MTN’s migration of its Enterprise Value Analytics (EVA) platform to Microsoft Azure completes a pivotal step in the operator’s multi-year digital transformation, delivering a cloud-native analytics backbone that the company says processes tens of billions of records daily, supports real‑time decisioning, and acts as a repeatable blueprint for telco modernization across Africa and the Middle East. The move—marketed as EVA 3.0—repackages MTN’s telemetry, OSS/BSS feeds, billing and customer signals into a Databricks-powered lakehouse running on Azure and protected by Microsoft Defender, promising faster insight-to-action paths for network operations, customer care, product teams, and AI initiatives.
Background
MTN’s cloud-first strategy, built under the banner Project Nephos and supported by a centralized Cloud Centre of Excellence (CCoE), has steadily pushed the group toward platform-led operations and data-driven services since the partnership with Microsoft began in 2022. The EVA migration is the most visible execution of that strategy: re-architecting an enterprise analytics system to run as a cloud-native lakehouse that emphasizes streaming plus batch analytics, model operationalization, and centralized security controls. MTN frames this as both an operational necessity—real-time visibility into a sprawling network—and a strategic capability that unlocks AI-enabled services and downstream monetization. The announcement arrived with headline metrics and executive commentary intended to signal scale and credibility. MTN describes EVA 3.0 as the “largest telco cloud implementation in the Middle East and Africa,” citing daily throughput and workflow counts, while Microsoft personnel highlight the role of hyperscale cloud and AI to accelerate telco digital transformation. These claims are echoed across trade media and industry outlets, reflecting broad media amplification of the corporate release.
What EVA 3.0 is and how it’s built
A cloud-native telco lakehouse
EVA 3.0 is described as a lakehouse architecture implemented on
Azure Databricks, backed by Azure storage and platform services. The lakehouse pattern combines the scale and cost-efficiency of data lakes with structured layering, ACID transactions and performance optimizations associated with data warehouses—enabling telco teams to run streaming ingestion, Spark-based transformations, ML model training, and low-latency analytics in a single fabric. MTN’s messaging emphasizes the data engineering and orchestration advantages of Databricks for complex telco workloads. Key architectural elements reported include:
- Ingestion and event streaming from network probes, call detail records, OSS/BSS systems and application telemetry.
- Persisted raw and curated layers (Delta Lake or equivalent) stored on Azure Data Lake Storage for cost-efficient scale.
- Compute and transformation using Azure Databricks (Spark), with orchestration across hundreds of analytic pipelines.
- Runtime protection and platform security provided by Microsoft Defender and Azure security controls to centralize threat detection and governance.
Why Databricks + Azure?
Databricks accelerates batch-to-stream convergence with first-class Spark optimizations, native Delta Lake semantics, and a managed runtime that simplifies cluster management at scale—capabilities that address telco needs for both throughput and consistency. On Azure, Databricks integrates with identity, networking, and security services, enabling secure data access patterns across MTN’s multi-region footprint. The vendor choices reflect a common hyperscaler + lakehouse pattern increasingly adopted by large network operators when replatforming OSS/analytics stacks for real-time operations.
Scale, performance, and the numbers that matter
MTN’s announcement contains several headline metrics designed to quantify EVA 3.0’s scale:
- Approximately 22 billion records processed per day.
- More than 800 analytics workflows in production.
- Integration of roughly 1,700 data feeds into the platform.
These figures are repeated in multiple independent trade and regional news reports following MTN’s release, which increases confidence that the numbers are MTN’s official, company‑reported metrics. Independent technical validation beyond corporate disclosures—such as third‑party audits or published performance benchmarks—has not been made public, so these remain
company‑reported operational metrics rather than independently verified measurements. Readers should treat the numbers as credible but self‑reported; they indicate scale and ambition but not precise, audited throughput characteristics. MTN and Microsoft state that EVA 3.0 delivers “significant improvements” in processing speed compared to the previous on‑premises setup. Faster ingestion, reduced time-to-insight, and earlier detection of service incidents are called out as primary operational outcomes—advantages consistent with cloud-native replatforming when architectures and ETL pipelines are modernized for stream processing and autoscaling compute. The exact performance delta (for example, percent reduction in pipeline latency or cost per record) is not disclosed in public materials, so precise benchmarks remain internal.
Operational benefits: from NOC to personalised offers
Earlier detection and faster remediation
By processing telemetry at higher velocity and providing near real-time analytics, EVA 3.0 offers network operations teams earlier visibility into degradations, congestion, and outage patterns. This should shrink mean time to detect (MTTD) and mean time to repair (MTTR) for many classes of faults, particularly those that are visible in streaming network probes and CDR-derived KPIs. The platform’s value accrues where rapid detection can reduce customer-impacting incidents and enable proactive capacity actions.
