5 minBusiness
AI Driving Game Demo Costs Developer $1,000 a Day, Forcing Bank Loan
The developer of Teach My Little Sister How to Drive is spending $1,000 daily on AI tokens to keep the demo running after player numbers grew more than twentyfold, prompting a bank loan and raising questions about the economics of generative AI in games.
A small independent game studio has taken out a bank loan to keep its AI-powered driving instruction demo running after daily operating costs reached $1,000, exposing the financial strain that generative AI can place on even modest game projects.
The developer behind Teach My Little Sister How to Drive, a driving instruction simulator where players guide a generative-AI-powered non-player character, said the demo's player base has grown more than twentyfold over the past month. That surge in traffic has driven up the cost of the AI models that power the game's dynamic NPC, which constantly queries services such as Google Gemini and ChatGPT as players interact with it.
«Over the past month, the number of demo players has grown more than twentyfold,» the developer said. «This growth has also increased our operating costs considerably.»
According to a blog post from developer Easy Fox, the team has taken out a bank loan to help cover the expense of keeping the demo online. AI models charge based on usage, consuming units known as tokens with each request. Because Teach My Little Sister How to Drive relies on continuous AI queries to let the NPC follow instructions, react, and even disobey, the cost scales directly with the number of active players.
The developer stressed that individual players are not especially expensive to support. The total bill has climbed because so many people are playing at once. For the eventual paid release, the team plans to factor expected AI usage into the game's price rather than charging players separately for their own token consumption.
«For the paid release, we plan to account for expected AI usage in the game's price,» the developer said. «Players will not be charged separately for their individual AI token usage after purchasing the game.»
The game originally launched its demo in February. Interest from streamers, including thinknoodles, gave it a significant popularity boost. It has also found an audience in Asia, where the developer suggested there is less stigma around generative AI than in Western markets.
Despite the loan, the studio's future remains uncertain. Easy Fox is a team of just four people, and the developer acknowledged that financial pressure could force it to close the demo earlier than planned, before the full release. No decision has been made yet.
«We want to continue operating the demo and would love for more people to enjoy it,» the developer said. «However, we are a small team of four, and financial pressure may mean we have to close the demo earlier than expected before the full release. We have not made that decision at this time.»
The studio is also exploring ways to let local AI models handle the NPC for players with sufficiently capable PCs, which could reduce reliance on paid cloud services and lower ongoing costs.
The situation reflects a broader reckoning for companies that build products around generative AI. In a recent Axios report, one unidentified company spent $500 million on AI tokens in a single month after failing to set usage limits on its account. Elsewhere, the creator of OpenClaw reportedly burned through $1.3 million in OpenAI API tokens in one month.
Larger game companies have also faced scrutiny over AI. Capcom recently had to clarify comments about its proprietary RE Engine being described as «AI powered.» The publisher behind Arc Raiders defended AI use in Embark's extraction shooter, though the team has since worked to replace stilted AI performances with human ones. Activision, one of the industry's largest players, has come under fire for AI usage in Call of Duty and its marketing materials.
For now, Easy Fox's experience offers a concrete example of how quickly AI-driven operating costs can outpace a small studio's resources, even when a game finds an audience.
4
