Mark Walker, director and co-founder of Transformation for Innovation (T4i).
Organisations should sequence AI spend as incremental to cloud modernisation, not as a parallel track – particularly because security, compute and financial control are all moving faster than the traditional allocation models can absorb.
This is according to Mark Walker, director and co-founder of Transformation for Innovation (T4i), who delivered a keynote at the ITWeb Cloud Summit 2026 in Riversands, Johannesburg, this week.
Walker pointed out that AI deployment accounts for a 20% share of the South African enterprise CIO IT budget in 2026, projected to reach 50% by 2030. The figures are based on T4i analysis.
He used cyber security allocation to illustrate his point: in traditional cloud, it represents a 5%-7% maintenance overhead, while the AI-era requirement is 10%-12% of total IT spend for zero trust and edge protection.
“Underfunding becomes exposure. Security can no longer be treated as a residual operating cost,” he added.
No more ‘lift-and-shift’
Walker said the era of basic “lift-and-shift” cloud migration is officially over in SA.
Driven by hyperscale data centre investments and the rapid maturation of GenAI, the local cloud market is projected to more than double, surpassing R100 billion by 2029, according to figures Walker presented at the summit.
This transition into AI environments presents a massive paradox for local enterprise leaders, said Walker.
“While the infrastructure imperative demands high-density, GPU-centric cloud environments to run autonomous workloads, CFOs are simultaneously battling severe cloud bill shock driven by US dollar-denominated pricing and rand volatility. Furthermore, deploying AI without deep systems integration and strict human guardrails is a recipe for systemic failure.”
Walker said SA recorded marginal year-on-year growth in ICT, which, according to ICASA’s State of the ICT Sector March 2026 report, stood at 0.8% in 2025.
This growth is supported by increasing demand for cloud and AI infrastructure, particularly in sectors such as banking, financial services and retail.
Research shows that AI adoption and use is increasing. According to the Global AI Diffusion Q1 2026 Trends and Insights report released by Microsoft, SA is ranked 46th out of 147 economies measured for AI adoption. The research estimates that 23.1% of SA’s working-age population used AI in Q1 2026, up from 21.1% in the second half of 2025.
“Cloud is the foundation of effective AI, and a mechanism to grow productivity,” said Walker, who added that productivity and efficiency have been identified as a constraint to AI cloud execution, followed closely by a shortage of technical skills, data readiness and quality, integration with legacy systems, and security, compliance and POPIA.
Walker said SA’s productivity rate – output per person – is lower than that of direct competitor markets such as China, India, Indonesia and Vietnam.
On productivity and jobs, the current picture shows micro-gains in developer velocity (eg, copilots) and basic customer-service automation.
However, Walker cited research from Stellenbosch University Business School, the Sanlam Benchmark 2026, the Altron Employee TEXIndex 2026 and T4i analysis showing that R30 million per 1 000 employees is lost annually to South African operational friction – any stage of business operation that is halted, slowed, stopped or repeated.
Cloud and AI present the primary structural mechanism to reclaim this value, he said.
T4i estimates, drawing on analyst research from Gartner, IDC, BMI-T and others, that 2027 will bring operational efficiency gains in back-office tasks, including finance, compliance and legal document indexing.
By 2031, T4i forecasts that the market will reflect hyper-productivity, with routine middle-management and clerical tasks fully run by autonomous AI workflows.
Walker stressed the need for stronger collaboration between CIOs and CFOs in driving an AI agenda.
“The CIO drives architecture and control, the CFO hardwires commercial discipline. Both own the outcome,” he added.
Walker said the CIO’s role comes down to three things. First, bring models to the data by keeping sensitive data in place and cutting unnecessary egress. Second, make the estate API-ready by modernising any core system an autonomous agent cannot query. Third, install an AI gateway to enforce identity, policy and logging at the edge, and shut down shadow AI.
For the CFO, Walker said three priorities stand out: make FinOps non-negotiable through hard caps, tagging and right-sizing before provisioning; hedge currency exposure by contracting in rand and locking in longer terms where the economics justify it; and buy outcomes rather than seats by rejecting per-user AI licences and tying fees to measurable business value.
He also urged organisations to stabilise power and energy, secure compute access, redesign for productivity and mobilise the workforce. That means ensuring grid resilience and dynamic load management, securing green hyperscale cloud capacity for priority workloads, rebuilding workflows around autonomous AI and measurable outcomes, and funding rapid upskilling and talent attraction. Organisations should also close the shadow-AI gap before the regulator forces the issue, and put governance structures in place ahead of the AI policy rather than in reaction to it.
