Operational inefficiencies always exist and always cost money. The questions that matter are where the inefficiencies are located and how much they cost. In the games industry, we can see the number: global consumer spending on games hit $195.6 billion in 2025, an all-time high, but studio operating margins have not followed and have declined around 7% every year since 2021. The spread lies in operational costs, and in game development, much of that cost lives in the pipeline
Development Costs Are Up 90%
At GDC 2026, Google Cloud's global director for games Jack Buser put a number on it: development costs are up 90% since 2017.
Matthew Ball's 2026 State of Video Gaming report confirms the compression. Of the $24 billion in net market growth since 2020, most studios captured almost none of it in margin. The studios that are shipping titles are failing to ship with business-sustaining margins.
This is a cost structure problem, and in game development much of the cost lives in the pipeline.
Handoffs are a silent velocity killer
Every hour a build sits waiting is an hour of blocked context. It might be waiting for a QA team to receive it, for an external partner to download it, or for a certification lab to acknowledge it. While build times can introduce delays, the friction is rarely in the build process itself. It is in the handoff. A producer sees it as schedule drift. A build engineer sees it as a Tuesday afternoon spent investigating why the Manila team received a build from last Thursday.
Outsourcing makes that surface area larger every year. Studios managing co-dev partners at scale have dozens of delivery destinations on different networks, often in different regions, each with its own access requirements. The pipeline cost shows up not just in headcount but in iteration speed, which is the number of feedback cycles a team can run in a sprint.
The build distribution problem is not a new one. What has changed is how much it costs when it does not work.
The last 10% is where it gets expensive
Also at GDC 2026, Buser named something build engineers already know: the last 10% of game development is where teams endure the most toil, and it happens largely during QA.
If QA is already the most painful phase, any friction in build delivery compounds directly in the most expensive window of production. A QA team waiting on the wrong build does not just lose the wait time. It loses the testing cycle. It loses the bug report. It loses the sprint.
Buser framed AI as an immune system for the game. That framing holds. But an immune system needs a circulatory system: the infrastructure that carries the right build to the right person at the right time. That delivery layer is where the cost accumulates when it is not working. You can have the best QA process in the world and it still fails if the build does not arrive.
AI Raises the Stakes
AI is compressing development time at the creation layer, and the volume impact is already visible in real workflows. Game studios are moving from experimental AI use to operational automation, and the real value now comes from pairing AI with structured tooling so teams can speed up repetitive work without losing control. Solsta CLI works great for studios looking to experiement with more automated deployments across all teams and partners.
AI is great but it still comes with its own set of bottlenecks when working with massive builds. The maintainers of the Godot game engine ran into exactly this problem. AI-generated pull requests were overwhelming their review process, and the volume outpaced the infrastructure around it. That is a preview of what studios will hit when AI-assisted creation meets a delivery layer that was not designed for the new tempo. AI investment in the creation layer without corresponding investment in the distribution layer shifts the constraint rather than removing it.
There are substantial gains available at the delivery layer. AI is accelerating the way we work, and now we are seeing once-reliable infrastructure respond in ways that were not happening just one year ago. GitHub recently reported explosive growth of commits, on pace for 14 billion commits in 2026, up from 1 billion in 2025. It is becoming increasingly clear that infrastructure that worked in the “before times” needs to be rethought for the post-2025 era.
Solsta Solve Delivery
Solsta is a collaboration platform built to automate game build workflows from any CI system to internal teams and external partners. It turns build delivery into a controlled layer instead of a chain of emails, shared drives, and manual uploads. Games studios that are experimenting with agents or AI workflows can now use the Solsta CLI for automated deployments across all platforms.
SOC 2 Type 2 and ISO 27001 certified.
FAQ
What is game development pipeline cost and why does it matter now?
Game development pipeline cost is the cumulative time, infrastructure expense, and labor between a code commit and a build reaching its destination. It matters now because development costs are up 90% since 2017 while revenuegrowth is not flowing to margin. As AI tools increase build frequency and outsourcing multiplies delivery destinations, the delivery pipeline becomes a primary cost driver, not a secondary concern.
Why are game development costs rising if the industry is growing?
Revenue growth and cost growth are not the same metric. Global consumer spending reached a record $195.6 billion in 2025, but most studios captured almost none of the $24 billion in net market growth since 2020 in margin. Development costs are up 90% since 2017. Revenue is growing, but the cost of making games is growing faster.
Why does pipeline time matter so much?
Every hour a build artifact spends in transit carries a direct cost in blocked context and delayed iteration — whether it is waiting to reach QA, an external partner, or a certification lab. A build that takes 48 hours to reach QA has differentiteration economics than one that arrives in 20 minutes. The handoff time is not invisible. It compresses the number of feedback cycles available in a sprint.
How does AI affect game build pipelines?
AI tooling compresses the creation phase, which means higher build frequency and more builds destined for QA, co-dev partners, and external destinations. A delivery pipeline not designed for higher frequency becomes the new bottleneck. AI investment in the creation layer without corresponding investment in the distribution layer shifts the constraint rather than removing it.
How does outsourcing affect build distribution?
Each external partner requires builds reaching a specific destination, with specific access controls, on a reliable schedule. Studios without a distribution layer designed for this absorb the complexity as manual overhead.
Sources
Buser, Jack. Quoted in Forbes-Calvin, Alex.
"This is not a sustainable business model" — Why Google Cloud's Jack Buser thinks AI can save the games industry.
GamesIndustry.biz, April 21, 2026.
Ball, Matthew.
The State of Video Gaming in 2026.
MatthewBall.co, 2026.
McAloon, Alissa.
Google Cloud's vision for AI in the game industry.
Game Developer, March 11, 2026.
Cherny, Boris.
Post on X, December 2025.
