GenAI as a service generates $5-10 mn revenue per project globally: study

Indian technology services providers had even lower revenue generation from GenAI at $1.5-2.2 mn

Generative artificial intelligence (AI), the big technology bet that was supposed to change the growth trend for the Indian IT services sector, is still too small to make any major impact on revenue. 

The reason, according to a recent survey by UnearthInsight, is because only a third of Forbes 2000 firms have modernised their ERP platform, and less than 10 per cent have mature data. 

Consequently, GenAI as a service generates only about $5 million-$10 million revenue per project, says the report. 

The report found that global technology services firms' average revenue per project is around $4.8 million to $10 million while Indian tech services firm average revenue per project is in the range of $1.5 million to $2.2 million. 

When it comes to the various stages of AI projects, the number of deals under proof of concept (PoC) are much higher than the actual projects. In case of global tech firms, the PoC projects are in the range of 2,500–2,800. In case of Indian IT services players it is 2,250–2,350, the UnearthInsight report found.   

Actual, live GenAI projects at global tech firms is in the range of 1,550-1,300, while it is much lower at 350-400 for Indian tech firms. 

When it comes to commercialisation of these deals, the number falls even further. Commercial pilot projects at gobal firms is around 900–1,050; for Indian tech players, this number ranges between 600-800. 

Most companies have said that GenAI deals, though happening, are still very small and do not have a material impact on revenue. 

K Krithivasan, CEO and MD, TCS has maintained that GenAI deals are small in size, for starters. During the Q2 FY25 analysts call, he said that such deals tend to be small in size. 

“It can vary, like deal size can vary based on the core size of the organisation, tenure of the deal, it could be $10 million to $20 million, $30 million based on the tenure and size of the organisation,” he said in reply to a question on what would be the size of projects that look to creating an AI office within an organisation. 

Corroborating the findings on the fact that the enterprise technology landscape is not yet ready for a GenAI kind of deployment, Phil Fersht, CEO of HfS Research said: “The real revenue with GenAI is with the 5 per cent of major enterprises which are going beyond pilot projects to scaled implementations. Remember, GenAI is only 18 months old, and many enterprises are only now allocating budgets dedicated to this, as opposed to eking out funding from existing tech projects.”