Smarter segmentation and personalization
EVA 3.0’s unified data layer enables marketing, product, and experience teams to combine network quality signals with usage and billing data to target offers more precisely and to surface timely retention or upsell opportunities. When properly governed, this kind of integration supports higher relevance in customer communications and more dynamic, context-aware product experiences, increasing the potential for revenue uplift and better NPS.
Platform for responsible AI
MTN frames EVA 3.0 as a foundation for
responsible AI—a phrase that combines governance controls, data protection, and model lifecycle management. A centralized analytics backbone simplifies model versioning, monitoring and rollback, and allows security and privacy teams to apply consistent policies across datasets used for training. However, responsible AI requires transparent governance, bias testing, and operational monitoring—areas where public disclosure in MTN’s release is limited. The platform can enable responsible processes, but the announcement does not provide operational detail on governance workflows, model explainability tools, or privacy-preserving techniques.
Security posture and data protection
MTN highlights
Microsoft Defender and Azure security controls as core to EVA 3.0’s protection model. Centralized security services on Azure provide capabilities such as threat detection, endpoint protection for managed compute, logging and SIEM integration, identity and access management via Azure AD, and role-based access controls required for telco compliance. These components are industry-standard building blocks for securing cloud telemetry and analytics workloads. That said, securing a telco-grade analytics platform goes beyond cloud-native tools. It requires:
- Strong network segmentation between production, analytics, and developer environments.
- Robust data classification, encryption at rest and in transit, and key management processes.
- Operational controls for ML model governance, lineage, and drift detection.
- Regular independent security assessments and compliance mapping to regional data protection laws.
MTN’s public materials affirm use of Defender and Azure controls but do not publish an independent security assessment or compliance mapping. External auditors or regulators typically hold the detailed evidence for compliance, so the high-level security claims should be regarded as credible but not independently validated in public.
People, skills, and the Cloud Centre of Excellence
MTN reports a deliberate, large-scale investment in cloud skills through its Group‑wide CCoE and the Microsoft Enterprise Skills Initiative. The company has publicly stated that it achieved more Microsoft Azure certifications in 2025 than any other organisation in Africa, with
over 1,350 certifications attained to date—figures that underscore the human-capability side of sustaining a complex cloud platform. Training and certification are essential to operate, optimize, and secure a cloud-native lakehouse at telco scale. Upskilling at this level addresses a perennial telco challenge:
the sourcing and retention of cloud and data engineering talent. However, certification counts alone do not guarantee operational maturity. Sustainable success depends on:
- Depth of hands-on experience (not just exam passes).
- Clear runbooks and operational playbooks for incident response.
- Cross-functional alignment between network, data, security, and product teams.
- Ongoing investment in advanced specialties such as site reliability engineering (SRE), edge analytics, and data governance.
MTN’s certification figures are a meaningful indicator of commitment and scale, but the real-world proof will be in MTTR improvements, model performance in production, and the ability to roll the architecture out across diverse regulatory environments.
Strategic implications: a blueprint for African telcos?
MTN presents EVA 3.0 as a repeatable model for other MTN markets, asserting that the architecture can be adapted across country operations to drive inclusion, resilience, and digital services. This is a logical goal: central architectures and reusable platform components reduce duplication of effort and accelerate time-to-market for AI-driven services in emerging markets. The partnership also benefits from Microsoft’s growing cloud footprint in Africa, including investments in local regions and partner programs that reduce latency and sovereignty friction. From a regional perspective, a successful, replicable model matters because:
- Many African telcos still run fragmented, on‑premises OSS/BSS landscapes.
- Hyperscaler investment in regional clouds lowers the technical barrier for massive analytics projects.
- A proven reference architecture from a major operator reduces perceived risk for peers considering similar transitions.
However, replication is rarely plug-and-play. Local technical teams must address regulatory differences, data residency rules, bandwidth constraints, and cost models that vary widely across countries. The platform is a credible template, but operational replication will require adaptation and investment in each local market.
Risks, questions and what the announcement leaves unsaid
The EVA 3.0 rollout is important, but the public narrative omits several details that enterprise architects and regulators will scrutinize:
- Vendor and cloud concentration risk. Heavy dependence on a single hyperscaler and integrated Databricks runtime creates operational efficiencies but increases exposure to vendor pricing, platform changes, and contractual terms. Multi-cloud or hybrid escape plans and careful contract structuring are prudent risk mitigations.
- Cost transparency. Large-scale cloud deployments can shift capital expense (CAPEX) to operational expense (OPEX) while improving agility. But without published cost-per-terabyte or cost-per-workflow metrics, it’s impossible to quantify total cost of ownership versus legacy systems.
- Data sovereignty and regulatory compliance. The announcement does not detail how cross-border data flows, lawful interception, and local data protection requirements are handled across MTN markets—matters that telecommunications regulators often require to be explicit.
