Microsoft’s decision to lead an AI and data adoption masterclass for African energy companies at African Energy Week 2026 puts a practical question at the center of the continent’s next infrastructure cycle: how can operators use cloud platforms, connected devices and advanced analytics to improve production and grid reliability without importing new operational, cyber and resource risks? The session is scheduled for October 12, 2026, at African Energy Week in Cape Town, with an agenda focused on AI-enabled operational excellence, energy-transition platforms, cybersecurity, data governance and responsible scaling. African Energy Week’s conference listing
The Microsoft Masterclass: AI & Data Adoption for African Energy Companies is more than another cloud-computing presentation on a conference agenda. It arrives as energy producers, utilities, developers and governments confront overlapping requirements: increase output, reduce downtime, integrate more variable renewable generation, protect industrial systems, and build the digital capacity needed to run increasingly automated operations.
African Energy Week’s announcement frames the masterclass as a practical session for upstream oil and gas, power and renewable-energy organizations. Microsoft Corporate Vice President for Energy Darryl Willis is expected to lead the discussion alongside industry participants, with the stated ambition of showing organizations how Azure-based services can support day-to-day operational transformation. African Energy Week
That focus matters because “AI in energy” is often presented in overly broad terms. For a utility, the priority may be forecasting demand and identifying network constraints. For an upstream operator, it may be interpreting seismic data faster, detecting anomalous equipment behavior or integrating data from a scattered fleet of wells and facilities. For a renewable developer, it may be coordinating batteries, solar assets and market signals in near-real time.
These are distinct workloads, with different latency, governance and safety requirements. A useful energy AI strategy cannot begin with a chatbot or a generic automation target. It has to begin with the quality of operational data, the location of the systems being connected, and the level of control the organization is prepared to delegate to automated processes.
Microsoft’s Azure Data Manager for Energy is designed around that integration problem. The platform combines Azure with the OSDU Data Platform standard and is positioned as a managed, industry-specific data foundation intended to reduce silos, support data federation and improve interoperability with first- and third-party energy applications. Microsoft Learn
For African energy companies, the significance is not merely that data can be stored in the cloud. The larger opportunity is to create a governed operational picture across exploration, drilling, production, maintenance, trading, emissions reporting and grid operations. If a company cannot reliably identify the source, ownership, quality and permitted use of a dataset, it will struggle to move from isolated dashboards to dependable AI-supported decisions.
That is why the scheduled masterclass emphasis on “how to start” adoption tracks is arguably more important than any single Azure feature. African Energy Week describes the session as covering customer examples, solution pathways and adoption tracks, rather than treating AI as a switch that can be turned on across an enterprise. African Energy Week
A cloud platform can provide elasticity for those workloads, but the value proposition depends on workload design. Microsoft says Azure Data Manager for Energy can autoscale for changing workload requirements and includes managed operations, security updates and support for multiple data partitions. Microsoft Learn
The opportunity is clear: a well-run cloud environment can shorten the time between raw operational data and an engineering decision. In exploration, that can mean more efficient evaluation of prospective basins. In production, it can mean moving from calendar-based maintenance toward risk-informed maintenance. In utilities, it can mean identifying issues earlier instead of relying only on post-fault investigation.
But cloud capacity is not the same as operational readiness. Models need accurate historical data, current telemetry, clear engineering ownership and a process for challenging recommendations. A predictive-maintenance model may flag a pump or compressor as high risk; it should not become an unquestioned authority over shutdown decisions in a safety-critical environment.
For upstream and midstream companies, AI-assisted reliability programs can combine vibration, temperature, pressure, flow and electrical data with maintenance history. The goal is to prioritize inspections, identify abnormal operating behavior and schedule interventions before a minor fault develops into an expensive outage.
This approach can be especially attractive where skilled maintenance teams cover large geographies or where spare parts and specialist contractors have long lead times. Still, the quality of the result rests on the quality of the inputs. Incorrect instrument calibration, incomplete maintenance records or inconsistent asset identifiers can turn an apparently advanced AI initiative into a faster way to repeat old data errors.
A responsible deployment therefore needs more than a model. It needs:
However, an AI model trained on one system’s historical load patterns cannot simply be transplanted into another. Electricity demand is shaped by local tariff structures, weather, industrial activity, informal supply constraints, grid topology and outage history. A sophisticated model with poor local context can create false confidence.
