itemised by sourcebacksC32
- S09primarycompany filing2026-09-11holders above 5% at August 31, 2026, the Renesas investor rights and registration rights; the executive officers and their start dates, the beneficial ownership of directors and officers, the fiscal 2026 bonus goals and results, the PSU design, and the annual meeting date and proposalsbacksC33C52C53C54C56
- S10primarycompany filing2026-08-25Renesas' board representative resigning effective September 27, 2026, replaced by an observerbacksC33
- S11secondaryvendor data2026-10-02Research pack, frozen at content/WOLF/research-pack.jsoncontent/WOLF/research-pack.jsonthe 2026-10-01 close, daily bars, the filed-competitor cohort multiples and margins, the price-move decomposition, analyst estimates and ratings, FINRA short interest, insider rows, the vendor news feedbacksC34C38C43C44C45C48
- S12primarycompany release2026-08-06the qualification and partnership; no volume, price or term. Business Wire refuses this client, so the company's newsroom copy was read
today announced a strategic partnership and the successful qualification of Wolfspeed's silicon carbide technology for deployment within LITEON's 800VDC power sidecar and compute rack PSU platforms.
backsC17 - S13primarycompany release2026-06-08a memorandum of understanding with GE Aerospace on 10 kV die and module formats; no volume
MOU includes the supply of Wolfspeed’s 10 kV MOSFET die and jointly determining standard high-voltage power module formats for commercial introduction.
backsC18 - S14primarycompany release2026-09-27the defect specification of the new 200mm substrate and that it is orderable now; no customer named
Wolfspeed’s Premium 200 mm n-type SiC substrate delivers fewer than 0.01 micropipes per square centimeter, or fewer than three micropipes per wafer
backsC21 - S15primarycompany release2026-07-07the five patents and the accused GaN and SiC product lines
The lawsuit asserts that a broad range of Navitas products infringes multiple Wolfspeed patents, including U.S. Patent Nos. 8,169,005, 10,998,418, 10,886,396, 10,749,443, and 11,888,392.
backsC20 - S16primarycompany filing2026-02-09a vertically integrated silicon carbide strategy inside its Power Solutions Group, selling into vehicles, energy and AI data centresbacksC12C38
- S17primarycompany filing2026-02-26STMicroelectronics Form 20-F, fiscal year ended December 31, 2025 STMicroelectronicsits own 150mm and 200mm SiC wafers at Norrköping, an integrated SiC substrate plant in Catania with EUR 292.5M of state aid, and SiC MOSFETs in its power portfoliobacksC12C38
- S18primarycompany filing2026-02-27Navitas naming Wolfspeed among its primary SiC competitorsread, but no claim rests on it
- S19primaryindustry data2025-10-13the fourteen silicon suppliers NVIDIA names for 800 VDC, which do not include Wolfspeed (the word appears 0 times on the page), and LITEON among the power-system makers
Silicon providers : AOS, Analog Devices, Efficient Power Conversion, Infineon Technologies, Innoscience, MPS, Navitas, onsemi, Power Integrations, Renesas, Richtek, ROHM, STMicroelectronics, Texas Instruments. ... Power system components : Bizlink, Delta, Flex, Lead Wealth, LITEON, Megmeet.
backsC16C17 - S20primaryindustry data2025-05-20NVIDIA Technical Blog - NVIDIA 800 VDC Architecture Will Power the Next Generation of AI Factories NVIDIAthe 2027 start of 800 VDC and the first partner list, also without Wolfspeed
NVIDIA is leading the transition to 800 VDC data center power infrastructure to support 1 MW IT racks and beyond, starting in 2027.
backsC16C25 - S21primaryindustry data2025-05-12TrendForce press release - 2024 N-type SiC substrate market and supplier shares, 12 May 2025 TrendForce2024 N-type substrate revenue of $1.04B, down 9%, and supplier shares; 8-inch over 20% of shipments by 2030
intensifying market competition and sharp price declines have pushed global revenue for N-type SiC substrates down 9% YoY to US$1.04 billion. ... Wolfspeed maintained its position as the top supplier in 2024 with a 33.7% market share, despite ongoing operational challenges. ... Chinese vendors TanKeBlue and SICC have rapidly emerged as major players, claiming 17.3% and 17.1% market shares respectively, placing them in second and third place.
backsC13 - S22secondaryindustry data2024-10-23the 6-inch price fall and level; 800V models at about 8.7% of China's new-energy car market with 72% SiC penetration. TrendForce's news desk citing industry reports, so secondary
the price of mainstream 6-inch SiC substrates has been consistently declining, with a drop of nearly 30%. As of mid-2024, industry insiders in China reveal that the price of 6-inch SiC substrates has fallen below USD 500, approaching the production cost of Chinese manufacturers.
backsC14 - S23secondaryindustry data2024-10-23TrendForce News citing Commercial Times - oversupply of 6-inch SiC substrate, 23 October 2024 TrendForce8-inch prices in China and Wolfspeed's 8-inch utilisation reported at 25%
At the end of 2023, the average quotation of 8-inch SiC substrates in China was approximately USD 3,000 to USD 4,000 per piece. By the second quarter of 2024, the price had dropped to just USD 2,000
backsC14 - S24secondaryindustry data2025-12-18Yole's read of the correction: utilisation, the Chinese share of wafer capacity, a downturn to 2027-2028 and device revenue near $10bn by 2030; yolegroup.com answers this client with an empty bot challenge, so the reprint was read
As of 2025, utilization rates have dropped to around 50% for upstream processes and 70% for device lines. ... In 2024, Chinese players already captured about 40% of SiC wafer and epiwafer capacity and are rapidly expanding into device manufacturing. ... The downturn is expected to persist until 2027-2028.
