What are Data Storage and AI Centres?
Data Centres are the physical and commercial consequence of our demand for virtual space and AI technologies.
Worth billions, applications for data centres are proliferating across the world – and protests are flooding Governments and, ironically, Facebook (Meta). Following last month’s article explaining BESS (Battery Energy Storage Systems), here we explain what Data Centres are and – in brief – the financial and environmental issues surrounding them.
As Roberts Szczerba puts it in his post: ‘A data center is where the digital world turns physical: the buildings behind streaming, banking, cloud software and AI.’ (https://www.forbes.com/sites/robertszczerba/ )
NB a lot of statistics in the articles that follow are from the USA where the big players (Amazon, Google, Microsoft, Meta) started hyperscaling ahead of the rest of the West. China also has hyperscale AI centres funded by the state and Chinese cloud providers such as Alibaba, Tencent and Baidu, but their statistics are harder to verify.
What is a Data Storage Centre?
When you tapped your phone to make a payment, stored a photo on your Cloud facility or played Dungeons and Dragons with friends far or near, you used data. Somewhere, far, far away, that action was facilitated, logged and processed digitally on a server.
The somewhere was a Data Storage Centre. Traditional data centres store files and run standard software and are relatively efficient.
AI data centres are purpose-built to run machine learning workloads. In order for us to ask Chat gpt or Claude to help with research, wizards needed a lot of tech skill and equipment to train them. These centres generate significantly more heat per square foot and consume a lot more electricity and water than previously. (https://www.brockovichdatacenter.com/facts-myths.html)
Why Are They Building So Many Now?
Short answer: AI. Use of, plus development and training of AI technology is driving a massive increase in the need for data storage centres. It is the future and big players are scanning the globe for appropriate sites to develop lucrative economy-of-scale hyper-sized sites.
What Does a Data Storage Centre Look Like?
At its smallest, a data storage centre can be like a converted van or shipping container.
On a larger scale, a data storage centre looks like a large blacked out building, such as the centres in London and Slough, blending in with city developments.
On a hyperscale, an AI centre can occupy many acres and look like a series of windowless industrial warehouses the size of cathedrals. See our article on the world’s 10 largest Data Storage Centres for dimensions.
What types of Data Storage Centre are there?
There are four main types of data storage centre:
- Enterprise 1-10 MW
These offer internal IT storage for the single organisation owning and using it. These tend to be less energy efficient as they are older and economies of scale are leading to greater use of colocation and hyperscale facilities. In 2023, only 3% of the US server energy consumption was by enterprise data centres compared with 60% in 2014.
- Edge 1-10 MW
These are located close to its end users or data sources. Linked to the expansion of 5G wireless networks, edge facilitates use of Internet of Things devices (such as smart home tech like phone-controlled heating or gates) the data for which is processed more efficiently near its source.
- Colocation 10-100 MW
These offer shared storage space for multiple clients who lease the space, power and cooling. Economies of scale have seen a rise in use of these compared to the Enterprise option. Major players here are Equinix, Switch, COPT and Digital Realty which operate facilities across multiple continents.
- Hyperscale 100+ MW
These provide cloud storage and AI development, built on a huge scale. The big players, Amazon, Microsoft and Google, together account for approximately 59% (and rising) of all hyperscale data centre capacity across the world. A hyperscale AI training cluster can use 100–500 megawatts — enough to power a city of 75,000–375,000 homes. (statistics from https://www.brockovichdatacenter.com/facts-myths.html).
For information, there are experiments with siting hyperscale AI data centres underwater and in cold locations to mitigate their need for cooling.
An additional form of data storage centre 
is Modular Data Storage. These are mobile units, for example, a specially adapted van or shipping container, that can be easily mobilised for transport or managing an IT disaster situation, such as storm damage.
What is in a Data Storage Centre?
Because Data Storage Centres provide a controlled environment for housing critical IT resources, they must be able to operate at a sustained high capability for longer than normal facilities (‘high availability’).
Data Storage Centres require:
- Power – which must be continuous, but also a contingency back-up;
- Temperature and humidity control. Overheating can cause components such as the silicon or copper in wires or circuits to melt leading to potential fire hazards.
- Cabling
- Fire safety
- Security – both for the facility and for the data stored within it.
Data Centres: The Benefits
For good and ill, our modern lives are already inextricably bound up in digital technology. From hospital beds that can tell nurses a patient is trying to get up, to cars that drive themselves, future AI developments have apparently limitless potential to improve our lives.
They must be trained and stored somewhere.
Data Centres: The Negatives
Increasing use of AI technology requires enormous storage space. Although cloud storage may one day come to be quite literally, in space, that technology does not exist yet and economies of scale are making hyper-scaling a fabulously lucrative market for those able to invest.
The need for (and commercial opportunities of) gigantic storage facilities bring major issues:
- Scale: these dwarf their environment and applications to build in rural communities causes anger and despair.
- Power: they need huge amounts of consistent power. Local communities are finding their electricity bills are rising steeply.
- Environment (see separate articles):
- Temperature: cooling systems require huge amounts of water, compromising local demand and that of wildlife.
- Wildlife: the huge acreage and mess of construction lead to large-scale destruction of wildlife habitats.
The noise generated by finished centres is thought to impact many species.
- Construction: huge lorries and site traffic further damage frail ecosystems and ever-eroding biodiversity.
- e-waste: besides risk of contamination of land and water supplies, the amount of superannuated tech equipment is immense. In 2022, e-waste totalled 62 million metric tons. Currently, only 22% of e-waste is formally collected and recycled. It is predicted that by 2030, generative AI will annually contribute 1.2 to 5 million metric tons. This includes precious metals like copper and gold, but also toxic substances like lead and mercury.
- Devaluation of property: although there is some suggestion that other industry is attracted by the new infrastructure, thereby raising the price of properties local to a data storage centre, most people see the impact as catastrophic to their investment both financially and emotionally.
- Employment: planning applications for data storage centres claim local employment will be significantly boosted. Experience suggests that post-construction (for which many specialist labourers are brought in by experienced construction companies), centres require very few staff to run them. It is possible there may be some extra jobs if other industry grows with the new infrastructure, or training and apprenticeship commitments are made by the developer.
Conclusion
As things stand, if we want the benefits of digital and AI technology, we have to accept the need for many and massive data storage centres.
The issue:
Investors (many huge and American) are scouring the world for possible lucrative Hyperscale AI centre locations;
Governments are excited by potential tax revenue and other benefits in the form of jobs and infrastructure investment;
Environmentalists and the man on the 
Torrington Ominbus, who are facing seeing open fields turn into a secured industrial site full of ugly containers the size of cathedrals, are tearing their hair out.
Lets ask Claude for the solution.