The better ambition is decision support, not automatic dispatch by default. AI can help planners recognize patterns and test scenarios. It can assist control-room teams by surfacing relevant signals. But dispatch, protection settings and contingency actions must remain subject to explicit engineering governance, especially where an error could affect customers, equipment or public safety.
The technology opportunity is real, but virtual power plants are not simply a software project. Their effectiveness depends on communications reliability, inverter compatibility, commercial agreements, meter accuracy, regulatory permission and an operational model for dealing with assets that do not respond as expected.
For utilities and developers, that means the software layer should be evaluated alongside the physical and contractual system. AI can optimize an incomplete or poorly governed asset pool only up to a point. It cannot compensate for unreliable communications, unclear dispatch rights or inadequate cyber controls.
This is relevant to African Energy Week because it demonstrates that digital transformation has a physical footprint. AI-ready infrastructure requires electricity, network connectivity, cooling, substations, backup systems and land-use planning. Data centers are not abstract “cloud” assets; they can become major industrial loads with long-lived implications for local networks.
The reported 174 MVA figure should be treated carefully. MVA and MW are not identical units, and the final real-power demand depends on the facility’s power factor and design. More importantly, the project’s reported water use remains undisclosed, and objections have focused on incomplete information around cooling, energy supply, emissions and backup generation. EnergiAfrica
That does not negate the economic case for hyperscale infrastructure. It does reinforce a basic principle: digital development must be planned as part of an energy system, not as a separate demand category that can be addressed after construction decisions are made.
But that potential should not be assumed. Data-center operators typically prioritize availability, and energy operators must understand the difference between a workload that can be delayed and a critical service that cannot. Any arrangement involving demand flexibility needs enforceable service levels, transparent incentives and contingency planning.
For the energy sector, the key lesson is that AI adoption and energy planning are converging. Electricity systems will increasingly power the data centers that run AI workloads, while AI platforms will increasingly help plan and operate those same electricity systems. The feedback loop is economically promising, but it makes integrated governance indispensable.
Microsoft describes ExpressRoute Metro as a high-resiliency configuration that connects an organization through two distinct peering locations in the same city, creating a dual-homed design intended to improve availability compared with a connection dependent on one edge location. Microsoft’s current documentation lists both Teraco JT1 and Africa Data Centres JHB1ADC in the Johannesburg Metro configuration. Microsoft Learn
For energy companies, this is not merely a networking footnote. A cloud-based analytics platform may be of limited value if field teams, regional offices and operations centers cannot access it reliably. More critically, organizations need to distinguish between:
Microsoft’s own documentation notes that standard redundant ExpressRoute links may still be vulnerable to disruptions that isolate an edge location; the Metro design addresses this with redundancy across two peering locations in a city. Microsoft Learn That distinction is valuable, but it should be part of a broader resilience architecture rather than treated as a complete answer to operational continuity.
Microsoft’s Defender for IoT platform is designed for agentless monitoring of IoT and operational technology environments. Microsoft says the service can identify devices, communications and industrial protocols, while integrating with security operations tools and supporting cloud-connected, on-premises and hybrid deployment approaches. Microsoft Learn
That architecture matters because many industrial devices cannot safely run endpoint agents. Operators may have legacy programmable logic controllers, sensors, gateways and engineering workstations that must be monitored passively rather than altered with standard IT security tooling.
Microsoft also documents integration between Defender for IoT and Microsoft Sentinel, designed to help security teams analyze, investigate and respond to operational technology incidents through more unified workflows. Microsoft Learn
An effective Azure security program for critical energy infrastructure should include:
Microsoft notes that Azure services generally allow customers to specify the region in which customer data is located, while some resiliency configurations may replicate information to other regions within the same geography. Microsoft Learn Organizations must therefore map their own regulatory requirements to the specific Azure services, regions, replication options, encryption settings and access controls they plan to use.
Microsoft’s sovereignty guidance also identifies Azure Policy as a way to enforce controls related to data residency and allowed services. Microsoft Learn That is useful, but implementation still requires local legal interpretation and an accurate inventory of the data being processed.
