This article is based on analysis originally published by Rapyder Cloud Solutions via DQ India in late 2019, covering forecast trends for cloud computing in 2020. Some statistics cited below originated from Gartner and other third-party sources referenced in that piece.
Why Cloud Investment Kept Growing Despite Economic Headwinds
IT budgets are historically among the first casualties when economic uncertainty looms. Yet heading into 2020, the data told a different story. Research firm Gartner forecast that business IT spending in India would grow by approximately 9% in 2020, up from around 6% growth in 2019. That acceleration pointed to a broader structural shift: cloud and automation were no longer discretionary investments but core operational infrastructure.
Analysts projected that more than 80% of enterprise workloads would move to cloud environments by the end of 2020. Rather than signaling a plateau in adoption, this milestone represented the moment cloud computing became the default platform for enterprise IT — not the exception.
Five Major Cloud Shifts Heading Into 2020
While no technology forecast is guaranteed, the following five trends had clear momentum and strong structural drivers. Understanding them remains relevant because they continued to accelerate well beyond 2020.
1. Cloud Security Moves to Centre Stage
Several high-profile public cloud data breaches in 2019 forced enterprises to revisit their security posture. Up to that point, cloud security conversations had been dominated by relatively surface-level controls: identity and access management (IAM) policies, network segmentation and basic compliance checklists.
As organizations migrated increasingly sensitive and regulated workloads to public cloud platforms, the security conversation broadened to include:
- Cloud workload protection platforms (CWPPs) — runtime security for containers, VMs and serverless functions.
- Data encryption at rest and in transit — including customer-managed encryption keys (CMEK).
- Threat intelligence integration — correlating cloud telemetry with known threat actor behaviour.
- Privileged access management (PAM) — tightening just-in-time and least-privilege access to cloud control planes.
- Shared responsibility clarity — ensuring teams understood exactly which security obligations fell to the cloud provider versus the customer.
The shared responsibility model, long documented by providers such as AWS, Azure and Google Cloud, became a practical training priority rather than a footnote in vendor agreements.
2. Edge Computing and Cloud: Complementary, Not Competing
Edge computing distributes data processing to the point where data is generated or consumed — a factory floor sensor, a retail point-of-sale terminal, a connected vehicle — rather than routing everything back to a central data centre. On the surface, this appears to conflict with cloud computing’s centralised model.
In practice, the two are complementary. Edge nodes handle latency-sensitive, high-frequency processing locally. Cloud infrastructure handles aggregation, long-term storage, model training, orchestration and global policy management. A practical example:
- An IoT sensor on a manufacturing line detects an anomaly and triggers a local shutdown in milliseconds (edge).
- The anomaly event is streamed to a cloud data lake for root-cause analysis and predictive maintenance model retraining (cloud).
As edge deployments scaled up heading into 2020, a reliable and well-architected cloud backbone became a prerequisite rather than an optional add-on. Frameworks such as AWS Outposts, Azure Stack Edge and Google Anthos were direct responses to this demand.
3. Hybrid Cloud Matures Into a Genuine Operating Model
On-premises data centres were not disappearing — and for many regulated industries and large enterprises, they never would entirely. The forecast suggested that close to 50% of Indian enterprises would adopt a hybrid cloud model in 2020, a significant jump from previous years.
The historic barrier to hybrid adoption was a lack of consistency: applications behaved differently on-premises than in the public cloud, operational tooling diverged, and networking complexity created latency and reliability problems. The maturation of hybrid cloud platforms began to dissolve these barriers by providing:
- Unified control planes — a single management interface for on-premises and cloud resources.
- Consistent Kubernetes environments — enabling workloads to run and be managed identically across locations.
- Predictable latency — keeping latency-sensitive workloads close to users or data sources while offloading elastic compute to the public cloud.
- Consistent security policies — applying the same identity, network and compliance controls across environments.
The practical result was that enterprises no longer had to choose between agility and control — a hybrid model could, when properly implemented, deliver both.
4. AI and ML Find a Home in the Cloud
Artificial intelligence and machine learning investments had been rising for years, but many projects remained fragmented: isolated proof-of-concepts, siloed data sets and inconsistent infrastructure. The problem was rarely ambition — it was platform fragmentation.
Cloud platforms addressed this directly by offering managed ML services that abstracted infrastructure complexity: data ingestion pipelines, distributed training clusters, model registries and inference endpoints. This allowed data science and ML engineering teams to focus on model quality rather than cluster management.
According to a survey cited in the original analysis, approximately 67% of IT professionals at the time believed AI and ML would be the primary drivers of cloud adoption in 2020. Whether or not that specific figure proves precise, the directional trend was accurate: cloud providers competed aggressively on AI and ML tooling, and managed ML platforms (Amazon SageMaker, Azure Machine Learning, Google Vertex AI) became significant adoption drivers.
5. Serverless Computing Moves Toward the Mainstream
Serverless computing — where the cloud provider dynamically allocates and manages the underlying compute resources in response to events — had existed since AWS Lambda launched in 2014. By 2020, it had matured from an experimental pattern to a production-grade deployment model.
The core value proposition of serverless is developer productivity: teams define functions and event triggers, and the platform handles provisioning, scaling and availability. This eliminates entire categories of operational overhead.
Gartner projected that more than 20% of global enterprises would have deployed serverless technologies by 2020, compared to roughly 5% in 2018. Paul Johnston, co-founder of ServerlessDays, described serverless as “the next iteration of cloud” — a characterisation that captures how fundamental the abstraction shift is.
Common serverless use cases that gained adoption heading into 2020 included:
- Event-driven data processing pipelines (S3 triggers, message queue consumers).
- API backends with unpredictable or spiky traffic patterns.
- Scheduled automation tasks replacing cron-based scripts on dedicated instances.
- Real-time stream processing for IoT and analytics workloads.
What These Trends Tell Us About Cloud Architecture Today
Looking back from the vantage point of a few years later, the five trends above proved durable. Cloud security matured into a dedicated discipline (Cloud Security Posture Management, zero-trust networking). Edge and cloud integration became standard in manufacturing, retail and telecommunications. Hybrid and multi-cloud became the dominant enterprise model. AI/ML workloads became one of the primary justifications for cloud migration. And serverless adoption grew steadily, complemented by container-based abstractions like AWS Fargate and Google Cloud Run.
For DevOps and platform engineering teams, these trends have concrete architectural implications: security must be embedded in pipelines and platform design, not bolted on. Workload placement decisions — edge, on-premises, or cloud — must be deliberate and repeatable. And abstraction layers (serverless, managed Kubernetes, managed ML) exist to eliminate undifferentiated heavy lifting so teams can focus on product outcomes.
The structural forces that drove these trends in 2020 have only intensified since.