Mobile Mavens on AI: doubled output, pricier RAM, and a bubble risk

Burny Games' organic growth team now handles at least twice as much work per week without overtime, after several weeks of optimising AI-driven processes — that's Alisa Eren's headline claim in the latest Mobile Mavens round-up, where seven industry figures assess whether a year of practical AI use has made the games business more optimistic.

Eren, Organic Growth Lead at the studio, describes a daily stack: custom skills and workflows in Claude, automated task creation in Asana, and automated user acquisition dashboards wired together so data, insights and next steps land in one workflow. "Faster production raises the baseline for everyone," she said, arguing that thoughtful, professional AI use is now a competitive requirement in mobile — and that the strongest teams will pair broad skills with systems that stretch them further.

Is the hype deflating or delivering?

Will Luton places the industry somewhere between the peak of inflated expectations and the slope of enlightenment on the Gartner hype cycle. His list of undeniables: consumers dislike public-facing AI in art, video and sound, especially on console and PC; it genuinely speeds up development; it's less impactful than some hoped; and it hasn't yet worked as a game mechanic.

At Village Studio, staff can use whatever AI tools they like, but everything in the final game — art, sound, writing — must be 100% human-made. Luton uses AI for coding, "mostly line completion and bits of refactoring," and admits he has "lost days and weeks of work from bad AI code." He flags cybersecurity as a double-edged case: AI is good at the boring, repetitive prodding that hacking involves, which means short-term problems but long-term security gains.

His sharpest warning concerns finance. "I think there is a real risk that we may be facing a dot-com-style bubble, where the wheels fall off of the circular infrastructure finance around GPUs and data centres," he said, adding that local, open models already passing the "good enough" threshold could strand trillions in infrastructure.

What does the supply flood do to attention?

Colan Neese argues attitudes haven't swung so much as split: coding assistants entered studios with little pushback, so concern has shifted to jobs and to anything players can see — "art is where the fights will be."

His attention share data quantifies the squeeze: the top ten games' share of tracked gaming attention fell from 36.5% to 27.8% between the Gears 5 launch week in 2019 and the E-Day launch week this autumn. AI pushes the number of games toward infinite while play hours stay fixed — a world mobile has inhabited for fifteen years. Neese is nonetheless an optimist, betting the next UGC wave comes from people who could never make a game before, though he concedes hobbyist AI work "raises serious IP questions" — early YouTube, built on pirated clips, is his precedent.

Where is AI actually paying for itself?

Jan Sommerfeld reports AI has had "a huge positive impact" across Almedia's departments, from analysing campaign performance to triaging bug reports, with the company working to let staff securely query internal data in plain English. His worry is cost spiralling as usage scales.

Kimberley Shearer points at infrastructure. AI-assisted management now lets developers deploy servers, troubleshoot and interact with infrastructure APIs in natural language. But the AI boom is also inflating the economics underneath games: rapid data centre expansion has driven significant RAM price increases and supply chain pressure. "We've effectively reached a point where AI is simultaneously making some aspects of game development cheaper and easier, while making the physical infrastructure supporting those games more expensive," she said — so AI's biggest multiplayer impact may be behind the game, not inside it.

Pascal Clarysse supplies the small-studio ledger. Big Karma, a mission-driven indie, runs three tools: Gemini bundled in Google Apps Workspace at $48 a month for the whole team; a Claude Code 20x Max account at $200 a month on the lead developer's machine; and Tripo, which first-drafts 3D renders from hand-drawn character art by Miguel Francisco (Angry Birds, Best Fiends) before animators correct and clean manually. Any generated output must be fully reviewed by the requesting team member — accountability stays human.

AI sits under 1% of the studio's monthly burn, yet Clarysse estimates it replaces 2-3 man-months of part-timer or freelancer work — and he doubts hired hands would iterate this fast. His economic worry is supply-side: "if the means of production enable faster development for all of us, the explosion on the supply side will shrink individual demand for each title," especially with organic discovery broken and UA costly for anyone without Turkish government subsidies or robust LTVs.

What comes after tool adoption?

"Almost every game team I speak to is already using AI somewhere in development or publishing," said Wenfeng Yang, citing Tripo for low-poly assets, Sensor Tower data combined with Codex for analysis, and AI-generated ad creatives with A/B testing. He notes AI-native companion products like Zeta, OOC and Tipsy show people will pay for AI interaction, and that game companies are becoming LPs in funds backing AI startups.

His longer-term call: AI-first companies where AI handles coordination, task tracking and follow-up, freeing people for creative direction. "I'm more confident about AI than I was about the metaverse," he said. "AI has passed a simple test: teams are using it to get real work done."

Part two of the round-table is pending, and the topic gets its own track at Pocket Gamer Connects Jordan on November 7-8 and PGC London on January 18-19, 2027 — worth watching for whether the optimism survives the bubble talk.