The responsible-AI dimension is equally important. An energy organization using AI for forecasting, maintenance prioritization or emissions analysis should be able to answer several operational questions:
Azure Data Manager for Energy may be relevant where organizations need an OSDU-aligned energy data platform and interoperability with specialized industry tooling. Microsoft describes the service as managed, encrypted in transit and at rest, and integrated with Microsoft Entra ID for authentication and authorization. Microsoft Learn
For organizations with facilities in Johannesburg, new ExpressRoute Metro capabilities could provide a stronger private-connectivity option for Azure workloads, but the architecture should still include local resilience and tested failover paths. Microsoft Learn
The success metric should not be “the model works.” It should be a measurable operational result: fewer unnecessary truck rolls, shorter investigation cycles, improved maintenance prioritization or reduced data-preparation time.
This is where Microsoft’s positioning around managed cloud platforms, security tooling and data governance will face its real test. The technology stack can support the journey, but successful transformation will depend on organizational discipline, skills development and local operating conditions.
The strongest potential benefits are practical:
The lasting value of the AEW 2026 Microsoft Masterclass will therefore not be determined by how persuasively it presents Azure, generative AI or cloud modernization. It will be determined by whether African energy companies leave with a sharper understanding of where AI genuinely improves operations, where human control must remain paramount, and how digital infrastructure can be built to strengthen — rather than strain — the energy systems it is meant to modernize.
From a conference session to an infrastructure story
The Microsoft Masterclass: AI & Data Adoption for African Energy Companies is more than another cloud-computing presentation on a conference agenda. It arrives as energy producers, utilities, developers and governments confront overlapping requirements: increase output, reduce downtime, integrate more variable renewable generation, protect industrial systems, and build the digital capacity needed to run increasingly automated operations.African Energy Week’s announcement frames the masterclass as a practical session for upstream oil and gas, power and renewable-energy organizations. Microsoft Corporate Vice President for Energy Darryl Willis is expected to lead the discussion alongside industry participants, with the stated ambition of showing organizations how Azure-based services can support day-to-day operational transformation. African Energy Week
That focus matters because “AI in energy” is often presented in overly broad terms. For a utility, the priority may be forecasting demand and identifying network constraints. For an upstream operator, it may be interpreting seismic data faster, detecting anomalous equipment behavior or integrating data from a scattered fleet of wells and facilities. For a renewable developer, it may be coordinating batteries, solar assets and market signals in near-real time.
These are distinct workloads, with different latency, governance and safety requirements. A useful energy AI strategy cannot begin with a chatbot or a generic automation target. It has to begin with the quality of operational data, the location of the systems being connected, and the level of control the organization is prepared to delegate to automated processes.
Why Azure adoption is becoming an energy-sector issue
Data has become a production input
The energy sector has always generated significant amounts of data. Production systems produce sensor readings. Grid operators collect telemetry. Exploration teams work with geological and geophysical data. Maintenance teams manage work orders, inspection records and parts inventories. The persistent challenge is that these records often reside in disconnected operational technology, enterprise software, contractor systems and local databases.Microsoft’s Azure Data Manager for Energy is designed around that integration problem. The platform combines Azure with the OSDU Data Platform standard and is positioned as a managed, industry-specific data foundation intended to reduce silos, support data federation and improve interoperability with first- and third-party energy applications. Microsoft Learn
For African energy companies, the significance is not merely that data can be stored in the cloud. The larger opportunity is to create a governed operational picture across exploration, drilling, production, maintenance, trading, emissions reporting and grid operations. If a company cannot reliably identify the source, ownership, quality and permitted use of a dataset, it will struggle to move from isolated dashboards to dependable AI-supported decisions.
That is why the scheduled masterclass emphasis on “how to start” adoption tracks is arguably more important than any single Azure feature. African Energy Week describes the session as covering customer examples, solution pathways and adoption tracks, rather than treating AI as a switch that can be turned on across an enterprise. African Energy Week
Cloud elasticity fits uneven energy workloads
Energy computing loads are rarely uniform. Seismic processing, reservoir simulation, imaging, portfolio modelling and large-scale forecasting can require concentrated bursts of compute capacity. Buying and operating enough dedicated hardware to cover peak demand can be costly, particularly where projects are intermittent or teams are geographically distributed.A cloud platform can provide elasticity for those workloads, but the value proposition depends on workload design. Microsoft says Azure Data Manager for Energy can autoscale for changing workload requirements and includes managed operations, security updates and support for multiple data partitions. Microsoft Learn
The opportunity is clear: a well-run cloud environment can shorten the time between raw operational data and an engineering decision. In exploration, that can mean more efficient evaluation of prospective basins. In production, it can mean moving from calendar-based maintenance toward risk-informed maintenance. In utilities, it can mean identifying issues earlier instead of relying only on post-fault investigation.
