For Azure, Ola’s departure would mean a daily revenue loss of INR 5-25 lakh. For AWS, this number would range between INR 30 and INR 40 lakh, as per some estimates and disclosures by Ola founder Bhavish Aggarwal on X.
However, losing a customer like Ola would just be a needle in the haystack for the two hyperscalers which have amassed 40% of the USD 8.3-billion public cloud market in India. They are unlikely to lose business due to Ola Krutrim’s freebie offering of cloud services, executives and market watchers said.
Microsoft, Amazon and Ola did not respond to ET's emails seeking comments till press time on Tuesday.
Two weeks ago, Ola founder Aggarwal had announced that the company is migrating from Microsoft Azure to its homebred Krutrim cloud after the latter’s social media subsidiary LinkedIn had deleted his post on “pronoun illness” of western AI models. “Any other developer who wants to move out of Azure, we will offer a full year of free cloud usage. As long as you don’t go back to Azure after that!” he wrote on X. So far, 2500 developers have joined Krutrim.
In another post last Sunday, Aggarwal compared the pricing of AWS (USD 213) and Azure (USD 93) for cloud computing. “No wonder the biggest cost for Indian startups is cloud costs. @Krutrim cloud will be lower and in ₹₹,” he said.
“Free cloud services for a year is a very lucrative offer for mid-tier organizations and I believe Ola must be counting applications in hundreds as of now,” said Bhavesh Goswami, founder and CEO of consulting and skilling firm CloudThat. “But no player can build infrastructure to support this kind of exodus overnight. Therefore, it remains to be seen whether Krutrim has that capacity and can support integration so rapidly.”
All hyperscalers, including Azure, AWS and Google Cloud Platform, have always been offering certain cloud storage and processing services free for an initial one year for any organisation, he added. “So it is not some new disruption in the cloud market!” he said.
Amazon, Microsoft
According to International Data Corporation (IDC), the public cloud services market in India was estimated to be USD 8.3 billion in 2023, with Microsoft Azure leading the market with 25% share followed by AWS’s 15%. Assuming that Ola’s daily spending was INR 75 lakh-INR 1 crore, it is less than 0.01% of what the hyperscalers collect as revenue, estimates suggest.
“Krutrim’s proposition could attract some niche players in the near term, but the scale and infrastructure built by hyperscalers is almost impossible to challenge,” said a partner at a top consulting firm. “Their geographical span across India, platform and software stack integrations, redundancy support, reporting and analytics, security are some features which would be difficult to compete with,” he added.
Experts have also questioned how Ola plans to fund free/lower cloud distribution for a year.
According to a report by Forrester, Indian organisations spent an average of INR 370 crore on public cloud in the year 2023. “Assuming that few mid-tier organisations join the Ola cloud, the company would still be running a bill of hundreds of crores by the end of year,” one executive at a cloud platform company said.
Back in January 2024, Ola’s AI venture Krutrim had raised USD 50 million in a fundraising round led by Matrix Partners valuing the company at USD 1 billion and making it the first AI unicorn in India within less than a year of its incubation.
“While AI adoption is slowly picking pace in India, global hyperscalers are collaborating with vendors like Nvidia and Intel to build the AI infrastructure and with other AI organizations for different language models” said Rajiv Ranjan, Associate Research Director, Cloud & AI, IDC India.
“Indian cloud service providers would have to come up with comparable solutions either through strategic alliances, joint ventures or their home-grown service offerings in order to compete with hyperscalers.”
As large organizations build their AI applications, factors such as integration of different language models with platforms, data quality and performance of the language models will matter more, Ranjan said.
Disclaimer: The copyright of this article belongs to the original author. Reposting this article is solely for the purpose of information dissemination and does not constitute any investment advice. If there is any infringement, please contact us immediately. We will make corrections or deletions as necessary. Thank you.
