Wolfspeed’s headline number is famous and almost universally misread. Roughly 24 million shares of the silicon carbide maker sit short, close to half the shares outstanding, and the crowd reads that as a squeeze waiting to happen. The capital structure says otherwise. Wolfspeed’s post-bankruptcy convertible notes convert into about 51.8 million new shares at strikes of $12.23, $18.35 and $20.14 – every one deep in the money at the 27 August close of $26.95. The short base is smaller than the hedge arbitrage desks would need against those bonds. That reframes the whole stock. Our bull case is $53, built on device revenue compounding and 200mm fabs filling up. Our bear case is $13, built on a company still losing money on every wafer. Both are arithmetic, not adjectives. This is analysis, not investment advice.
Here is the take nobody else is making. Almost every WOLF write-up treats short interest and dilution as two separate risk bullets. They are the same bullet. A convertible arbitrage desk buys the bond and shorts the stock; reported short interest is the mechanical output of that hedge. Wolfspeed’s own 10-K shows why desks would bother: the convertible notes carry $849.7 million of principal but a fair value of $2.49 billion, because the equity option inside them is worth roughly three times the debt. When a company’s bonds are priced as equity call options, the squeeze narrative and the dilution narrative collapse into one trade – and the 34.7 million free float stops being the relevant denominator. The relevant denominator is 106.7 million.
Key facts
- Wolfspeed reported Q4 and full-year fiscal 2026 results after the close on 19 August 2026; the 8-K was filed at 16:40 Eastern that day and the 10-K followed on 20 August (SEC EDGAR).
- Q4 FY2026 revenue was $149.6 million, down 24.1% year on year. GAAP gross margin was -25% and non-GAAP gross margin -20%, against -1% non-GAAP a year earlier (Wolfspeed Q4 FY2026 release, 19 August 2026).
- Adjusted EBITDA was -$62.4 million and free cash flow -$60.9 million. Cash, equivalents and short-term investments stood at $1,088.6 million at 28 June 2026 (same source).
- Q1 FY2027 guidance: revenue of $140-160 million, non-GAAP gross margin still negative, non-GAAP operating expenses of $62-66 million.
- Wolfspeed emerged from Chapter 11 on 29 September 2025 and adopted fresh-start accounting. Total debt was cut by approximately 70% versus pre-emergence levels; principal outstanding was $1,782.0 million at 28 June 2026 (Wolfspeed FY2026 Form 10-K).
- Shares outstanding were 52,995,396 as of 13 August 2026 (10-K cover page). Screener data retrieved on 28 August 2026 from Finviz showed 24.00 million shares short – 69.15% of a 34.71 million free float, 4.53 days to cover.
- FinanceFeeds reported on 15 July 2026 that Wolfspeed topped the US short interest rankings at 46.09% of shares on loan, per S&P Global Market Intelligence Securities Finance. Six weeks on, the number has barely moved.
What actually happened on 19 August
Wolfspeed’s fiscal year ends in late June, so the 19 August print covered both the June quarter and the full year. Revenue of $149.6 million landed at the midpoint of guidance. Underneath it, the two product lines went in opposite directions. Power Products – the MOSFETs and modules – delivered $106.3 million against $118.6 million a year earlier, a 10.4% decline. Materials Products, the silicon carbide substrate business, fell to $43.3 million from $78.4 million, down 44.8%.
Stitch the 13-week predecessor stub to the 39-week successor period and full-year fiscal 2026 revenue comes to roughly $665.1 million against $757.6 million in fiscal 2025, a 12.2% decline. Fresh-start accounting means the two periods are not formally comparable and Wolfspeed says so; we add them only to get a run-rate, and flag the method. On that basis Power Products grew about 10.3% for the year while Materials Products fell about 39.3%.
That split is the industrial story. Chinese suppliers have collapsed the price of the raw wafer, the commodity end of silicon carbide, so Wolfspeed is trying to stop selling the commodity and sell the device instead. Fiscal 2026 is where that pivot shows up, painfully.
