IT companies using artificial intelligence models at scale are likely to benefit as AI firms move from the per-seat to usage-based model of pricing, though they may face some near-term margin concerns, industry experts said.
Demand for productivity tools such as Anthropic’s Claude Enterprise and GitHub has increased over the past year. The higher demand increases inference or compute cost for AI firms, which say it is not sustainable in the long run for them to absorb this. To address this issue, Anthropic and GitHub have moved from fixed price to consumption-based models.
For large IT companies, which are big consumers of AI technologies as they restructure their business, this triggers additional margin pressure in the short term.
Currently, India’s $300 billion software service industry is at a transitional period, with companies rewiring their deals, operating models and services to incorporate agentic and generative AI offerings, attracting high complexity tasks in their processes.
“The complex tasks will likely consume more tokens, and IT companies will have to end up paying higher amounts (for the same),” said Karan Uppal, vice-president and lead analyst for IT services at PhillipCapital.
While IT firms could offset higher costs with tailwinds from factors like the rupee’s depreciation, Uppal highlighted that large-cap firms bidding for deals with thin margins amid a highly competitive environment might not hold the pricing power to pass on entire costs to customers.
But in the long term, IT companies are expected to benefit from economies of scale, said industry experts. As task complexity increases, along with token usage, the cost per token will fall, said Manish Tandon, chief executive of mid-tier IT company Zensar Technologies. “As AI becomes more mainstream, most of the use of AI will be through APIs or technology calls, rather than through individuals sitting on machines. So, as the market becomes more competitive, (usage-based pricing) will be good,” he told ET.
IT firm Mphasis’ CEO Nitin Rakesh likened the situation to the early days of cloud computing, when cost optimisation was a primary concern among service providers. “Even as recently as a year or two ago, there used to be a lot of focus on driving cost reduction on cloud compute using FinOps,” he said. “To me, this (AI tokens) is no different than that.” FinOps is a management practice that helps companies better understand and optimise their cloud spending.
Coforge CEO Sudhir Singh said his company primarily uses tokens in two scenarios: internal productivity improvements and client-facing AI solutions. “We have already guided that our G&A costs next year will remain flat, even as the market expects us to grow 30-35%. That will happen because of AI,” Singh said, adding that token consumption costs remain below the savings generated through automation.
Hexaware CEO R Srikrishna said enterprises are becoming more sophisticated in managing AI economics and are increasingly evaluating token usage against productivity gains.
“At most software engineering services engagements, token costs are borne by clients…if productivity improves by 40%, clients may not expect a 40% reduction in pricing. They may ask for 25-30%,” he said, noting that part of the gains could flow into profitability.