AI Token Spending: How One Startup Learned to Optimize Costs (2026)

In the fast-paced world of startups, where every decision can make or break a business, Sarthak Dhawan, a 21-year-old co-founder of Turbo AI, shares a cautionary tale of his company's unexpected $30,000 AI token expenditure in a single month. This story, originally published in Business Insider, highlights the fine line between innovation and financial prudence in the age of AI. While Dhawan doesn't view this as a total mistake, it serves as a valuable lesson in managing AI spending and maintaining a healthy balance between speed and cost.

The AI-Driven Shift in Coding

Dhawan's journey began with a simple idea: an AI learning tool app. However, the landscape of coding has undergone a profound transformation with the advent of AI. He explains, "AI has shifted my job from writing code to reviewing it." This shift is not unique to Dhawan; it's a trend that many software engineers are experiencing. The once-essential task of writing code from scratch is now often replaced by AI-generated code, leaving engineers to focus on reviewing and understanding the AI's output.

This development raises a critical question: as AI takes over more tasks, will engineers' coding skills atrophy? Dhawan acknowledges this concern, stating, "I can feel my coding skills atrophying." He further elaborates on the historical context, noting that 20 years ago, engineers deeply understood the ins and outs of their codebases because they wrote them. However, with AI-generated code, the codebase becomes an entity that engineers may not fully comprehend, potentially leading to a loss of control over their work.

AI Spending: A Balancing Act

Turbo AI's approach to AI spending is pragmatic. They don't have a set budget for AI tokens, but they do monitor their usage. Dhawan mentions, "We keep a loose eye on it, but there's no approval process for using tokens." This flexibility allows them to adapt to the evolving needs of their business. However, it also means that they must be vigilant in managing their costs.

The $30,000 AI token expenditure in a single month was a wake-up call. Dhawan explains, "We crossed $13 million in lifetime revenue this year. We're still in the mentality that if it makes us more productive, it's probably worth it for the business in the long run." This mindset is crucial for startups, as it encourages them to invest in tools that can drive growth, even if they come with a price tag. However, it also requires a fine balance to avoid overspending.

The Easy Fix: Fast Mode

One of the key lessons from this experience is the importance of paying attention to AI settings. Dhawan's story highlights the impact of a simple setting in Claude Code: fast mode. He explains, "The setting on Claude Code, called fast mode, makes it feel faster, but it also costs way more per token." By leaving fast mode on, Dhawan inadvertently drove up the company's AI token bill.

The solution was straightforward: turn off fast mode when it's not needed. Dhawan notes, "Switching out of fast mode barely made a difference in the speed of output." This simple fix demonstrates that even in the world of AI, small adjustments can lead to significant cost savings. It also underscores the importance of being mindful of AI settings and their impact on spending.

The Way Forward: Balancing Innovation and Cost

As AI continues to evolve and become more integrated into the software development process, startups must navigate the fine line between innovation and cost management. Dhawan's story serves as a reminder that while AI can drive productivity and growth, it's essential to be mindful of the financial implications. By adopting a pragmatic approach to AI spending and learning from experiences like this one, startups can harness the power of AI while maintaining a healthy balance between speed and cost.

In conclusion, the journey of Turbo AI is a testament to the power of innovation and the importance of financial prudence. As AI continues to shape the future of software development, startups must be agile, adaptable, and mindful of the impact of their decisions. By learning from experiences like Dhawan's, they can navigate the challenges and opportunities of the AI-driven world with confidence and foresight.

AI Token Spending: How One Startup Learned to Optimize Costs (2026)
Top Articles
Latest Posts
Recommended Articles
Article information

Author: Margart Wisoky

Last Updated:

Views: 6310

Rating: 4.8 / 5 (78 voted)

Reviews: 93% of readers found this page helpful

Author information

Name: Margart Wisoky

Birthday: 1993-05-13

Address: 2113 Abernathy Knoll, New Tamerafurt, CT 66893-2169

Phone: +25815234346805

Job: Central Developer

Hobby: Machining, Pottery, Rafting, Cosplaying, Jogging, Taekwondo, Scouting

Introduction: My name is Margart Wisoky, I am a gorgeous, shiny, successful, beautiful, adventurous, excited, pleasant person who loves writing and wants to share my knowledge and understanding with you.