Chief executive Robert Feurle framed the year around exactly that: “We continued to expand our device business, highlighted by strong growth in AI data center applications and the launch of our fifth-generation SiC MOSFET. These achievements strengthen our technology leadership and confidence in our long-term growth opportunities” (Wolfspeed Q4 FY2026 release, 19 August 2026). AI data centre revenue more than doubled year on year in fiscal 2026 and rose roughly 20% sequentially in Q4, though the company calls the opportunity “moderate but growing” – more honest than most AI-adjacent disclosures.
The market’s verdict was blunt. The stock closed at $29.09 on 19 August, then finished 20 August at $26.35, a 9.42% fall on 9.72 million shares. From the 17 August close of $34.96 to the 24 August close of $24.76, WOLF lost 29.2% in five sessions.
The number nobody quotes: 106.7 million shares
Screeners report Wolfspeed’s market capitalisation at about $1.43 billion. That figure uses 52.2 million shares and is the most misleading number attached to this stock.
The 10-K debt note lists three convertible instruments outstanding at 28 June 2026: the 2.5% second-lien non-Renesas notes, $267.1 million of principal remaining, converting at $12.23; the 2.5% second-lien Renesas notes, $203.6 million at $18.35; and the 3.5% 1.5-lien notes issued in March 2026, $379.0 million at approximately $20.14. Add 2.0 million pre-funded warrants from the March equity placement.
Against a $26.95 share price, all three strikes are in the money. Convert every one and Wolfspeed issues roughly 51.8 million new shares on top of the 53.0 million already outstanding. Fully diluted, the count is about 106.7 million shares, or 2.0 times the reported figure.
Run the enterprise value both ways. On basic shares plus all debt at carrying value less cash, EV is roughly $2.12 billion, or 3.19 times trailing revenue. Treat the converts as equity – which is what their strikes say they are – and fully diluted equity value is $2.88 billion; add the $932.3 million of genuinely non-convertible debt, subtract $1.09 billion of cash, and EV is roughly $2.72 billion, or 4.09 times trailing revenue. That is a 28% valuation gap hiding inside a screener field.
The conversion is already underway. Chief financial officer Gregor van Issum noted in the same release: “During the fourth quarter, our capital structure further improved as holders of $46 million of our convertible notes exercised a voluntary conversion of their debt to equity.” Cumulatively, $64.3 million had converted into 5.3 million shares by 28 June 2026.
The bull case: $53
The bull path does not require a demand miracle. It requires the device business to keep compounding while fixed costs stay where they are.
Start with Power Products at roughly $456.5 million for fiscal 2026 on the combined-period basis. Grow that at 20% a year for three years and it reaches about $789 million by fiscal 2029; assume Materials stabilises and recovers to around $230 million as Siler City runs at proper utilisation. Fiscal 2029 revenue: approximately $1.02 billion.
The margin leverage is the point. Wolfspeed closed its 150mm device fab in Durham during fiscal 2026 and now makes all power devices at the 200mm Mohawk Valley fab in Marcy, New York. The 10-K says underutilisation costs at Mohawk Valley and Siler City will persist “until market demand for our products meets or exceeds our production capacity” – a fixed cost already in the run rate. At $1.02 billion of revenue through fabs built for materially more, a 35% non-GAAP gross margin gives roughly $357 million of gross profit against about $270 million of operating expense, so positive operating income and roughly $217 million of EBITDA after depreciation.
Apply 6.0 times enterprise value to sales – rich for hardware, but far below the 12.6 times the market itself paid at the May 2026 peak – and enterprise value is about $6.1 billion. All converts have converted by then, leaving $932.3 million of senior secured and second-lien non-convertible debt. Add roughly $750 million of cash, divide by about 112 million fully diluted shares including equity awards, and you get $53, about 97% above spot.
Three things have to go right: 200mm yields hold, AI data centre design wins turn into volume rather than samples, and the substrate price war stops getting worse. All three are observable quarter by quarter.
The bear case: $13
The bear path is simpler because it requires nothing new to happen. Annualise management’s own Q1 FY2027 guidance: about $600 million of revenue with non-GAAP gross margin still negative. That is a company selling $600 million of product at a loss while carrying $1.78 billion of debt principal and $270 million of annual operating expense – roughly -$200 million of adjusted EBITDA a year and free cash flow closer to -$250 million.
