The GCC has been experimenting with, piloting, and adopting artificial intelligence technologies for some time. Saudi Arabia has the National Strategy for AI and Data, Qatar has an Artificial Intelligence Committee, and the United Arab Emirates was the first nation to establish a ministry dedicated to AI. All this government apparatus was in place before generative AI arrived, which, in many ways, redraws the roadmaps.
AI has a powerful pitch in a region where governments and businesses are focused on health, public safety, sustainability, and economic stability. It can accelerate time to market and time to value; it can enhance efficiencies from the back office to the factory floor and from the warehouse to the field.
It can alleviate risk by making market volatility, cyber threats, and ROI more predictable. It can speed up R&D. AI is the new standard foundation of competitiveness for enterprises everywhere.
It is understandable to be inspired by AI’s seemingly endless possibilities. However, if the right upfront investments are not made in resources and technology, value can become elusive in implementation.
The GCC’s focus on sustainability and societal enhancement also means that organisations must pay due attention to the environmental impact and energy strain associated with AI, especially the large-language models found in GenAI. However, hope lies in some research that suggests AI may be an important tool in reducing greenhouse gas emissions by up to 10 per cent by 2030.
So, enterprises must design an AI journey that measurably adds the right value for their unique business model and, therefore, can justify past and future budgets. They must also invest in infrastructure that will allow them to offset carbon emissions and meet regulatory requirements.
The right foot forward
The transformative benefits and value of successful AI projects far outweigh the challenges. Most industries are still in the early stages of adoption. Still, implementation is gathering steam as new use cases are defined and we move beyond the conservative thinking that prevails within many organisations. In preparation for this shift, regional enterprises must start thinking about what is required to ensure solid foundations are in place for an AI-based future.
To ensure the success of the AI journey, here are the key issues organisations must address:
Access to GPUs
Supply chains must be assessed and factored into any AI roadmap. Access to GPUs is critically important as without them, AI projects cannot succeed. As regional AI adoption soars, the already significant demand for GPUs will affect the supply chain, and some organisations planning AI implementations may need to look to service providers for access to the technology.
Power consumption and space in data centres
Successful AI projects need massive datasets, which creates challenges for already stretched data centres, particularly in power consumption. Modern AI implementations can demand power densities of 40 to 50 kilowatts per rack — well beyond the capability of many data centres. AI is a game changer for today’s data centres’ network and power requirements.
A much higher density of fibres is required, together with greater, higher-speed networking than traditional data centre providers can deliver. Power- and space-efficient technologies will be crucial to the success of AI projects. Flash-based data storage can help mitigate this problem, as it is considerably more power- and space-efficient than HDD technology and requires less cooling and maintenance than traditional hard drives. Every watt allocated to storage reduces the number of GPUs powered in the AI cluster.
Model variance
Unlike other data-based projects that can be more selective in data storage and access, AI projects need extremely large data sets to train models and extract insights to fuel new innovation.
This presents major challenges, especially when fully understanding AI models and predicting how introducing new data may change outcomes. AI professionals are still grappling with the issue of repeatability, but a best practice to help understand data models and very large datasets is to introduce “checkpointing”.
This ensures models can be easily returned to earlier states, facilitating a better understanding of the implications of data and parameter changes. The ethical and provenance aspects of using data from the Internet in training models are also yet to be sufficiently addressed. The same goes for the possible impacts of removing selected data from an LLM or RAG (retrieval augmented generation) vector dataset.
Skills gaps
Any GCC organisation will face talent shortages on its AI journey. There is a worldwide shortage of data scientists and other AI professionals. As a result, AI-skilled people have difficulty securing and commanding premium salaries. This is likely to remain a significant issue throughout the coming decade. So, organisations will need to invest heavily in talent through recruitment and upskill or reskill their existing workforce.
The GCC shows remarkable maturity in its AI journey, including identifying use cases, recognising the need for infrastructure investment, and prioritising the skilling and upskilling of its workforce.
Progress has been made partly by identifying a fundamental truth — that the AI journey is best taken in allegiance with others. Partnerships can be the difference between success and failure for skills, infrastructure, or consultation.
Across the region, cloud service providers, managed service providers, and others can step forward and join hands with AI innovators to smooth their journeys. In collaboration, AI partners can build the region’s future—insights-driven, sustainable, and globally competitive.
