Wolfspeed (WOLF) Earnings Aug 19: Up 29%, Sales Falling, And One AI Number With No Denominator

TL;DR

• Wolfspeed reports fiscal Q4 and full-year 2026 on Wednesday, August 19, with the call at 5:00pm ET.
• The stock closed at $31.51 on August 13, up 29.4% from $24.35 on August 3. The biggest session was August 7, +19.1%, after a silicon carbide partnership with LITEON aimed at 800 VDC power for hyperscale AI data centers.
• Revenue has fallen three quarters running: $196.8m, $168.5m, $150.2m. Management guided FQ4 to $140-160m, so the midpoint is another decline and roughly 24% below the $197.0m Wolfspeed did in the year-ago quarter.
• Gross margin was -27% GAAP in Q3 and the company said it stays negative in Q4. Adjusted EBITDA was -$61.7m.
• The number the entire rally rests on has never been published in dollars. Wolfspeed said AI data center revenue grew "approximately 30%" sequentially and has not sized the base. That disclosure, not the headline revenue, is what August 19 is worth watching for.

When Does Wolfspeed Report Earnings?

Wednesday, August 19, 2026, with the conference call at 5:00pm ET, hosted by CEO Robert Feurle and CFO Gregor van Issum. Wolfspeed confirmed the date in an August 5 release. It covers fiscal Q4 and the full fiscal year, which ended in late June. The rest of that week's slate is in the earnings calendar.

The Board

Two facts that both check out, and they point opposite ways.

Why The Stock Ran 29% In Eight Sessions

Silicon carbide is the switching material for high-voltage power conversion, and the AI build-out has spent 2026 discovering that it needs a great deal of it. Wolfspeed announced a partnership with LITEON on August 6 to qualify its MOSFETs into 800 VDC power systems sold to cloud providers. The stock closed +9.5% that day and +19.1% the next.

The rest of the summer points the same way. Wolfspeed launched a 10 kV SiC MOSFET aimed at grid and AI infrastructure, released a fifth-generation MOSFET family it claims lands the lowest on-resistance in the 1200 V and 750 V classes, and stood up a data center team with a Silicon Valley office to sit closer to hyperscalers and ODMs. In July it sued Navitas Semiconductor for patent infringement.

The sell side has not followed. Consensus sits at a hold-to-sell rating with a $30.00 twelve-month target, below where the stock closed. That is the sell-side average; this site carries no target on WOLF.

The 30% With No Base

Here is my problem with the rally. In the Q3 release, Wolfspeed disclosed that AI data center revenue grew "approximately 30%" sequentially. It did not say 30% of what.

Total revenue that quarter was $150.2m. If the AI line is $10m, then 30% growth is $3m of incremental sales. If it is $50m, 30% growth is $15m and something real is happening to the mix. Both readings are consistent with everything Wolfspeed has published, and they support completely different valuations. A growth rate with no denominator does the same work as a lock-up headline that quotes a dollar value and skips the float, which is the shape I picked apart in the Hut 8 lease fine print.

So the disclosure I want on August 19 is one line: AI data center revenue, in dollars, for the quarter. If management gives it, the segment can be valued. If they give another percentage, I would read that as a base too small to help them.

A Fourth Straight Quarter Of Shrinking

The revenue line is doing something the narrative is not. $197.0m in the June 2025 quarter, then $196.8m, $168.5m and $150.2m through fiscal 2026. The guide is $140-160m.

Wolfspeed has explanations and they are reasonable ones. Customers pulled orders forward from the Durham fab before its device production wound down, some second-sourced away during the bankruptcy, and EV demand has been weak. Durham now runs as a materials site. All of that is true, and it still nets out to a company selling less each quarter while the market prices it as an AI winner.

Gross margin is the harder number. -27% GAAP and -21% non-GAAP in Q3, guided negative again. Operating cash flow was -$84m against $5.0m of net capex, so the burn is coming from operations rather than from building anything. Cash and short-term investments were $1.2bn, which buys a few years at that rate.

Emergence Doubled The Share Count

Wolfspeed filed Chapter 11 on June 30, 2025 and emerged on September 29, cutting debt by about 70%, pushing maturities to 2030 and dropping annual cash interest by roughly 60%. Old shareholders were nearly wiped: around 1.3m new shares at an exchange ratio near 0.0083.

Emergence issued 25.8m new shares. Since then Renesas took 16,852,372 shares after CFIUS clearance, 871,287 went to pre-petition holders, about 1.5m came from note conversions, and a March raise added roughly $96.9m of stock and pre-funded warrants. Shares outstanding are now about 52.0m. That is a doubling in under a year, and it is not finished: Renesas holds a warrant for 4,943,555 shares, and there are $379m of 3.5% converts due 2031 plus Renesas's 2.5% converts sitting outside the count.

Which matters for the multiple everyone quotes. At $31.51 the market cap is $1.64bn. Add the $922.2m of long-term debt and $798.3m of convertible notes from the Q3 balance sheet, subtract the $1.2bn of cash, and enterprise value is about $2.16bn. Against a fiscal 2026 revenue run of roughly $665m on the guide midpoint, that is about 3.2x sales for a business with negative gross margin. Call it the floor, since the warrant and the converts are not in the share count.

What Would Change My Mind On August 19

Three things, in order of how much they would move me:

• A dollar figure for AI data center revenue. Anything above roughly $40m of the quarter and the re-rating has a foundation under it.
• A gross margin path with a date on it. Negative margin is survivable while the balance sheet is repaired. Negative margin with no stated crossover quarter is not.
• FY27 revenue guidance above $700m. That would break the four-quarter decline and say the second-sourcing losses have stopped.

I am not convinced they deliver the first one. Companies that have a good number tend to print it, and Wolfspeed has now had two quarters to put a dollar sign in front of this one.

The Options Angle

I could not source a live WOLF chain tonight, so nothing below is priced off a real quote and I am quoting it against realised moves instead. That is still scoreable.

What is specific to this name: with about 52.0m shares out, WOLF has been realising enormous single sessions on press releases alone. In the eight sessions to August 13, with no earnings in the window, it printed +11.1%, -6.9%, +9.5%, +19.1%, -10.9%, +2.9%, +5.2% and -0.6%. Four moves near or above 10% on partnership news.

The house reflex is to look at an earnings straddle and call it expensive. July taught the opposite lesson here: realised beat implied repeatedly, and the losing calls were the ones that talked readers out of buying volatility. On a name already moving 10-19% on a press release, an earnings strangle priced much under 15% of spot looks fair rather than rich to me. The distribution is bimodal on one disclosure, which favours owning options over picking a direction.

The trade I do not want is a covered call. Writing upside on a 52m-share float into a binary disclosure collects a small premium and caps the only outcome that pays.

Trade log

#	Stance	Structure	Strikes and expiry	Cost or credit	Spot at writing	Implied move	Breakeven
1	Long volatility	Long strangle	Nearest strikes above and below $31.51, expiry after Aug 19	No live chain sourced; constructive at any cost under ~15% of spot	$31.51, close Aug 13	Not sourced; realised ran 10-19% on news in Aug	Needs a move above ~15%
2	Pass	Covered call	Any strike, expiry after Aug 19	Premium not sourced	$31.51, close Aug 13	Not sourced	Caps the AI-disclosure upside

Row 1 gets scored against the actual August 19 to August 20 move versus the 15% threshold it was written on. Row 2 gets scored on the whole position, per the house rule, not on the premium alone.

The One-Line Read

A shrinking company priced as a growing one, on a growth rate nobody has sized. Give me the AI revenue in dollars on August 19 and I can value it. Another percentage and it stays a story stock.