D E E P D I V E
• Micro-Local AI: Small-scale AI solutions that live at the edge, keep data sovereign, and avoid GIGO pitfalls.
• Rock-Solid Data Foundations: No flashy LLM is meaningful without clean, structured inputs.
All three items are inextricably linked and major areas of investment to support the exploding requirement for data centres, not only to facilitate the growing Cloud, but more significantly the rapidly growing AI market, for which every Generative Model, LLM or other AI requirement for compute, huge amounts of rack space is required.
We foresee Rack density exceeding 200kW per rack within the next two to five years and AI will be a huge area of investment, not only to facilitate the use of AI to make tasks more efficient, but to do so safely, whilst avoiding the fallibilities that are present in many current AI platforms and solutions. Micro Local AI will also likely become a major area of investment, either developing smaller solutions from scratch, or developing a solution that will safely utilise the power of another existing platform, whilst ring fencing sensitive data.
What are the region-specific challenges you encounter in your role?
The list is full and varied as every region is different. However, it generally includes; local building regulations, local contract forms, power availability, planning consents and public perception.
In addition, data sovereignty, where any entity cannot allow their data to leave the constraints of their country, which then has a direct effect on the data centres that can be used and the redundancy protocols( ensuring that copied data is not sent out of area) is a challenge. The same applies to Data Protection Regulations( GDPR or equivalent) where the categories of data must be considered and no protected category data used in contravention of the requisite regulations.
At BCS we have experienced and navigated all of these region-specific challenges in many countries, helping many of our international clients navigate the transition from building a data centre in one country, to another.
What changes to your job role have you seen in the last year and how do you see these developing in the coming months?
The biggest change I have seen in the last year, by far, is the arrival of AI.
This has been accelerating rapidly, mostly in the form of GenAI and LLMs, as organisations race to jump on the AI train.
Whilst I support and applaud digital transformations, which AI accelerates, there are some significant fallibilities that should not be ignored in the fervent race to board this train. All users should understand that when using default interfaces, the data entered will be held by the learning model, to improve its base of data.
This can lead to unintended invalidation of confidentiality agreements, or otherwise lost control of sensitive data, or IP.
The quality of any AI output is wholly dependent upon the quality of the data that they are trained on and have access to. If data foundations are not sound, one cannot expect even the best AI systems in the world to produce accurate and meaningful outputs.
The Mantra“ Garbage In: Garbage Out”,( GIGO), succinctly captures this.
I see the use of AI developed by the leading users, to safely configure applications, or develop their own AI models, to protect against inadvertent loss of data. �
www. intelligentdatacentres. com 69