- Independent validation of scale and performance. The throughput and workflow numbers are company-reported and echoed by media. Independent third-party audits or benchmark studies would strengthen confidence in those claims.
- Model governance and privacy safeguards for AI. MTN emphasizes responsible AI, yet the public statement lacks specifics on fairness testing, privacy-preserving ML techniques, and external oversight—areas of growing regulatory and public concern.
- Sustainability and energy footprint. Large-scale compute for analytics and ML has a measurable carbon and energy profile. The public materials do not disclose energy or sustainability metrics or how Azure’s regional infrastructure choices impact MTN’s carbon reporting.
Each of these areas is solvable with proper governance, contracts, and operational discipline—but they are the natural follow-up questions that enterprise users, customers, and regulators will need answered.
A critical read on the “largest telco cloud implementation” claim
MTN’s characterization of EVA 3.0 as the “largest telco cloud implementation in the Middle East and Africa” is a strong marketing claim that has been widely repeated. The phrase is defensible as a company statement of scale, but it is ultimately a comparative assertion that depends on how “largest” is measured—records processed, number of workflows, data volume, or region footprint. Publicly available independent audits or cross‑operator comparisons that would conclusively validate that ranking are not available, and media reporting repeats the claim as company-sourced. Readers should therefore treat the “largest” claim as a company-declared milestone rather than an independently verified industry ranking.
How operators should evaluate similar projects
Operators considering a move to a Databricks-plus-hyperscaler lakehouse should evaluate success across both technical and non-technical axes. Key evaluation criteria include:
- Technical fit and future-proofing:
- Can the platform support both streaming and long-term analytical workloads at acceptable latency and cost?
- Is the data architecture resilient to node failures, and does it support disaster recovery across regions?
- Cost and commercial terms:
- Are pricing models (compute vs storage) aligned with traffic patterns?
- Do contracts contain predictable pricing, volume discounts, and escape clauses for future vendor changes?
- Security and compliance:
- Are encryption, key management, and access controls implemented end-to-end?
- Is there a compliance framework mapping platform controls to local regulation?
- People and processes:
- Does the operator have an achievable upskilling plan with hands-on training?
- Are runbooks, incident response and SRE practices in place?
- Measurable business outcomes:
- Are there KPIs for MTTR, customer satisfaction, revenue uplift from personalization, or fraud reduction that will be tracked and published?
Operators should require independent performance verification where possible and structure projects in phases with clear acceptance gates to reduce risk while delivering measurable outcomes early.
The continent-wide angle: inclusion, infrastructure, and local capability
MTN positions the EVA 3.0 rollout as more than an internal modernization: it’s presented as an enabler of broader digital inclusion and a step toward deeper technological sovereignty across Africa. By investing in both cloud architecture and local skills, the company claims to be strengthening the continent’s digital foundations—helping governments, enterprises and consumers benefit from more resilient and data-driven services. Whether that promise materializes will depend on downstream outcomes: lower latency services, local data processing for public-sector programs, and broader availability of advanced digital services for small businesses and developers. Realistically, the transition also depends on Microsoft’s regional infrastructure investments—data centre regions, edge locations, and partner ecosystems—to ensure services can run with acceptable latency and residency controls across multiple African markets. The combination of a major operator building a reusable platform and a hyperscaler expanding capacity creates a practical path to deliver higher-order digital services more widely.
Conclusion: significant progress, measured expectations
MTN’s EVA 3.0 migration to Microsoft Azure is a consequential milestone for African telecommunications. It demonstrates a credible, enterprise‑scale adoption of a lakehouse architecture and hyperscaler services to ingest, process, and operationalize the torrent of data generated by modern networks. The company’s claims—22 billion daily records, 800+ analytic workflows, 1,700 data feeds, and a large corps of Azure‑certified engineers—are consistent across MTN and multiple independent trade reports and signal both scale and ambition. At the same time, several elements remain appropriately opaque in the public narrative: detailed performance benchmarks, independent validation of the “largest” claim, granular security assessments, cost breakdowns, and operational metrics that demonstrate sustained business outcomes. Those are not weaknesses of the architecture itself, but they are the yardsticks by which the broader industry—and regulators—will judge whether EVA 3.0 is a durable, replicable model or primarily a large, well-executed migration.
For technology leaders and telco architects, the real lesson is pragmatic: modernizing telco analytics at scale requires an aligned combination of cloud architecture, vendor discipline, robust security and compliance frameworks, and serious investment in people. MTN’s EVA 3.0 is a useful, high‑visibility case study that will inform competitor strategies and vendor offerings across Africa and beyond—provided the next phase of transparency includes audited performance metrics, cost and sustainability disclosures, and clear governance for AI and sensitive customer data.
Source: Communications Today
MTN powers real-time intelligence at large scale through Microsoft Azure partnership | Communications Today