backsC13C23 - S25secondaryindustry data2025-07-01Yole's 2030 SiC device market, the automotive share and the data-centre opportunity
Despite a slowdown in EV sales which is delaying demand for SiC devices and wafers, Yole reckons the SiC market will have a 20% CAGR 2024-30 to reach $10.3 billion in 2030. Despite the slowdown, the automotive & mobility segment is expected to retain about 70% of SiC demand over the next 5 years. Yole also sees a: $200 million market opportunity for SiC in datacentres.
backsC22 - S26primarygovernment data2026-05-20electric car sales and share, the 2026 expectation, and that higher pack voltage moves power electronics to SiC and GaN; iea.org pages answer 403, the PDF on the agency's own blob storage answered
Electric car sales grew by 20% globally to exceed 20 million in 2025, meaning one-quarter of all new cars sold were electric. ... Global electric car sales are expected to grow to 23 million in 2026, representing 28% of total car sales.
backsC24 - S27primarycompany filing2025-08-20the named rival's own capacity ramp, its 16.7% share per Frost & Sullivan, its 8-inch mass production and 12-inch introduction, and industry substrate prices falling in 2024; the document header reads 'draft form'
In 2022, 2023 and 2024, our total production capacity was approximately 75,000 pieces, 270,000 pieces and 420,000 pieces of SiC substrates, respectively; our actual production was approximately 71,000 pieces, 262,000 pieces and 410,200 pieces of SiC substrates, respectively
backsC15 - S28primaryindustry data2026-04-01the 2026 read: capacity digestion over 2 to 3 years, pressure on substrate prices, 8-inch as the entry ticket to automakers
Profit is being redistributed along the value chain, where upstream material prices are under pressure, while device and IDM manufacturers show stronger earnings resilience.
backsC23 - S29primaryindustry data2024-06-20the top five device suppliers at 91.9% of 2023 revenue, ST first at 32.6%, onsemi second, and Wolfspeed's setbacks in devices
The top five suppliers accounted for approximately 91.9% of total revenue; ST lead the pack with a 32.6% market share while onsemi rose from fourth place in 2022 to second place. ... Wolfspeed’s operational strategy missteps have caused it to miss market opportunities over the past two years, leading to setbacks in its power device business.
backsC30 - S30primarycompany filing2026-08-14semi-insulating SiC substrates sold for GaN-on-SiC radio amplifiersread, but no claim rests on it
- S31primarycompany filing2026-09-02no open-market purchase or sale by any officer or director; three purchases and one sale by Citigroup as a listed 10% holder, all discretionary; the rest grants, vesting and withholdingbacksC47
- S32primaryindustry data2025-11-27TrendForce press release - AI to Reshape the Global Technology Landscape in 2026, 27 November 2025 TrendForcethe rate at which wide-bandgap chips displace silicon in data-centre power
TrendForce predicts that the adoption of SiC and GaN in data center power systems will reach 17% by 2026 and exceed 30% by 2030.
backsC25 - S33primarycompany filing2026-02-06product-line revenue for the quarter: Power $118.3M, Materials $50.2Mread, but no claim rests on it
- S34primarycompany filing2026-05-07product-line revenue for the quarter: Power $100.1M, Materials $50.1Mread, but no claim rests on it
- S35primarycompany filing2024-08-22income statements and cash flows for fiscal 2022 to 2024: research and development, sales, general and administrative, factory start-up costs and purchases of property and equipment; read through content/WOLF/financials.json and checked against the filed statementbacksC50C51
- S36primarycompany filing2026-10-07the commitment letter's size, tranches, maturity, rate range, conditions precedent, warrants and the refinancing of the 2030 first-lien notesbacksC57
- S37primarycompany filing2026-11-04first-quarter fiscal 2027 revenue by product, gross margin, adjusted EBITDA, cash flow, cash and investments, debt and the second-quarter targetbacksC26C41C58
- S38primarytranscript2026-11-05Wolfspeed Q1 fiscal 2027 earnings call transcript (FMP), held November 4, 2026Loan commitment status and conditions, breakeven nearer $200M a quarter, Mohawk Valley utilisation, data-centre revenue more than doubled on a trailing basis, BorgWarner and Infineon agreements, GaN entry, Q2 guide.
Last call, I mentioned that you should think about the company as being around $200 million a quarter to be break-even. Given the improvements we are making and the favorable end market mix, we think that break-even point is a little bit lower than the $200 million.
backsC06C57 - S39primarycompany filing2026-11-0553,581,878 shares at October 30, 2026; debt principal of $1,772.0M at September 27 after $10.0M of convertible notes became 0.8M shares; 50.9M shares issuable on conversion and 4.9M on the warrant; operating cash flow of minus $84.2M and capital spending of $10.6M in the quarter; the 2026 employee stock purchase plan approved at the October 27 annual meeting; the conditional loan commitment in Note 14backsC27C28C37C56
- S40secondaryvendor data2026-11-11FMP end-of-day close for WOLF, 2026-11-10, via thesis_deltadata/cache/delta/WOLF.jsonthe 2026-11-10 close of $26.23
26.23 (2026-11-10)
backsC37