Value it at 2.4 times forward enterprise value to sales – fair for a sub-scale, negative-gross-margin components business whose two largest customers are 38% of revenue – and enterprise value is about $1.43 billion. In this scenario only the $12.23 notes stay in the money, so the share count is roughly 76.8 million rather than 106.7 million and the $18.35 and $20.14 notes stay debt. Equity value is then approximately $999 million, or $13 per share – about 52% below spot, and a retest of the $13.49 intraday low printed on 5 February 2026.
Note the asymmetry in the capital structure itself: dilution is self-limiting on the way down, because notes out of the money stay as debt, which means the downside carries more leverage rather than less. That is the opposite of the intuition most investors bring to a converted balance sheet.
What the market is actually pricing
Wolfspeed’s post-emergence equity has 229 trading sessions of history, beginning 30 September 2025. Within that clean series the intraday high is $80.82 on 22 May 2026 and the intraday low is $13.49 on 5 February 2026. Spot is 66.7% below that high and up 54.8% from the 31 December 2025 close of $17.41.
Treat vendor 52-week fields with suspicion here. On 28 August 2026, stockanalysis.com reported a 52-week low of $8.05 while Finviz reported a 52-week high of $421.47. Neither appears anywhere in the traded post-emergence series; both are artefacts of chaining pre-bankruptcy and post-emergence equity across a corporate action that wiped out the old shareholders. Any “down 98% from its high” framing on WOLF is measuring a security that no longer exists.
The volatility is the real signal. On log returns from that series, annualised realised volatility is 130.3% over the last 30 sessions, 143.1% over 90 and 118.2% across the whole sample. The average absolute daily move is 5.92%; 45.9% of sessions moved more than 5% and 18.3% moved more than 10%.
Put a one-standard-deviation twelve-month band around $26.95 at 118% volatility and you get roughly $8 to $88. Our $13 to $53 range sits comfortably inside it. On this stock, a 97%-up and 52%-down band is not a wide call. It is a narrow one.
The runway, and how we calculated it
Wolfspeed held $1,088.6 million of cash, cash equivalents and short-term investments at 28 June 2026. The number that matters more is the covenant: the New Senior Secured Notes indenture requires at least $350.0 million of unrestricted cash in secured accounts on the last day of every calendar month. Usable cash is therefore about $738.6 million, not $1.09 billion.
Against Q4’s free cash flow of -$60.9 million, that is 12.1 quarters, or roughly three years. But Q4 property and equipment spending was only $5.0 million net of incentive reimbursements, which flatters the run rate. Use the successor period’s gross capital expenditure of about $24.3 million a quarter alongside operating cash outflow of $54.1 million and quarterly burn is $78.4 million – 9.4 quarters, or about 2.4 years. Add the $200.9 million of investment tax credit receivables still on the balance sheet and it returns to roughly 12 quarters.
So: approximately 2.4 to 3.0 years of usable cash above the covenant floor at the current burn. The first maturity is 23 June 2030 on the senior secured notes; everything else runs to 2031. The 10-K carries no going-concern qualification. A second bankruptcy is not the near-term risk. Dilution is.
Bull versus bear at a glance
| Driver | Bull case ($53) | Bear case ($13) |
|---|---|---|
| Revenue path | Power +20%/yr to ~$789m, Materials ~$230m, FY29 ~$1.02bn | Flat at the guided run rate, ~$600m in FY27 |
| Gross margin | Turns positive, reaches ~35% as 200mm fabs fill | Stays negative through FY27, as guided |
| Multiple applied | 6.0x EV/sales on FY29 revenue | 2.4x EV/sales on FY27 revenue |
| Share count | ~112m, all converts convert, debt falls to $932m | ~76.8m, only the $12.23 notes convert |
| Cash position | Burn stops before the covenant floor bites | Usable cash drains toward $350m by 2028 |
| Short base | Arb hedges unwind into conversion, not buying | Same hedges persist, capping any rally |
| Implied move | +96.7% from $26.95 | -51.8% from $26.95 |
Disconfirmation triggers
These are the specific, checkable conditions that would break each case.
The bull case fails if:
- 1. Non-GAAP gross margin is still worse than -10% by the fiscal Q3 2027 report. Q4 FY2026 was -20%; a path to 35% needs visible improvement within two quarters, not four.
