Over the last few weeks, we have seen a flurry of optimism surrounding the adoption of Artificial Intelligence (AI) in Mauritius. From the use of AI in the workplace, to Mauritius Telecom's bold new vision of an AI corridor with GPU architecture within its data centres, and to the forthcoming National AI Strategy from the Ministry of ICT, we are seeing the right headlines and, indeed, a real desire for actual and structural digital transformation.
These are certainly exciting and optimistic times in the Mauritian digital space. However, beneath this excitement, there is a distinct and conspicuous lack of discussion regarding our energy requirements and our energy reality. Of course, I do not prejudge the content of the forthcoming National AI Strategy, but the absence of key energy actors (Ministry of Energy, the URA, MARENA, and the CEB), in the public discourse around AI and data centres raises a legitimate concern.
Back in April 2025, in an article published on LinkedIn, I noted that there is a crucial need to synergise reforms in telecoms and energy if we are to unlock the economic potential of AI for Mauritius. I argued that any sovereign AI strategy would be ineffective if the digital transformation was not mirrored by a parallel transformation in the utility sectors, particularly energy (and water, or a cooling strategy, which is also key).
Our AI vision
Keep up with the latest headlines on WhatsApp | LinkedIn
Ultimately, the success, and indeed the type, of our AI ambitions will be determined not by the digital infrastructure or software layer, but by the energy we can generate, transmit and sustain. Importantly, the transformation (or lack thereof) in the energy sector will determine the AI model and framework Mauritius is able to pursue. And this is where we must distinguish, quite clearly, between two very different visions of what it means to be an AI nation.
On the one hand, there is the AI consumer nation. This is where government and businesses deploy AI tools, automate processes, improve service delivery, and enhance efficiency and analytics. Under this model, we efficiently leverage AI and incorporate external models. For these use cases, the existing energy infrastructure, with incremental improvements, should generally be sufficient.
The second vision is AI infrastructure. This model implies local capability to host, train and process data at scale using GPU clusters and specialised architecture. It moves Mauritius from being merely a consumer of AI to becoming a host and potentially a hub.
In fact, the Government's Digital Transformation Blueprint explicitly aims to create a "secure sovereign cloud platform," proposes "PPP initiatives to set up an AI Tech Park", and anchors its national data strategy in the principle of data sovereignty. This is, by definition, an AI infrastructure model.
When coupled with Mauritius Telecom's new strategic vision which commits to deploy GPU clusters and establish an "AI & Computer Corridor" for the continent, it is quite clear that the intention is moving towards industrial hosting of AI infrastructure. And this is precisely where energy becomes the critical constraint.
Public discourse tends to brush over the energy issues with references to adding more solar panels for mobile sites or implementing generic renewable solutions for data centres. But AI compute infrastructure requires continuous, high-quality baseload power and cannot tolerate fluctuations or interruptions, not to mention the real estate footprint required if one were to rely heavily on solar farms alone.
The risk, and the need for synergy
Without doubt, the Government's Blueprint and Mauritius Telecom's new vision are forward-looking. But we must guard against misalignment especially if the energy conversation is limited to "simplified notions of renewable solutions" and "solar panels".
The risk is clear: if investment in the digital and application layers (data centres, GPUs) outpaces the investment in the power layer (generation and transmission), we will end up with stranded assets. In other words, state-of-the-art facilities that cannot operate at full capacity because the energy infrastructure cannot support them.
It is therefore imperative that our digital and energy strategies are aligned and truly synergised. To realise the vision of the Digital Transformation Blueprint andMauritius Telecom's roadmap, we will need clear structural alignment between the Ministry of ICT and the Ministry of Energy. This also demands that regulatory authorities (ICTA, URA, MARENA, etc) align and handshake on a common plan. Operationally, it requires close synchronisation between digital operators such as Mauritius Telecom and the power sector, including both the CEB and Independent Power Producers.
Pursuing digital reform without parallel energy reform is an economic contradiction. If AI is to function as a platform industry, the energy sector must be strategically reoriented to support it.
At a strategic level, this requires a shift in our energy infrastructure model and energy provisioning model. Our regulatory framework must become proactive and anticipatory and our institutions must have both the independence and the capacity to plan for large, non-incremental demand.
It is also crucial for our energy procurement model to become more transparent, competitive and cost-reflective, ensuring that future digital growth is not built on legacy inefficiencies. We must prioritise reliability and performance, aligning incentives with the technical realities of AI infrastructure. Ultimately, this comes down to clear institutional responsibilities and an energy system that can meet our renewable goals without compromising the reliability required for AI infrastructure (...).