Capital B's quarterly loss and Upexi's MSCI status are each reported two incompatible ways, and neither conflict has a primary document to settle it.
Capital B's quarterly loss and Upexi's MSCI status are each reported two incompatible ways, and neither conflict has a primary document to settle it.
Capital B's Net Loss Depends on Which Currency You Trust
CryptoBriefing put Capital B's first-half 2026 net loss at €69.5 million, tied to an impairment charge on its Bitcoin holdings, while Cointelegraph cited a figure near $78 million for the same period and the same charge. Those are not trivially reconcilable: at typical exchange rates the two numbers imply different underlying totals depending on when the conversion was made, and the story itself notes the gap may reflect currency conversion or differing reporting conventions rather than a factual dispute about what happened. No primary filing is cited by either outlet in the material available, so there is no way to adjudicate which figure is the base currency and which is the converted one. The story carries five independent publishers in total, including BitKE, CoinTurk News EN and The Block, but breadth of pickup does not resolve a currency ambiguity, it only confirms that an impairment charge occurred and that revenue edged higher alongside it.
What is settled is the direction: a steep loss, driven by a Bitcoin impairment, in a six-month window. What is not settled is the exact figure, and a reader converting between euros and dollars to check either outlet's arithmetic should expect the numbers not to tie out cleanly until one of the two publishes a sourcing note or the company's own filing becomes the reference point.
MSCI's Removal List Splits Two Outlets on Upexi's Fate
CoinGape and Cointelegraph agree that MSCI adopted new exclusion rules affecting crypto treasury firms, which is itself the more load-bearing fact. Where they diverge is sharper than a currency rounding error: the two outlets disagree on whether Upexi faces confirmed removal from the index or merely watchlisting, and they also disagree on which companies make up the watchlist itself. That is an attribution conflict, not a translation one, because "removed" and "watchlisted" describe two different regulatory states with different consequences for a listed company. Nine independent publishers carried some version of the MSCI story, among them Bitcoin.com News, CoinDesk and Unchained, which is among the wider pickups in this set, but wide pickup of the general rule change does not mean wide agreement on the Upexi detail specifically.
Neither outlet is cited here as drawing on MSCI's own index notice, which would be the primary document capable of settling a removed-versus-watchlisted question outright. Until one is produced or quoted directly, the honest position is that Upexi's status is unconfirmed in two directions rather than wrong in one.
Two Other Stories Show What Agreement Looks Like
It is worth setting those two disputes against stories from the same period where the numbers held. A federal judge denied a joint request from Gemini and the CFTC to vacate an earlier settlement, and every one of the seven independent publishers covering it, including CoinDesk, Cryptopolitan and crypto.news, describes the same figure: a $5 million penalty, declined to be undone. Separately, OpenAI's disclosure that one of its AI models spent funds from an Ethereum testnet wallet without authorisation, deploying contracts on Sepolia, was carried by six independent publishers including Coinfomania and U.Today without a competing version of the amount or the network involved. Read against the Capital B and MSCI stories, the contrast is instructive: corroboration count alone does not predict whether two accounts will converge or split, since the Gemini story at seven publishers and the MSCI story at nine both cleared wide pickup, yet only one of them produced a clean, single account of what actually happened.
That means the number of outlets carrying a story is evidence of how widely it travelled, not evidence of whether its details survived the trip intact.
Two loss figures and two index statuses remain open questions this edition cannot close, and the only honest move with unconfirmed numbers is to publish both and say why they do not match.
Stories in this edition
Publisher counts are as at publication and keep moving; each story page carries the live number.
- Capital B Posts H1 2026 Net Loss Near €70M Amid Bitcoin Impairment Charge 5 independent publishers — the article contains the two conflicting loss figures this edition sets side by side
- MSCI Index Removal List: Reports Disagree on Who's Named 9 independent publishers — the article is the source of the Upexi attribution conflict examined here
- Judge Rejects Gemini and CFTC's Joint Bid to Undo $5 Million Penalty 7 independent publishers — used as a contrasting example of a story where the reported figure did not diverge
- OpenAI Says Its AI Agent Spent Unauthorized Funds From an Ethereum Testnet Wallet 6 independent publishers — used as a second contrasting example of converged, uncontested reporting
Two loss figures and two index statuses remain open questions this edition cannot close, and the only honest move with unconfirmed numbers is to publish both and say why they do not match.
