New Russian B2B Solutions Enter the Cloud Services and Virtualization Market
The Russian IT market has been enriched with several new B2B products. Turbo Cloud has introduced a service based on the Basis ecosystem, Selectel and Deckhouse have announced a joint Kubernetes solution, and MWS Cloud has launched a PaaS service, Managed Kafka, for working with streaming data and event-driven architectures.
Seeing the simultaneous announcement of Managed Kafka from MWS Cloud, the Kubernetes service from Selectel and Deckhouse, and the cloud platform on Basis from Turbo Cloud, professional experience immediately suggests: these are not three disparate launches but a coordinated transition of the Russian PaaS market into a phase I call "vertical expansion with a focus on data-driven architectures."
[Essence]: What Is Really Happening
The Russian cloud market is completing a structural shift from the "virtual machine rental" model (IaaS) to the "platform as a service" model (PaaS). The fact that these announcements are coming almost simultaneously indicates that the industry has reached a critical mass of maturity. Providers have realized: IaaS margins are falling, and differentiating by gigabytes and core count is no longer possible. New margins are forming at the level of services that manage data flows and containers.
Selectel and Deckhouse (via Flant) have launched DKaaS—a ready-made platform for deploying, operating, and scaling Kubernetes in the cloud. This is not just "another managed k8s" but a strategic maneuver to capture the segment of enterprise clients who need hybrid scenarios: DKaaS's global cluster router allows combining cloud and on-premise infrastructure with centralized management of dozens of clusters.
In parallel, MWS Cloud has brought Managed Kafka into production—a fully managed Apache Kafka service for streaming data. The timing is critical: late April 2026 is the moment when companies that have completed their basic infrastructure migration begin seeking tools for real-time analytics, microservice integration, and IoT.
The third piece of the puzzle is Turbo Cloud with a service on the Basis ecosystem, which addresses the import substitution of virtualization tools. According to data from the CNews conference, the share of domestic solutions in this segment has grown from 26% to 60%.
Timeline and Context
The timeline I reconstruct looks like this:
2022-2023: First wave of migration—companies urgently leave Western platforms (VMware, AWS, Confluent). Demand is for basic IaaS and virtualization.
2024-2025: Stabilization phase. Businesses adopt Russian solutions and accumulate operational experience. At this stage, it becomes clear that "just virtual machines" are not enough.
September 2025 – January 2026: Regulatory deadlines. Large companies are required to import-substitute virtualization tools (by September 1, 2025) and database management tools (by January 2026). This spurs the market.
April-May 2026: The very announcements we are analyzing. The Deckhouse Kubernetes Conf in Moscow on April 9 served as the launchpad for DKaaS. MWS Cloud announced the GA of Managed Kafka on April 20.
A parallel trend I track from industry sources: demand for GPU computing in public clouds is growing 45% year-over-year, and for platform AI/ML services, 35%.
Who Wins and Who Loses
Winners:
- Selectel and Flant (Deckhouse): DKaaS is designed to offload all routine tasks from development teams—from control plane administration to updates and support for basic modules. SLA 99.98%, compliance with Federal Law 152 and PCI DSS, ISO/IEC 27001 certification—these are arguments that tip the scales in favor of a Russian solution for compliance-oriented clients. Deploying a cluster with GPU for AI tasks takes less than 40 minutes.
- Enterprise customers (fintech, retail, industry): They gain infrastructure for real-time scenarios that previously required either expensive Western expertise or compromise batch processing. MWS Cloud's Managed Kafka, with support for multinode clusters and Private Link, provides private connectivity within the cloud and deploys in minutes.
- Russian virtualization vendors (Basis, RED SOFT, etc.): The virtualization market has entered a consolidation phase. Demand is shifting from functionality to stability and compatibility.
Losers:
- Western PaaS providers (AWS, Azure, Confluent): Their Russian segment is shrinking to the limit. Each new announcement of DKaaS or Managed Kafka reinforces the sovereign standard.
- Companies that delayed migration: The gap between those already building event-driven architectures on Managed Kafka and those still migrating virtual machines is becoming competitive. The cost of delay is rising.
What the Media Isn't Saying
Now for the promised non-obvious insight, which I derived from analyzing technical documentation and market signals. No one openly states that these launches are preparation for the era of AI-driven operations, where PaaS becomes not infrastructure but an "operating environment" for artificial intelligence agents.
Here's why. MWS Cloud's Managed Kafka is tailored for Kafka → ClickHouse / DWH / ML pipelines and IoT telemetry. DKaaS from Selectel and Deckhouse offers cluster deployment with GPU for AI tasks. This is no coincidence. Companies are building infrastructure where streaming data from thousands of devices and microservices flows into Kafka, is processed by ML models in Kubernetes clusters, and the results influence business decisions in real time.
But there is a deeper layer. 76% of cyberattacks today target cloud storage. Managed Kafka is designed with isolation at the virtual machine level to reduce the impact of neighboring workloads. DKaaS supports hybrid scenarios where the critical core remains on-premise and AI workloads go to an elastic public perimeter. Together, these solutions create an architecture that solves not today's problem but tomorrow's: how to safely run AI agents that will have access to a company's streaming data. Security shifts from perimeter to Zero Trust embedded in the PaaS architecture.
A second underestimated point: the pricing model. DKaaS operates on a pay-as-you-go basis, allowing payment only for actually used resources and creating clusters for short-term projects of 3-6 months. This radically changes the approach: companies can test AI hypotheses without capital expenditure. A successful hypothesis scales; a failed one is shut down without losses. This is exactly the flexibility the Russian market lacked in the IaaS phase.
Forecast: Next 30 Days and 90 Days
Next 30 days (until June 5, 2026):
Selectel and Flant will announce the first enterprise deployments of DKaaS—likely in fintech or retail, where the demand for real-time analytics is most acute. MWS Cloud will actively promote Managed Kafka through the MTS partner network, emphasizing integration with Private Link for closed perimeters. I expect at least one major contract for building an event-driven architecture for an industrial IoT project.
Next 90 days (until August 4, 2026):
The Russian PaaS market will move toward consolidation around three or four ecosystems. Selectel/Deckhouse, MWS Cloud, and Basis will form competing standards, and enterprise clients will have to choose an ecosystem rather than a product. A key signal will be the emergence of the first cross-cloud solutions that allow managing a Kafka cluster in one ecosystem and a Kubernetes cluster in another.
But the main shift I predict by August: one of the major Russian banks or retailers will publicly announce the launch of an AI agent fully running on the Russian PaaS stack (Managed Kafka + DKaaS + domestic virtualization). This event will be a point of no return, cementing the ability of sovereign cloud infrastructure to serve not just legacy systems but cutting-edge data-driven products.
— Editorial Team
No comments yet.