But cloud capacity is not the same as operational readiness. Models need accurate historical data, current telemetry, clear engineering ownership and a process for challenging recommendations. A predictive-maintenance model may flag a pump or compressor as high risk; it should not become an unquestioned authority over shutdown decisions in a safety-critical environment.
AI’s real value is operational, not theatrical
Predictive maintenance and equipment reliability
The strongest case for AI in energy is often neither consumer-facing nor glamorous. It is the ability to identify patterns that are difficult for humans to see consistently across thousands of signals, maintenance records and operating conditions.For upstream and midstream companies, AI-assisted reliability programs can combine vibration, temperature, pressure, flow and electrical data with maintenance history. The goal is to prioritize inspections, identify abnormal operating behavior and schedule interventions before a minor fault develops into an expensive outage.
This approach can be especially attractive where skilled maintenance teams cover large geographies or where spare parts and specialist contractors have long lead times. Still, the quality of the result rests on the quality of the inputs. Incorrect instrument calibration, incomplete maintenance records or inconsistent asset identifiers can turn an apparently advanced AI initiative into a faster way to repeat old data errors.
A responsible deployment therefore needs more than a model. It needs:
- A verified asset register and consistent naming conventions.
- Defined ownership for data correction and maintenance workflows.
- A method for engineers to validate model recommendations.
- Clear thresholds for alarms, escalation and human intervention.
- An audit trail showing which data informed a recommendation.
Grid intelligence needs local context
Power systems pose a different but equally compelling challenge. Utilities increasingly need to manage demand growth, changing generation patterns, distributed solar, battery installations and constrained transmission corridors. Connected infrastructure and advanced analytics can help operators forecast demand, identify potential congestion and improve situational awareness.However, an AI model trained on one system’s historical load patterns cannot simply be transplanted into another. Electricity demand is shaped by local tariff structures, weather, industrial activity, informal supply constraints, grid topology and outage history. A sophisticated model with poor local context can create false confidence.
The better ambition is decision support, not automatic dispatch by default. AI can help planners recognize patterns and test scenarios. It can assist control-room teams by surfacing relevant signals. But dispatch, protection settings and contingency actions must remain subject to explicit engineering governance, especially where an error could affect customers, equipment or public safety.
Virtual power plants are a coordination challenge
The event announcement also highlights virtual power plants as a means of coordinating distributed solar and battery resources. African Energy Week In principle, this model can help aggregate smaller assets so they behave more like a coordinated resource, supporting balancing, resilience and market participation.The technology opportunity is real, but virtual power plants are not simply a software project. Their effectiveness depends on communications reliability, inverter compatibility, commercial agreements, meter accuracy, regulatory permission and an operational model for dealing with assets that do not respond as expected.
For utilities and developers, that means the software layer should be evaluated alongside the physical and contractual system. AI can optimize an incomplete or poorly governed asset pool only up to a point. It cannot compensate for unreliable communications, unclear dispatch rights or inadequate cyber controls.
The digital infrastructure beneath the AI narrative
Cape Town’s data-center debate makes the stakes tangible
The relationship between AI, cloud adoption and electricity infrastructure is becoming visible in Cape Town itself. The City of Cape Town’s Municipal Planning Tribunal approved land-use applications that could enable two hyperscale data centers near Cape Town International Airport, with reported planned electrical demand of approximately 174 MVA, commonly described in coverage as roughly 174 MW. The approval remains conditional, with further site-development and building-plan processes still required. EnergiAfricaThis is relevant to African Energy Week because it demonstrates that digital transformation has a physical footprint. AI-ready infrastructure requires electricity, network connectivity, cooling, substations, backup systems and land-use planning. Data centers are not abstract “cloud” assets; they can become major industrial loads with long-lived implications for local networks.
The reported 174 MVA figure should be treated carefully. MVA and MW are not identical units, and the final real-power demand depends on the facility’s power factor and design. More importantly, the project’s reported water use remains undisclosed, and objections have focused on incomplete information around cooling, energy supply, emissions and backup generation. EnergiAfrica
That does not negate the economic case for hyperscale infrastructure. It does reinforce a basic principle: digital development must be planned as part of an energy system, not as a separate demand category that can be addressed after construction decisions are made.
