The Mispriced Assets Super Bowl

Last night around 1am EST I was screenshotting ERN pages off the terminal into an email to myself, which passes for pregame ritual. Three of my positions report tomorrow and all three inside the same fifteen minutes. Groupon, which is 18% of the model portfolio and >50% of my net worth. Then Corsair. And then StepStone, one of my favorite shorts. Oh, and 3 other names… the earnings gods gave us a doozy.

Corsair and StepStone both take questions tomorrow at 5:00. Groupon posts its release after hours and then nobody from the company says a word until the call at 8am Friday. The opportunity for volatility from the print to the commentary is yuge.

The number Groupon has to answer for is North America Local billings. Five quarters of growth: 11, 20, 18, 9, 2. The 2 is Q1. On the May 7 call management said April had already turned back up, managed and organic channels both, and tomorrow we get to see what that means. Everything else in the release is commentary.

What do I actually want? $17 or $18 million of real EBITDA, severance added back, against a $13-15 million guide. Consensus is $14.55M, and the entire range of estimates runs $14.3 to $14.8, three analysts huddled inside half a million dollars of each other. Revenue guided $126-128M, consensus $127 million, dead midpoint. These estimates are so tight that it feels like homework copying!

GRPN Q2 revenue consensus, off the terminal.

And the EBITDA screen: $14.55M mean

A year of Q2 EPS revisions, one direction.

The thesis is frequency, I wrote a whole piece on it in July. 16.2 million actives buying 2.3 times a year, and each turn of frequency is worth roughly $100M of EBITDA, more than the company earned in all of 2025. Right now actives are up 5% and units are down 5%. Sophisticated bulls want to see units up, and new users is a bonus. It’s pretty obvious the paid channel is working, and we are before the redesign, so this is really a 2H thing. A simple revenue beat would be nice.

The restructuring lands in this print too. Up to 400 positions, $20-25 million annualized, roughly $5M net this year. And a quirk worth knowing before the number crosses the tape: Groupon doesn’t back severance out of adjusted EBITDA. Most companies would consider that a reasonable adjustment (definitely one-time). Last quarter about $2 million of it sat inside the reported figure, this quarter carries the heavy end of the $7-13 million of charges, so whatever adjusted EBITDA prints, add the severance back to get the real one, and opportunity lies in the delta here. A reported 14 with three or four million of severance inside it is a 17, 18 underlying. That would be a fantastic buy-the-dip on an algo/short-seller flash crash. GAAP will be ugly, it was ugly in May, minus 32 cents against minus 3 expected, and the stock closed up 18% that day anyway. That number doesn’t really matter in a transition.

Eight quarters of GRPN EPS against consensus. The stock only cared Q3 2025.

And the setup. 13.4 million shares short at the July 15 settlement, up from 11.96 million at June 30. Fintel says SI is 75% of the float, Benzinga 60%, so it’s nine to twelve days to cover on any of them. Options price a 14% move, which is about $23.50 to $31.50 which should be support and resistance. Past 31.5 and it gets exciting.

The vol surface into the print. Front month north of 100, normal again by January 2027.

My targets are $63 base in a year, $90 in two years. Tomorrow doesn’t have to prove the model, and Friday’s commentary will be more useful for a fine-tune.

I own Corsair from $5.60 in February. What I wrote then was that the stock was priced like a business that’s dying while the financials said it was transforming, and since then it mostly has: record 32.7% gross margin in Q1, EBITDA of $35.8 million against a guide topping out at $15.5M, the stock at $11.37 and the 52-week high is $13.10. The one-standard-deviation upward earnings move gets you almost exactly to the 52-week high.

Two straight demolitions after the 2024 misses that buried the stock.

Now the crucial, which is in the segment bridge of the Q1 10-Q where analysts will be paying attention. Components gross margin was up 670 basis points, and 520 of those were memory price increases and mix. Another 140 was from cancelled promotions as demand was too strong (!) to bother discounting. Translation: Corsair bought DRAM cheap before the AI buildout repriced it and spent the quarter selling appreciated inventory into a rising market. Great inventory management!

A tailwind made of inventory has an expiration date, and only the company knows the date. That’s tomorrow’s whole question. Craig-Hallum downgraded in June on exactly this, memory costs eating second-half margins, and I don’t dismiss it. What breaks the bear case on CRSR 0.00%↑ is if demand stays strong despite the price increases in this segment.

The guide is $295-320M of revenue and $12.5-15.5M of EBITDA. Consensus sits at $310.6M and $14.33M, upper half of both, nine estimates and nine estimates. The street leans beat. In-line probably sells off. I want gross margin over 31 with a bridge I can follow, and I want the full year raised, because $1.33-1.47B and $100-115M haven’t been touched since May and the first half earned it.

CRSR Q2 revenue consensus against the $295-320M guide.

The EBITDA screen.

The GAAP line the value screens still see.

Also on my list for the call: Logitech beat its June quarter by 52% on EPS last week and named gaming as the driver, and Logitech is the comp I valued Corsair against in February when it traded at a 50-87% discount on everything. A double later, the discount has narrowed but not closed. And the AI workstation. Management put a $22 billion market on it in May, said prosumer demand was starting to ramp. I’d really like to hear commentary on that.

From the February piece.

The put went exercisable June 30. Tomorrow StepStone reports the quarter that ended June 30.

If you missed the piece in May: four executives hold profits interests in StepStone Private Wealth, with a forced buyout the company has to fund, payable up to 75% in newly issued stock, at up to 20 times the unit’s earnings. When I wrote it the liability estimate had gone from $664 million to somewhere between $2.2 and $3 billion in nine months, and it grew because the business performed. The flywheel and the noose are the same rope. That sentence is doing the same work it did in May.

A word on the theater around this name. StepStone earnings calls are a clown show. The sell side only started asking the right questions after a short report showed up, cough, and the business itself runs on unrealized gains, incentive fees booked against marks nobody has sold into. When that unwinds it is ugly for investors, SPCX showed everyone how that movie ends. Management is promotional as hell and will have spin ready for whatever this quarter says. I don’t know which spin. I know I’m shorting any pop it produces.

Since the window opened, five weeks of nothing. No 8-K, no exercise notice, no amendment. I’ve gone back and forth on what the quiet means. It fits executives who are waiting. It fits just as well a notice already delivered privately with the disclosure parked for the 10-Q. Tomorrow’s filing updates the liability number either way, that part isn’t optional.

My reading order, in the release and then the Q: redeemable non-controlling interest, and what the put revalued to. Control-F exercise, amendment, negotiation. SPRING flows, and the marks-to-realized ratio, which was running about 300 to 1 when I counted in May. Then fee-related earnings against actual cash, because nine months of operating cash flow lately covered about 40% of what went out as dividends and distributions.

Other names reporting: HEPS (majority owned by KSPI), WRBY, and TRIP. Will be watching each but nothing smart to say on any of them.

Not financial advice. The author is long GRPN and CRSR, and short STEP via shares and puts. Figures are approximate and drawn from company filings and releases, the Q1 2026 earnings calls (May 2026), Bloomberg consensus and options data, FINRA short interest settlements, and market data as of August 5, 2026; Q2 figures are street estimates and implied-move scenarios are illustrative. Do your own research and understand your downside before considering your upside.