Games · Answer

Is live service or premium release the better model for a mid-size games publisher?

The disclosed shares argue for live service and two of the three largest of them rose for the wrong reason. Electronic Arts reports 71% of net revenue from live services — while that line shrank 1% and full-game revenue grew 7%. A share that rises because the numerator grew and one that rises because the denominator collapsed look identical on a slide and mean opposite things.

The evidence

The three disclosed shares.

Electronic Arts: live services and other, 71% of net revenue — $5,383M of $7,531M. Take-Two: recurrent consumer spending, 78% of net bookings. Ubisoft: player recurring investment, 60.6% of net bookings, against 43.7% the prior year.

Two of them are arithmetic, not momentum.

EA's live-services line fell by $78M, down 1%, while full-game revenue grew 7%. Ubisoft's ratio rose because total net bookings fell 17.4%.

Recurring is not the same as compounding.

Roblox guided Q3 2026 bookings to $1,576–1,653M, a 14–18% year-over-year decline, with free cash flow guided from −$60M to +$5M against $443M a year earlier. Epic laid off more than 1,000 people in March 2026 citing reduced Fortnite engagement. Sony cancelled seven live-service titles and closed three studios.

The counter-case is a single clean instance.

Clair Obscur: Expedition 33 — reported budget under $10M, a core team of about thirty, released day-one on Game Pass, past 8 million units within a year and Game of the Year at five award bodies. It defeats two claims at once: that premium single-player is dead, and that day-one subscription placement kills premium sales.

There is no credible failure rate, and we will not supply one.

The widely quoted figure that 83% of launched live-service games close within three years is a vendor survey of mobile developers published by a company whose business is extending the life of underperforming games, and it predates the current window. The defensible statement is a census of named shutdowns, not a rate.

What would have to be true for this to be wrong

If your studio already has a live audience large enough that retention economics beat acquisition economics, the share question is settled for you and the real decision is cadence. The evidence above bears on a publisher choosing a model, not on one already inside a working one.

Sources: SEC EDGAR full-text search and company filings, with each figure's period stated above. AI-answer measurements are CULT+MATH primary research, criteria registered before retrieval. Published 2026-10-21 by Ipalibo Da-Wariboko, CULT+MATH LLC. Figures carry the date they were pulled; where a number could not be sourced, this page says so and gives none.

This is one of six decisions covered in the games publisher decision brief — each with the evidence behind it, who owns it, who it is contested with, and what would have to be true for the answer to flip.

See the brief →