Using artificial intelligence to automate coding tasks once seemed like a no-brainer for tech companies. The idea was that AI could cut costs by reducing the need for human developers or boosting productivity. But as AI tools become more widely used, the financial and practical realities are starting to look less promising.

Surging Expenses and Unexpected Bills

In theory, deploying AI models to generate code is a cost-effective strategy. Companies can save on salaries and healthcare costs by replacing or supplementing human workers with AI. However, the actual expenses are climbing rapidly. Some businesses are spending over $150,000 a month just on AI tokens and usage fees, which is far more than they anticipated.

AI providers are feeling the strain too. The high demand for their coding tools is taxing server capacities, prompting some to raise prices or restrict access. For example, Microsoft’s GitHub Copilot has shifted to a pay-per-use model, making it more expensive for developers to rely on AI for everyday coding tasks. These rising costs are forcing many organizations to reconsider whether AI-driven coding is truly economical.

The Productivity Myth and Hidden Downsides

Many studies are casting doubt on the assumption that AI improves productivity. Some research shows that companies adopting AI saw little to no revenue growth, and others report that AI often creates more work instead of reducing it. This phenomenon, dubbed “workslop,” describes how AI can generate tasks that need fixing or revising, adding to employees’ workload instead of easing it.

Furthermore, employees are experiencing increased stress and burnout due to the constant demands of managing AI tools. Instead of freeing up time, AI sometimes complicates workflows, leading to frustration and resentment. As the costs of using these tools rise and their benefits remain uncertain, companies are questioning whether AI coding is worth the trouble and expense.

Overall, the promise of AI as a cheap, efficient coder is facing serious hurdles. The financial burden of AI usage is growing, and the supposed productivity gains are being questioned. It appears that for many organizations, the economics of AI-driven coding are starting to look worse than ever before. This could lead to a reevaluation of how and when to integrate AI into software development in the future.

Inspired by

• https://futurism.com/artificial-intelligence/economics-ai-churn-code

Sources

• nytimes.com
• claude.com
• axios.com