- 2. Power Products revenue declines year on year for two more consecutive quarters after Q4 FY2026’s 10.4% fall – that would mean the device pivot is failing too.
- 3. Quarterly revenue guidance is cut below $140 million, the bottom of the current Q1 FY2027 range.
- 4. Wolfspeed issues equity below the $20.14 conversion price, signalling the March 2026 refinancing did not solve the funding problem.
The bear case fails if:
- 5. Non-GAAP gross margin turns positive in any quarter of fiscal 2027. That single event invalidates the “loses money on every wafer” premise.
- 6. AI data centre revenue is disclosed as a named, material line rather than a growth-rate mention – a dollar figure above roughly 15% of revenue changes the multiple.
- 7. Quarterly free cash flow burn falls below $30 million, roughly half the Q4 rate, pushing the covenant-adjusted runway beyond five years.
- 8. A strategic investor takes a stake, or a supply agreement lands with a named hyperscaler. Renesas already holds a board designation right and the August 8-K confirms it retains the right to re-designate a director.
Frequently asked questions
Did Wolfspeed go bankrupt, and is the current stock the same company?
Wolfspeed filed a prepackaged Chapter 11 on 30 June 2025 and emerged on 29 September 2025, adopting fresh-start accounting. The current NYSE-listed shares are post-emergence equity; the pre-bankruptcy shares were largely wiped out. This is why the traded price history only runs back to 30 September 2025, and why any long-term chart of WOLF mixes two different securities.
Is Wolfspeed’s short interest still elevated?
Yes. FinanceFeeds reported on 15 July 2026 that Wolfspeed led US short interest rankings at 46.09% of shares on loan, per S&P Global Market Intelligence. Finviz data retrieved on 28 August 2026 showed 24.00 million shares short – 46.0% of shares outstanding, 69.15% of the free float. Options Trading Report cited a 932.81% annualised borrow fee on 15 August 2026, without naming a securities-lending provider. The short base has not decayed.
Does the high short interest mean a squeeze is likely?
Not mechanically. Wolfspeed’s converts turn into roughly 51.8 million shares at strikes below the current price, and convertible arbitrage strategies short the underlying as a hedge against those bonds. On our estimate, even a partial delta hedge across the stack would account for most of the reported short interest. Hedged shorts do not panic the way directional shorts do, and can settle by converting rather than buying.
When does Wolfspeed’s debt actually come due?
The New Senior Secured Notes mature on 23 June 2030. The second-lien non-convertible PIK toggle notes and the 2.5% convertible second-lien notes mature on 15 June 2031, and the 3.5% 1.5-lien convertible notes on 15 March 2031. There is no maturity wall before 2030. The nearer constraint is the $350 million minimum monthly cash covenant on the senior notes, not a repayment date.
How does Wolfspeed compare with other semiconductor turnarounds?
The closest structural analogue is Intel, whose bull and bear case also hinges on fab utilisation rather than demand; the difference is that Intel can absorb years of underutilisation and Wolfspeed has about three. The AI power theme behind its data centre pitch runs through the same season as Nvidia’s $96.2 billion quarter and Marvell’s record revenue – both fell on guidance, not results, as Wolfspeed did on 20 August.
What is the single most important number to watch next quarter?
Non-GAAP gross margin. Revenue is guided within a $20 million band and is unlikely to surprise much. Margin is where the 200mm transition either shows up or does not. Q4 FY2026 was -20%; the direction of that number in the Q1 FY2027 report decides which of these two cases is live.
What we are watching
Wolfspeed enters fiscal 2027 as a genuinely different company from the one that filed in June 2025: 70% less debt, no maturity before 2030, device production consolidated at 200mm, and a securities class action over the Mohawk Valley disclosures dismissed on 12 August 2026. What it lacks is a positive gross margin, and everything else is downstream of that.
The stock is priced for the transition to work eventually and for the dilution to happen soon. Both are probably right. The mistake to avoid is reading the short interest as a coiled spring when the capital structure says it is mostly a hedge, and accepting a $1.43 billion market capitalisation when the honest figure is closer to $2.9 billion. Get those two right and the $13-to-$53 band stops looking like a guess. None of this is investment advice; Wolfspeed is a loss-making company running at 118% annualised volatility, and position sizes should reflect it.