Why medium-sized, high-growth businesses are going to have to ‘go co-lo’ in an AI era…

Growing pains

Medium-sized, high-growth businesses in the UK are currently caught in a set of crosswinds. In the year to March 2023, the sector outperformed both the FTSE 350 and smaller businesses, boosting revenue by 12% to £1.5tn.

Yet, this growth faced great friction. More than half (56%) of UK mid-sized businesses say record levels of inflation or increased overheads across energy bills, rent and payroll are the biggest challenges they face. In addition, access to skills and talent is a barrier to expansion for approximately 24% of these businesses.

The unwelcome news is that technology is set to join this list of challenges – particularly the fight to secure technological resources needed to drive growth. Artificial Intelligence (AI) is top of this pile. The Generative AI market alone is expected to report annual growth in excess of 40%, built on technical innovation set to outperform Moore’s Law.

A reality check?

It is important to remember that despite such immense potential, AI is not a self-sustaining entity, and it does not happen in a vacuum. AI capabilities depend on infrastructure, which is both finite and in short supply.

As AI demands and use cases come to dominate infrastructure requirements, the impact of AI on data centres and connectivity, cannot be underestimated.

The first consideration is that AI workloads can substantially alter how a data centre operates.  AI applications are hungry. The increased compute power demands specialised, advanced hardware.

This hardware demands more power. A ‘traditional’ data centre is designed around five to ten kilowatts per rack as an average density. AI can require more than sixty kilowatts per rack. This increased density needs more cooling to prevent overheating and disruption.

This leads to the critical issue of power, specifically the need to find and secure sustainable power. The International Energy Agency (IEA) expects global electricity demand for data centres, driven by AI growth, to double by 2026.

And it is not just an issue of the power going into the data centre.  AI applications create far more data than other types of workloads.  As a result, they need more capacity.

Thus, as the market demands more AI capabilities, the data centre must not only adapt to the hardware needed for that, but also the profound changes to accommodate that hardware. AI has the capacity to push the data centre to its limits.

A mid-market focus

So, what has all this to do with medium sized businesses?  Firstly, AI presents a unique opportunity to this sector. The World Economic Forum has recently noted that AI: “can be especially beneficial for mid-market companies.” It also noted that, “A relatively small, but well-thought-out, investment in AI can help mid-market companies rapidly develop and scale.”

Secondly, any mid-market business not already looking into AI is at a huge disadvantage: In a 2023 Technology Trends Report from Deloitte, a survey of five hundred private, mid-market businesses found that nearly all respondents were on the AI adoption curve. Forty-one per cent of respondents were ‘exploring’ AI. Thirty per cent claimed they are piloting solutions, and one-quarter said they have active AI applications.

This mixture of opportunity and commercial and technical challenges is hitting medium-sized businesses at a time when they need to focus on core activity.

Co-Lo is critical

It is this imperative that has led UK businesses to seek colocation partners that can deliver data centres and infrastructure that is sustainably powered.  At the same time, this infrastructure must remain operationally capable and available long-term.

However, there is one final consideration: the combination of speed and agility that powers mid-market innovation. These businesses need to be able to exploit AI (and other technologies such as analytics) in real-time, with competitive, transparent costs.

As a result, this substantial compute resource needs to remain close to end users – especially as the demand for more data increases. Proximity means that a regional presence, connected where necessary to other resources, is the only viable topology for the mid-market.

As far back as 2021, more than three-quarters of mid-size companies were looking to invest in AI. The difference today is that not only is AI almost mandatory amidst intense competition, but also that the infrastructure, partners, and capability are available to enable these companies to maximise that investment.

Pulsant is the leading UK data centre and edge infrastructure provider. Its portfolio of 12 edge data centres, interconnected through a low latency network fabric, delivers access to cloud, connectivity and compute services across the UK and beyond.