Data centers can become partners in grid planning
There is also a more constructive reading. Large data centers can create anchor demand that supports investment in transmission, generation, energy storage and private-power arrangements. Where market rules allow it, flexible computing workloads may eventually become a resource that can shift at times of stress or high carbon intensity.But that potential should not be assumed. Data-center operators typically prioritize availability, and energy operators must understand the difference between a workload that can be delayed and a critical service that cannot. Any arrangement involving demand flexibility needs enforceable service levels, transparent incentives and contingency planning.
For the energy sector, the key lesson is that AI adoption and energy planning are converging. Electricity systems will increasingly power the data centers that run AI workloads, while AI platforms will increasingly help plan and operate those same electricity systems. The feedback loop is economically promising, but it makes integrated governance indispensable.
Connectivity is now part of operational resilience
In July, Cassava Technologies said its Africa Data Centres JHB1 facility in Johannesburg had been designated a Microsoft Azure ExpressRoute Metro peering location. The company described the development as expanding regional cloud-resilience capacity, with Liquid C2 offering managed connectivity and integrated cybersecurity services around the deployment. Cassava TechnologiesMicrosoft describes ExpressRoute Metro as a high-resiliency configuration that connects an organization through two distinct peering locations in the same city, creating a dual-homed design intended to improve availability compared with a connection dependent on one edge location. Microsoft’s current documentation lists both Teraco JT1 and Africa Data Centres JHB1ADC in the Johannesburg Metro configuration. Microsoft Learn
For energy companies, this is not merely a networking footnote. A cloud-based analytics platform may be of limited value if field teams, regional offices and operations centers cannot access it reliably. More critically, organizations need to distinguish between:
- Enterprise data connectivity, such as reporting, document management and planning.
- Operational data connectivity, such as telemetry, historian replication and remote monitoring.
- Control-path connectivity, where commands can directly influence physical assets.
Microsoft’s own documentation notes that standard redundant ExpressRoute links may still be vulnerable to disruptions that isolate an edge location; the Metro design addresses this with redundancy across two peering locations in a city. Microsoft Learn That distinction is valuable, but it should be part of a broader resilience architecture rather than treated as a complete answer to operational continuity.
Cybersecurity cannot be bolted on after cloud migration
Connected operations expand the attack surface
The masterclass’s emphasis on cybersecurity and data governance is well placed. African Energy Week Connecting industrial equipment, smart meters, substations, field sensors and legacy operational systems can create significant visibility benefits. It can also expose networks that were designed for isolation, longevity and predictability rather than modern internet-connected threat environments.Microsoft’s Defender for IoT platform is designed for agentless monitoring of IoT and operational technology environments. Microsoft says the service can identify devices, communications and industrial protocols, while integrating with security operations tools and supporting cloud-connected, on-premises and hybrid deployment approaches. Microsoft Learn
That architecture matters because many industrial devices cannot safely run endpoint agents. Operators may have legacy programmable logic controllers, sensors, gateways and engineering workstations that must be monitored passively rather than altered with standard IT security tooling.
Microsoft also documents integration between Defender for IoT and Microsoft Sentinel, designed to help security teams analyze, investigate and respond to operational technology incidents through more unified workflows. Microsoft Learn
Visibility is not the same as security
Security platforms are valuable, but they do not create security by themselves. A dashboard can show every device in a plant and still leave the organization exposed if it lacks trained responders, asset owners, tested incident plans and authority to act quickly.An effective Azure security program for critical energy infrastructure should include:
- Asset discovery to establish what is actually connected to the network.
- Network segmentation between corporate IT, operational technology and external access paths.
- Identity controls with least-privilege access for staff, contractors and service providers.
- Continuous monitoring that recognizes unusual device behavior without disrupting operations.
- Incident exercises that test coordination among cyber teams, engineers, executives and regulators.
- Recovery planning that includes manual fallback procedures when cloud or communications services are unavailable.
Data sovereignty and responsible AI need concrete controls
Energy data can involve national resources, critical infrastructure, commercially sensitive geological information and personal data from customers or employees. As a result, data sovereignty cannot be treated as a generic legal checkbox.Microsoft notes that Azure services generally allow customers to specify the region in which customer data is located, while some resiliency configurations may replicate information to other regions within the same geography. Microsoft Learn Organizations must therefore map their own regulatory requirements to the specific Azure services, regions, replication options, encryption settings and access controls they plan to use.
