Why Companies Are Returning and How Demand Is Formed
After 2022, many international companies suspended Russian operations: some sold their business, others froze assets, some maintained legal entities on pause. Since then, the market has adapted: new hardware supply channels emerged, Russian cloud platforms matured, and operational constraints were navigated. Now we see a reverse trend: companies — cautiously, step by step, often without public announcements — are exploring return options.
A key feature of this market: demand is almost never framed as "we need service X". It comes as tasks: "how to quickly deploy infrastructure", "where to move services", "which data center to choose", "how to restore employee access", "where to store personal data now". Behind each task lies an entire project — with architecture, procurement, migrations, legal requirements, and yesterday deadlines.
Three Typical Return Scenarios
Although every project is unique, most returns fall into three scenarios:
Scenario 1: "Warm Return" — Partial Infrastructure Remains
The company kept its legal entity, some equipment, possibly an office and staff. Services were moved to foreign clouds or frozen. Task: audit remnants, determine what can be revived, move critical services back to Russia, restore access and documents. Fastest scenario: working infrastructure in 3–6 weeks.
Scenario 2: "Cold Start" — Starting from Scratch
Legal entity liquidated or sold, no equipment, no team. Everything needed: legal structure, office, equipment, cloud or data center, network, domain, email, workstations. Realistic timeline: 6–12 weeks to full operation of the first office.
Scenario 3: "Insurance" — Standby Site Without Immediate Return
Return decision not yet made, but board wants readiness: backup site in Russia, tested migration plan, clear budget and timeline. We design target architecture, deploy minimal core (domain, VPN, critical data replicas), and conduct quarterly recovery tests.
12-Week Roadmap: From Decision to Launch
| Stage | Weeks | Deliverables |
|---|---|---|
| 1. Audit & Architecture | 1–2 | Approved project plan and timeline |
| 2. Sites & Procurement | 2–5 | Signed contracts, hardware in transit |
| 3. Basic Contour | 3–6 | Working infrastructure core |
| 4. Services & Data | 5–9 | Critical services in production |
| 5. Workstations | 7–10 | Employees operational |
| 6. Stabilization | 9–12 | SLA, operations, growth plan |
Deployment Model: Cloud, Colocation, or Hybrid
Start in the cloud, stabilize in hybrid. Cloud gives speed at launch; when loads stabilize, heavy persistent parts (databases, storage, backups) are more economical on dedicated hardware in a data center, while cloud keeps frontends, peaks, and managed services.
Russian Cloud Provider Comparison
| Provider | Strengths | Considerations |
|---|---|---|
| Yandex Cloud | Broad managed services, strong data/ML ecosystem | Cost at high loads, default limits |
| VK Cloud | Rapidly growing, strong K8s and DB services | Younger ecosystem |
| SberCloud | Mature platform, enterprise and public sector focus | Onboarding may take longer |
| MTS Cloud | Cloud + telecom integration, own data centers | Check managed services availability |
| Selectel | Cloud + colocation + dedicated in one contour | Fewer managed services than hyperscalers |
Data Centers: What to Look for in 2026
Key players: DataLine, IXcellerate, Selectel, 3data, MMTS-9 facilities. Evaluate each site on: Tier certification, power redundancy (N+1/2N), connectivity providers, remote hands SLA, physical security, documentation, and total cost of ownership.
Hardware: Procurement, Logistics, Warranty
Russian-assembled servers (Yadro, Kraftway), parallel import for known brands, and quality refurbished equipment. We maintain a compatibility matrix and verify stock availability before specifying. Typical lead times: 1–2 weeks for stock items, 4–10 weeks for ordered items.
Network, VPN and Connectivity
Office networks with VLAN segmentation, site-to-site VPN (IPsec/WireGuard), dedicated L2 channels for critical loads, remote access with MFA or ZTNA, DDoS protection from day one. Document everything from day one.
Core Services: Active Directory, Email, Workstations
For "cold starts" we recommend a new domain with clean rights model and user migration by approved lists. Email: SPF/DKIM/DMARC, anti-spam, archiving. Workstations: golden OS image, disk encryption, EDR, MDM — new employee ready in one hour.
Data and Regulations: 152-FZ, Cross-Border Transfer
152-FZ requires PDn processing databases to be located in Russia. Practical scheme: Russian "core" with personal data (HR, CRM, client bases), global systems receive only anonymized or aggregated data. We design the technical scheme; your lawyers define the legal requirements.
Security and Resilience
Network segmentation, MFA for all admin access, EDR on all endpoints, 3-2-1 backup with isolated copy and regular restore tests, patch cycle, DDoS protection, admin activity logging. RTO/RPO defined with business, not IT.
Common Return Mistakes
- Starting with hardware purchase instead of architecture
- Relying on a single "hero" engineer
- Ignoring data requirements until the last week
- Not testing backup restores
- Forgetting about access credentials
- Choosing provider based on presentation only
- Not planning for ongoing support
- Chasing perfection instead of "good enough"
Checklist: IT Return to Russia
- Project owner assigned and budget approved
- Scenario defined: warm return / cold start / standby
- IT audit completed: equipment, data, access, licenses
- Legal requirements assessed (152-FZ, sector-specific)
- Deployment model chosen: cloud / colocation / hybrid
- Cloud provider or data center selected and contracted
- Hardware specification aligned with supply channels
- Network designed: segmentation, VPN, DDoS
- Active Directory plan: new domain or restore
- Backup 3-2-1 configured with isolated copy
- MFA enabled for all administrative access
- Documentation complete and support model defined
- Load and failover tests passed before full launch