Easy Fox Takes Bank Loan to Cover $1,000 Daily AI Costs for Driving Sim Demo

3 min read
Source: Kotaku
Easy Fox Takes Bank Loan to Cover $1,000 Daily AI Costs for Driving Sim Demo
Photo: Kotaku
TL;DR

The developers of the driving simulator 'Teach My Little Sister How To Drive' have taken out a bank loan to keep their free Steam demo running after AI token costs exceeded $1,000 per day. The four-person studio, Easy Fox, faces financial strain from a twentyfold surge in players, forcing them to consider early closure or a shift to local AI models for the full release.

Key points

  • Easy Fox is spending over $1,000 daily on AI tokens to maintain the demo of 'Teach My Little Sister How To Drive.'
  • The studio took out a bank loan to cover these costs after player numbers increased by more than twentyfold in the past month.
  • The game uses external AI services like Google Gemini and OpenAI to generate real-time conversations with the in-game character.
  • Developers are considering shutting down the demo early or implementing local AI models to reduce costs for the full release.
  • The full release is expected to factor AI usage into the game's price, though players will not be charged individually for token usage.

Background

This incident highlights the financial risks of integrating cloud-based AI into free-to-play game demos. While previous coverage noted the rise of AI-generated content in gaming, this case illustrates the direct operational costs for small studios relying on external API services for core gameplay mechanics.

How outlets are covering it

Kotaku emphasizes the broader industry trend of AI firms struggling to generate profit, noting that the high resource costs outweigh public willingness to pay for AI-driven content. FRVR corrects earlier reports, clarifying that the peak concurrent player count was approximately 405, not 62,000, which makes the financial burden even more striking given the modest player base. Automaton focuses on the game's popularity among Japanese streamers due to the AI's chaotic and sharp-tongued responses, while Tom's Hardware highlights the technical shift from Google Gemini to OpenAI models and the potential for local hardware solutions.

Why it matters

This case serves as a cautionary tale for indie developers integrating real-time AI into their games. It demonstrates that even modest player counts can incur unsustainable costs if the game relies on paid cloud AI services, potentially forcing studios to alter their business models or technical architecture to ensure viability.

What to watch

Easy Fox is currently evaluating whether to close the demo early or implement local AI models to reduce costs. They plan to factor AI usage into the pricing of the full release, which is expected next year, to ensure the game remains financially viable without charging players individually for token usage.

Share this article

Want the full story? Read the original reporting

Read on Kotaku