South Africa’s telecommunications operators are committing extraordinary sums to physical network transformation. Vodacom’s Vision 2030 programme carries an R85.2-billion price tag, and MTN’s ongoing network modernisation targets AI-driven operations, 5G coverage expansion, and core network upgrades. A new research report from Cloudera argues that the data foundations required to make those investments productive are not ready for what operators are asking of them.
The warning isn’t without merit. Cloudera’s Data Readiness Index 2026 finds that 60% of telecom leaders globally cannot access the data required for their strategic AI initiatives, despite near-universal confidence in their data visibility. The gap between what operators believe they can see and what they can actually use is the central problem the report documents, and it’s one that tracks closely with what other readiness surveys have found in adjacent areas. When Cisco assessed South African cybersecurity readiness last year, the finding was structurally similar: organisations were confident about their posture and underperforming against it in practice.
The EMEA-specific numbers add needed regional granularity. While 89% of EMEA IT leaders claim they have complete visibility into where their data resides, 42% admit that complicated access requirements act as a primary barrier to utilising it. Access friction of that scale directly affects the AI use cases South African operators are investing towards. Predictive network maintenance, automated fraud detection, and real-time customer experience personalisation all depend on data being reachable when the system needs it, not just theoretically accounted for.
Integration is the second fault line. Only 34% of EMEA organisations report their data sources are fully integrated across hybrid environments. The remaining 66% operate with fragmented data pockets that create latency issues and drive up public cloud egress fees, a point with specific financial weight locally given the rand’s volatility against dollar-denominated cloud pricing. The currency exposure makes the cost of moving large volumes of network data to public cloud infrastructure a genuine operational risk, not just a theoretical one.
This creates an awkward question for both Vodacom and MTN. Vodacom signed a multi-year partnership with Google Cloud in November 2025, with data modernisation and unification as its first strategic pillar. The deal was structured to create a unified data layer using BigQuery to support real-time insights, strengthen governance, and provide a single source of truth for business decision-making. That’s precisely the kind of data architecture Cloudera’s report suggests is missing. What the report questions is whether moving critical data to public cloud infrastructure is the right mechanism to achieve it, or whether the egress costs and sovereignty implications create new problems in solving the old ones.
Cloudera’s answer is predictable given who’s selling it. The report advocates for a Private AI approach, deploying AI models directly alongside the data rather than moving datasets across cloud environments. By bringing AI to where the data is stored, operators can protect data sovereignty, satisfy POPIA requirements, and avoid cloud egress costs entirely. That’s a direct counterposition to the public cloud data consolidation strategies both major local operators have already announced.
The governance numbers are the most consequential finding for South African operators specifically. Only 26% of EMEA IT leaders believe their enterprise data is fully governed. Globally, only 18% describe their data as fully governed, while 85% claim to have a clear data strategy. The distance between strategy and governance is where POPIA compliance risk lives. Deploying AI models on poorly governed datasets under South Africa’s Protection of Personal Information Act isn’t a hypothetical problem; it’s an audit finding waiting to happen. POPIA’s enforcement teeth have sharpened since 2021, and the Information Regulator has been increasingly active.
What gives the report more credibility than a standard vendor survey is the size of the contradiction it documents. Despite the significant barriers, 93% of telecom respondents say their executives prioritise the infrastructure necessary to enable AI at scale, and 91% have a clearly defined data strategy that supports their broader business objectives. The issue isn’t intent. Operators understand what they’re trying to do. The fundamental challenge is moving beyond perceived capability to consistent, real-world execution.
There is a reasonable counterargument to Cloudera’s framing. Vodacom’s Google Cloud deal is specifically structured around data unification and governance — it’s not a raw infrastructure play, it’s an attempt to resolve exactly the fragmentation the report describes. The question isn’t whether operators need better data governance; it’s whether public cloud consolidation or private hybrid platforms are the more effective route to it. That’s a genuine architectural debate, and Cloudera’s interest in one answer over the other doesn’t make the question less real.
The research found that 96% of EMEA organisations are willing to adopt new governance frameworks to improve data readiness. Willingness and execution are different things. But for an industry that has historically been strong on physical infrastructure and weaker on data architecture, the conversation is shifting in the right direction. The R85.2 billion going into South Africa’s networks over the coming years will only produce the AI returns operators are projecting if the data those networks generate can be accessed, governed, and used at the speed the applications require.