Netflix closed five game studios in five years and kept every category that behaves like an episode. Then it premiered the GTA VI reveal through its documentaries department, and a trailer for an unreleased game became the most-watched thing on the service that week in 87 of 93 countries.
This report tests one hypothesis across three companies: games are being reclassified from products you seek out into programming that arrives, and the filter being applied is not genre or quality but whether a thing behaves like an episode. Amazon and Microsoft moved on the same afternoon in July, independently, toward adjacent versions of the same shift. Six weeks later Microsoft put a meter on it.
Part I tests it against what Netflix shut and kept. Part II tests it against Amazon, Microsoft, and market data that disagrees with itself by nearly six times. Part III tests it against the compute the whole model runs on — and against this report's own failed drafts.
None of these four numbers is what it looks like. A Netflix view is hours watched divided by runtime, not a headcount; the 11x is growth off a base never published; the 4% is Microsoft's own unaudited figure. Each is handled where it is used.
The instrument that would settle most of this argument is the one nobody publishes: cloud session length, and Netflix game hours per subscriber. Peters was asked for the second in July and declined. Until those numbers exist, every thesis in this space is inference dressed as measurement, including this one.
Abbas Saleem is a Principal Consultant at Llama & Griffin, advising game studios, publishers, and investors across MENA, South Asia, Southeast Asia, and Europe. He writes Recognizing Patterns: gaming industry intelligence 12 to 24 months before it becomes consensus. LinkedIn