MicroAlgo enters the second half as a WiMi-controlled Shenzhen algorithm shop whose public equity is no longer a bet on advertising software. The first-half print showed a still-profitable services book sitting underneath a treasury that just destroyed more value than the operating company earned in a full year. That is the debate: whether the reported cash pile is a floor for minority holders or a balance-sheet item the market has already written off. Parent control, dual-class voting, and mainland capital locks sit between the Class A quote and any residual claim on that cash.
Services revenue in the first half fell to $20 million from a year-ago level near $26 million as mainland advertising budgets kept tightening. Operating income still printed, because cost of services and research spend came down with the book. The entire net loss of $34 million is an investment-book event. Realized and unrealized marks on short-term holdings erased more than a year of prior net profit in six months. Combined cash and short-term investments still exceed $300 million against bank debt of only $4 million. The Nasdaq quote capitalizes the Class A float at roughly $37 million. The entire swing sits outside the services franchise and says more about how management deploys idle cash than about algorithm demand.
The next several prints resolve one question. Does the advertising-algorithm franchise stabilize, or does mainland contraction keep eating the only cash-generating line while the treasury keeps leaking. A second half that shows a smaller investment loss and a slower revenue fade would argue the discount is mostly a governance and jurisdiction haircut. Another large mark and another double-digit top-line drop would argue the market is correctly treating the cash as unreachable.