Bullish prints a $280m loss and a tripled EBITDA in the same release. The two stories are not the same story.

On 13 August 2026, Bullish, the cryptocurrency exchange operator, told two opposite stories in the same quarterly release. The reported net loss for the second quarter came in at $280 million. Adjusted EBITDA, the figure the company itself highlights, more than tripled. Subscription, services and other revenue hit a record $62.7 million. The stock moved in both directions across the next 36 hours: roughly 10% higher on the day of the print, then softer the following session.

The two lines are pointing in opposite directions, and which one you privilege depends on what you think the print is actually measuring. That the same release can land as a loss event in one headline and a beat in another is the structural fault line this piece is going to walk down.

Two reads of the same release

CoinDesk's coverage, published on 13 August 2026, leads on the loss. Its headline frames the quarter as a $280 million net loss reported against a slower trading backdrop, with adjusted revenue of $92.6 million and refined full-year guidance. The story's pitch is that the print is the cost of running a venue through a soft tape.

Cointelegraph, covering the equity tape the same day, leads on the beat. Its headline frames the same release as a roughly 10% share-price jump, with adjusted EBITDA more than tripling and subscription, services and other revenue hitting a record. The story's pitch is that the operating business has stepped up a gear regardless of the headline number.

Both stories describe the same print. The split is editorial: one treats the GAAP result as the headline, the other treats the adjusted result as the headline. The distinction matters because institutional readers price on different lines depending on mandate. A mark-to-market loss is a different thing from an EBITDA result, and the wire coverage on 13 August is, in effect, two different desks reading the same filing through two different lenses.

Monexus assessment: the more useful read for an institutional reader is the adjusted one, because the adjusted result is what strips accounting noise off the underlying franchise. The $280 million line is the headline; the EBITDA and services lines describe the operating engine. Neither is wrong, but the wire split suggests the distinction is not yet settled in coverage.

The one-day pop and the next-session drift

Cointelegraph's read is reflected in the equity tape on the day of the release: shares gained around 10%. That is the bull case taking a turn at the wheel. The reading on the day of the print is that the operating story has surprised to the upside, that services have scaled to a record even with transaction volumes softer, and that the firm's full-year guidance refinement is incremental on top of an already constructive picture.

An explainer published on 14 August 2026 by investing.com under the headline "Why is Bullish stock sliding today?" reflects a different read in the subsequent session. The available source items do not specify the article's exact framing beyond its headline and subject, so the precise editorial case it makes about the next-session move is not entailed by the cited evidence. What the headline tells us directly is that the tape was softer in the second session, and that this softer tape was treated as newsworthy enough to warrant an explainer the day after the print.

The honest read of the two-day tape is that the same release produced two different reactions in two consecutive sessions. Our assessment: that is most naturally explained by algorithmic and short-term retail flows re-pricing the headline number in the second session after the initial pop on 13 August absorbed the operating beat. The evidence does not specify the precise mechanism; that interpretation is this publication's read, not a sourced claim, and should be treated as such.

What the services line proves, and what it does not

The $62.7 million subscription, services and other revenue figure is the part of the print that does not depend on spot volatility. CoinDesk and Cointelegraph both note the figure is a record. It is contract-style revenue of the kind that gets signed for a year and renewed quietly, and it is the line that lets the EBITDA beat stand up against a softer transactional environment.

Monexus assessment: if the services line has reached a record in a soft tape, then the franchise is being carried by recurring contracts at a moment when transaction revenue is mechanically weaker. The relevant comparison is not last quarter's services number but last year's, and the cited evidence establishes only that the current quarter is a record, not the rate at which the line is scaling. We can say from the cited evidence that the line has expanded to a record; we cannot say from the cited evidence what the run-rate implies for revenue mix or contribution margin in subsequent quarters, and we are not going to estimate those here.

A separate observation, drawn from CoinDesk's framing of the period as a slower one for crypto trading generally: the venue-side business is currently running through a quieter tape, and the adjusted picture is, by construction, a picture of the franchise with the cyclical element dampened.

What the evidence does and does not support

The cited sources establish the following without ambiguity. Bullish reported a $280 million net loss for Q2 2026 (CoinDesk). Bullish's adjusted revenue was $92.6 million (CoinDesk). Bullish refined full-year guidance alongside the release (CoinDesk). Subscription, services and other revenue reached a record $62.7 million (CoinDesk, Cointelegraph). Adjusted EBITDA more than tripled (Cointelegraph). Bullish shares gained around 10% on 13 August 2026 (Cointelegraph). An investing.com explainer was published on 14 August 2026 asking why the stock was sliding that day.

The cited sources do not specify, and this article does not establish, the following: the precise composition of the $280 million loss; the geographic split of the services book; the segment-by-segment cost base; the annualised run-rate implied by the services figure; the specific reason for the next-session move beyond the headline direction; or the firm's full-year guidance range in numerical terms. These are not gaps in the argument; they are the boundaries of what the cited evidence entitles us to say.

What to watch in the next 90 days

Three prints would resolve the operating-versus-headline tension. First, the next monthly volume disclosure: whether spot volumes re-accelerate from current levels or stabilise where they are will determine how much of the franchise is being carried by services alone. Second, the next quarterly release on services revenue: a second consecutive record would confirm the line as a structural shift rather than a one-quarter print. Third, the next round of full-year guidance: whether the refinement on 13 August is the first step in a multi-step upward revision or a single adjustment will tell us how Bullish's management reads the second half.

The bull case, as Cointelegraph framed it on 13 August, is that an exchange has tripled its adjusted EBITDA while scaling subscription and services revenue to a record. The bear case, as CoinDesk framed it the same day, is that the venue's headline number printed red in a quieter tape. The two-day tape reflects both reads in sequence. What the next 90 days of disclosure resolves is which of the two is the steadier signal.

Desk note: Monexus treats the EBITDA-versus-GAAP gap as the substantive story, not the headline loss in isolation. CoinDesk's coverage on 13 August led on the loss; Cointelegraph's coverage on 13 August led on the beat; investing.com's explainer on 14 August addresses the next-session drift. The cited sources do not specify the precise composition of the $280 million loss, and this article does not estimate it.

Wire provenance

This editorial synthesis draws on the following public wire/social posts:

• https://cointelegraph.com/news/bullish-shares-surge-as-q2-adjusted-revenue-surges-62
• https://www.coindesk.com/business/2026/08/13/bullish-reports-usd280-million-q2-net-loss-as-crypto-trading-slows
• https://www.investing.com/news/stock-market-news/why-is-bullish-stock-sliding-today-93CH-4861364
• https://www.investing.com/news/company-news/braskem-q2-2026-slides-ebitda-surges-on-supply-shock-spreads-to-normalize-93CH-4861430