The shift from fixed pricing to usage-based models by AI companies like Anthropic and GitHub is impacting IT firms, particularly in India's $300 billion software service industry. As demand for AI productivity tools rises, IT companies face increased inference costs that could pressure margins in the short term. Industry experts, including Karan Uppal from PhillipCapital, note that while firms can offset some costs through factors like currency depreciation, competitive bidding may limit their ability to pass on expenses to clients. However, as the complexity of tasks increases, the cost per token is expected to decline, offering long-term benefits for IT firms. Manish Tandon, CEO of Zensar Technologies, emphasizes that as AI becomes more mainstream, usage-based pricing will likely become more advantageous. CEOs from various IT firms, including Mphasis and Coforge, highlight the importance of managing AI-related costs and productivity gains, suggesting that while clients may not demand proportional price reductions, they will expect some concessions as productivity improves.