Microsoft’s sovereignty guidance also identifies Azure Policy as a way to enforce controls related to data residency and allowed services. Microsoft Learn That is useful, but implementation still requires local legal interpretation and an accurate inventory of the data being processed.
The responsible-AI dimension is equally important. An energy organization using AI for forecasting, maintenance prioritization or emissions analysis should be able to answer several operational questions:
- What data trained or informed the model?
- Who owns the model’s output and validates it?
- How are errors detected and corrected?
- Which decisions remain entirely human-controlled?
- What happens if the model, cloud service or data feed is unavailable?
- Can the organization explain a high-impact recommendation to regulators, partners and affected communities?
A practical Azure roadmap for African energy companies
The most useful outcome from the Microsoft masterclass would be a disciplined adoption roadmap rather than a catalog of services. The organizations best positioned to benefit will likely be those that move in stages and attach each stage to a measurable operational problem.Stage one: establish the data foundation
Begin by identifying the highest-value data domains: production, asset maintenance, grid telemetry, customer meters, emissions, trading or geoscience. Then define common identifiers, ownership rules, retention requirements and quality checks before attempting large-scale AI deployment.Azure Data Manager for Energy may be relevant where organizations need an OSDU-aligned energy data platform and interoperability with specialized industry tooling. Microsoft describes the service as managed, encrypted in transit and at rest, and integrated with Microsoft Entra ID for authentication and authorization. Microsoft Learn
Stage two: connect selectively and securely
Not every asset needs to be connected immediately. Start with a defined operational use case where remote data access creates clear value and the cyber exposure can be bounded. Implement secure connectivity, segmentation and monitoring before expanding the scope.For organizations with facilities in Johannesburg, new ExpressRoute Metro capabilities could provide a stronger private-connectivity option for Azure workloads, but the architecture should still include local resilience and tested failover paths. Microsoft Learn
Stage three: deploy narrow, measurable AI use cases
A pilot should have a business owner, engineering owner, baseline performance measure and explicit safety constraints. Examples might include reducing unplanned downtime for a specific class of rotating equipment, improving forecast accuracy for a constrained network area or accelerating a defined seismic interpretation workflow.The success metric should not be “the model works.” It should be a measurable operational result: fewer unnecessary truck rolls, shorter investigation cycles, improved maintenance prioritization or reduced data-preparation time.
Stage four: scale governance with the technology
Once a pilot works, scaling requires standard patterns for security, model monitoring, identity, data classification, cost management and disaster recovery. Scaling an ungoverned prototype merely expands the blast radius of its weaknesses.This is where Microsoft’s positioning around managed cloud platforms, security tooling and data governance will face its real test. The technology stack can support the journey, but successful transformation will depend on organizational discipline, skills development and local operating conditions.
The bigger opportunity — and the boundary conditions
Microsoft’s African Energy Week masterclass is timely because the continent’s energy and digital ambitions are increasingly inseparable. AI can help operators make better use of data that already exists, improve visibility across complex systems and create more efficient ways to plan, maintain and operate infrastructure.The strongest potential benefits are practical:
- Faster access to trustworthy operational information.
- Improved equipment reliability and maintenance prioritization.
- Better grid forecasting and network awareness.
- More coordinated integration of distributed energy resources.
- Stronger cybersecurity visibility across industrial environments.
- A more standardized foundation for data governance and reporting.
The lasting value of the AEW 2026 Microsoft Masterclass will therefore not be determined by how persuasively it presents Azure, generative AI or cloud modernization. It will be determined by whether African energy companies leave with a sharper understanding of where AI genuinely improves operations, where human control must remain paramount, and how digital infrastructure can be built to strengthen — rather than strain — the energy systems it is meant to modernize.
References
- Primary source: African Energy Week 2026
Published: 2026-07-27T14:37:00+00:00
Microsoft to Lead AI Masterclass at AEW 2026’s Renegade Intel Platform
Microsoft will lead an AI and Azure masterclass at AEW 2026, showcasing practical digital solutions to modernize African energy operators, grids, cybersecurity and cloud infrastructure.aecweek.com
