Whalesbook
India's artificial intelligence sector is taking a different path than global trends, favoring diversified cloud infrastructure and open-source tools. A report from Esya Centre shows that almost 90% of Indian AI companies are using or planning to use hybrid or multi-cloud strategies. This approach helps them avoid becoming too reliant on any single major cloud provider. The study included 227 companies involved in AI.
Spreading Out Cloud Use
About 62% of the companies surveyed are using services from multiple smaller cloud providers alongside their own data centers. More than 80% of firms do not have exclusive contracts with cloud vendors, showing they manage their infrastructure flexibly. This competition seems to be lowering costs, as nearly 99% of companies reported lower compute expenses. Using multiple vendors and managing costs are major reasons for adopting multi-cloud setups.
Prioritizing Performance and Security
Contrary to what might be expected, the report emphasizes that performance and security are more critical than price when choosing cloud providers. For three-quarters of the companies, these factors were rated as "extremely important," ahead of cost. The ability to scale and the range of services offered were also highly valued, indicating a focus on reliability and flexible deployment as AI workloads grow.
Open-Source Models Are Growing
The Indian AI sector is increasingly using open-source technologies. Roughly 83% of organizations use open-source foundation models, 67% use proprietary models, and 63% use custom-trained systems. Companies also deploy models flexibly, with three out of five using multiple models in the same category and 81% combining small and large language models based on what they need for specific tasks.
India's generative AI startup scene has grown substantially, increasing 3.6 times from the first half of 2023 to the first half of 2024. However, companies face challenges like poor data quality, limited access to government data, unclear copyright rules, and regulatory uncertainty regarding the Digital Personal Data Protection Act (DPDPA), 2023.
"The survey data shows an AI ecosystem that is much more active and competitive than many believe," said Meghna Bal, Director at Esya Centre. The report suggests that the main barrier to wider AI adoption in India is not market concentration or funding, but a significant lack of skilled AI talent.