Cloud computing is the engine that powers several emerging technologies, including artificial intelligence (AI). These services deliver data storage and computing digitally. The key value proposition driving its rapid enterprise adoption is its ability to enable access to IT resources on demand without the need for significant upfront capital investments. It also has other advantages such as scalability, rapid provisioning, advanced security, and minimal management effort. The Indian public cloud services market was estimated at $7.51 billion in 2023, and its growth is expected to have a significant impact on India’s economy, contributing $310-380 billion by 2026 (~8% of GDP) and 14 million new jobs.
The cloud computing market is dominated by a few “hyperscalers” — large cloud service providers (CSPs) such as Amazon Web Services, Microsoft Azure, and Google Cloud Platform which provide services such as computing and storage at enterprise scale. The rapid growth of the cloud computing market must be viewed within the context of the increasing market shares of these hyperscalers. Potential competition concerns are heightened by vertically integrated hyperscalers, some of which also have a dominant presence in adjacent markets such as on-premises software, search engines, and e-commerce. Certain commercial policies that are employed by legacy software vendors (who also offer cloud infrastructure services) to exclude cloud competitors and gain market share are increasingly falling under the scrutiny of competition authorities.
Creating a competitive CSP market is critical to facilitate a multi-cloud approach that optimally combines the services of multiple providers based on factors such as cost, speed, scalability, and security requirements. While Indian CSPs differentiate themselves from hyperscalers in terms of cost-efficiency and personalised support, the latter’s business practices create technical and financial barriers for switching and multi-clouding.
On the technical side, limited interoperability (i.e. ability of services hosted on different CSPs to seamlessly interconnect) and restrictions on the portability of data and applications due to the use of proprietary application programming interfaces (APIs) lead to vendor lock-in. Since services hosted on different CSPs cannot interact, enterprises tend to use multiple CSPs in silos to serve different purposes. Self-preferencing licensing restrictions pose another set of technical barriers, with cloud players who are dominant in the legacy software market, limiting the compatibility of their essential software with a third-party cloud infrastructure or increasing the costs for third-party deployment through excessive licensing fees.
On the financial side, cloud credits (i.e. the option provided to new customers, such as start-ups, to utilise services up to a predefined limit without any cost during trial) and committed spend discounts (i.e. volume discounts targeted towards high-end enterprise users, offered as a percentage upon achieving cloud spending targets) have emerged as significant barriers. These incentives encourage single sourcing from large providers, and CSP entrants are unable to profitably match the value and coverage offered by hyperscalers.
Antitrust intervention based on a deeper understanding of the cloud market may be the first step to correct any market failures. Global antitrust and regulatory authorities, including the Federal Trade Commission, Office of Commissions, and Autorité de la concurrence, have conducted studies on the cloud computing market to gain a better understanding of competition concerns. There have been cases such as Microsoft’s acquisition of Activision Blizzard, where antitrust intervention to address competition concerns in cloud gaming led to modifications. However, merger regulations and ex-post interventions would bring little succour to competitors and may be insufficient to address the market failures of a proprietary cloud ecosystem where entry will be effective only at the fringes. Such entry will not be sufficient for customers to adopt a multi-cloud to realise the complete benefits of the ecosystem and of different applications integrating and transferring data across cloud systems.
In India, the paradigm of digital public infrastructure (DPI) is being increasingly adopted in areas such as digital payments, e-commerce, and e-health to overcome the interoperability challenge and ensure the improved aggregation of demand and supply by using open-source technologies. DPI is also being adopted in the cloud computing space. The Open Cloud Compute project aims to decentralise and expand computing infrastructure by connecting cloud providers of various sizes through open and interoperable APIs. A decentralised network that connects smaller providers would ease entry and create incentives for improving technical synergies between providers, potentially reducing lock-in for proprietary ecosystems. This could also provide for more cost-efficient and customised solutions, in addition to enhanced security and reliability from a distributed network vis-à-vis centralised data centres.
Unlike previous technological revolutions, India has embraced the AI revolution at the same time and with the same intensity as the rest of the world. Cloud computing capacity is the shovel that we need to emerge as one of the winners in the AI gold rush. Computing capacity, or compute, is a fundamental component of large AI systems as well as algorithmic innovation and vast data sets. It is also one of the most difficult elements for smaller businesses to set up in-house. We also need creative strategies beyond digital regulation to combat the concentrated power of hyperscalers, and providing open, sustainable, and effective cloud access and DPI in cloud computing seems like our best bet. Enabling cloud computing innovation that serves interests beyond the commercial gains of Big Tech requires public goods investments in digital infrastructure and models that have the capability to scale and support decentralised power.
The author is a visiting professor at ICRIER. Coauthored with Harishankar Jagadeesh and Bhargavee Das, respectively, fellow (consultant), and research assistant, ICRIER.
Disclaimer: Views expressed are personal and do not reflect the official position or policy of Financial Express Online. Reproducing this content without permission is prohibited.
Comments