BALANCEb180rel/funding-ma4 min read
AI Cuts Game Costs Least Where the Money Is, M&A Advisor Warns
M&A advisor Rhys Dekle told PGC Summit Montreal that AI savings are real for indies but "mostly a mirage" in triple-A, and sellers must now prove humans made their work.

Patch notes
Strategic Alternatives partner Rhys Dekle called AI cost savings real for indies, mixed in mobile and "mostly a mirage" in triple-A at PGC Summit Montreal.
Square Enix aims to have AI perform 70% of its QA and debugging by the end of 2027.
Purely AI-generated content cannot be copyrighted in the US or Europe, reducing sale value at closing.
Dekle's sell-side deals include Spry Fox to Netflix, System Era to Devolver and PeopleFun to AppLovin.
"The studios that look clean in 2028 are the ones keeping records in 2026."
AI savings are "realest exactly where there's least to buy, and thinnest where the money is," Strategic Alternatives partner Rhys Dekle told the Global Trends track at PGC Summit Montreal, arguing that cost reductions from AI tools translate unevenly across the games industry and rarely raise a studio's sale price on their own.
Dekle, a certified M&A advisor who works exclusively in games, framed his talk as "the money conversation, not the creative-AI debate." His sell-side deal sheet includes Spry Fox to Netflix, System Era to Devolver and PeopleFun to AppLovin. Before joining Strategic Alternatives, he spent 11 years in business development at Microsoft, where he helped establish ID@Xbox, the publisher's independent developer program.
His assessment of where AI actually reduces production budgets was blunt: real savings for indie developers, mixed results in mobile, and "mostly a mirage" in triple-A, where expanded scope absorbs the efficiency gains.
Are production costs actually falling?
Dekle described AI arriving in two waves. The first brought generative tools that produce single assets such as art or dialogue. The current wave delivers agentic tools that can handle entire job functions such as QA. He cited Square Enix's stated goal of having AI perform 70% of its QA and debugging by the end of 2027 as the scale of ambition now in play.
The budget math is less favorable than the tooling suggests. Saving time on one task barely moves an overall budget, Dekle said, and someone still has to verify everything AI produces. In triple-A development, time saved typically gets reinvested in making the game bigger rather than reducing spend. He also argued that some studios currently labeling layoffs as "AI efficiency" are in fact simply cutting costs after the post-pandemic boom.
The takeaway: cost savings concentrate in indie development, where budgets are smallest and buyer interest thinnest.
Why cheaper games do not mean richer studios
Dekle compared AI adoption to outfitting every bakery with an oven that bakes twice as fast. Bread gets cheaper for all customers, but no bakery profits unless it owns something rivals lack.
"AI relocates value. It doesn't create it," he said. "It moves value away from anything that can be prompted, toward anything that can't."
He gave studios four questions to test whether AI-created value stays with them:
- Is the tool exclusively yours, or can a competitor buy it tomorrow?
- Does it sit on something only you own, such as IP, community or data?
- How long would it take a rival to copy?
- Who collects the money: the studio, the platform, or the AI vendor?
The outcome varies sharply by studio type. A co-development shop without its own IP loses ground because AI now performs cheaply the work it used to charge for. A studio holding a major franchise comes out ahead. And a production pipeline that exists only in three people's heads is, in Dekle's words, "not an asset, a risk."
He compressed the logic into one line: "Zero moat times AI equals faster zero."
What buyers check during diligence
The acquisition market adds a legal complication. Dekle noted that content made purely by AI cannot be copyrighted in the US or Europe, which effectively means anyone can copy it. Sellers who cannot demonstrate clear ownership of their content, or confirm it does not infringe third-party rights, face buyers holding back more money at closing.
The burden of proof sits with the seller, according to Dekle. Documentation he expects buyers to demand includes source history, layered working files and an asset register distinguishing what a human made, what had AI assistance and what was fully AI-generated.
He pointed to the backlash against Square Enix's Kingdom Hearts collection cover as a cautionary case. Fans flagged it as AI-made within hours, and the company's denials failed to persuade across three revised covers. If one of the world's largest publishers cannot prove its art is human-made, smaller studios face a harder standard in front of a buyer's lawyers.
"The studios that look clean in 2028 are the ones keeping records in 2026," he said.
Does skipping AI carry a penalty?
Dekle said avoiding AI hurts a studio's value only when it cannot adopt the technology. Studios that deliberately choose human craft can command a premium as backlash against "AI slop" grows. For studios whose value rests in their IP, adoption matters little either way.
"A clean, all-human studio can be worth more," he said, citing fewer diligence obligations and nothing requiring insurance carve-outs.
Three actions for studio leadership
Dekle closed with three operational steps: name your moat, the one thing AI cannot hand a competitor; get your IP records in order by tracking what is AI-made; and do not expect AI adoption alone to lift a sale price.
"Adoption is table stakes now, not an edge. It protects a moat; it doesn't make one," he said.
His closing line framed the valuation question in full: "You don't get paid for the horsepower everyone can buy. You get paid for the track only you can run on."
With buyers already pricing AI-provenance risk into deal structures, the asset registers studios build this year will likely determine what clears diligence when deals close in 2028.
via pgconnects.com (Original)
More from James Calloway
Show full bio
Senior reporter covering money, funding and market shifts around video game industry at Game Dev Wire.
223 articles