operational oversight, scenario plan-

ning and system optimisation. They

enable more accurate forecasting of

renewable input, better balancing of

local and national supply and efcient

integration of new technologies like

heat pumps and EVs.

This work is underpinned by the

‘Digital Spine Feasibility Study’. It

entailed a 6-month feasibility study,

conducted by Arup, in partnership

with Energy Systems Catapult and

the University of Bath, to explore the

concept of a ‘digital spine’ identifying

the needs case and challenges for the

energy sector to facilitate data sharing

through a digital infrastructure.

Through a combination of stake-

holder engagement, market research,

and the consortium’s internal exper-

tise, the concept was explored

through the lens of priority energy

sector use cases, such as eibility

and vulnerable customers, to under-

stand the technical and non-technical

requirements of a data sharing infra-

structure. It dened the technical

architecture, security considerations,

governance models, and the path-

ways and delivery routes necessary

to enable a data sharing infrastruc-

ture within the energy sector. By

doing so, it highlighted the chal-

lenges for the energy sector to facili-

tate data sharing and how these

challenges could be overcome

through an enabling infrastructure.

The completed study presents the

cumulative thinking of the consor-

tium, along with the 100+ individuals

and cross-sector organisations that

were consulted in the co-creation of

what has now become the concept of

a data sharing infrastructure. Indeed,

the consortium developed a concep-

tual technical architecture which

brought the study to life by illustrat-

ing the user journey, key components,

their interactions, and how they

would support identied use cases.

Real-world impact

While AI is often spoken of in abstract

terms, its most valuable contributions

are practical. It does, and should, im-

prove project outcomes, keep time-

lines and budgets under control and

enhance the performance of built

infrastructure.

For example, in the United States,

our team has developed machine

learning models with Whole Foods

and the National Resources Defence

Council (NRDC). To do so, we used

machine learning to train models to

he energy industry is at an un-

precedented turning point: cli-

mate change, population uc-

tuations, and economic and lifestyle

shifts present challenges that tradi-

tional energy planning models were

not designed to address.

On the one hand, we’re racing

against the clock to achieve net zero

by decarbonising as quickly as pos-

sible. On the other, the energy de-

mands of data centres are booming

and we need to build resilience into

our energy supply. There is no silver

bullet for this increasingly complex

situation; however, when paired with

human innovation, articial intelli-

gence (AI) presents a vital tool to help

meet these challenges head-on.

We know from Arup’s recent survey

– ‘Embracing AI: Reshaping Today’s

Cities and Built Environment’ – AI is

already being widely used by engi-

neers, city planners, and digital of-

cers across the globe.

We have seen very high take-up of

the technology and overwhelmingly

positive attitudes. Many respondents

are already using it to enhance energy

efciency and believe it can help re-

newable energy optimisation and

decarbonisation.

With its right-time data analysis and

predictive capabilities, AI can support

in maintaining grid stability and resil-

ience amid rising demands and in-

creasing risks. In addition, it can help

to optimise costs and ultimately, drive

systemic transformation across the

sector.

Cities as energy actors

For more than a century, the role of

cities in the energy system has re-

mained largely unchanged. Electricity

is generated at thermal power stations

and then consumed by urban centres.

In fact, cities consume three-quarters

of global primary energy supplies.

However, in the face of increasingly

scarce resources, AI is being used to

rethink this model.

But now we see AI and digital solu-

tions helping to turn cities and their

residents into active nodes in the

electricity grid rather than endpoints.

For example, it is enabling local en-

ergy systems to anticipate demand by

integrating buildings’ battery energy

storage systems (BESS) and electri-

cal vehicle (EV) batteries into the

grid, controlling rooftop solar panels

and optimising heating and cooling

networks.

In this model, where cities and their

residents are active participants, grid

operators have complete transparency

over energy supply and demand to

anticipate future needs. What’s more,

surplus revenue from energy can

subsidise poorer residents or be in-

vested in community initiatives, re-

ducing wastage.

Interoperability is key

Embedding AI into our energy sys-

tems and realising the benets of this

relies on interoperability. This means

that data can be shared across plat-

forms, projects and sectors so that all

the different components of the energy

system, from solar panels to electric

vehicles and grid operators, can com-

municate with one another. By con-

trast, when AI systems operate in

isolation, and are conned to specic

projects, they create and operate in

silos of data. This hinders comprehen-

sive planning.

Interoperability is, therefore, at the

heart of an intelligent energy system.

It facilitates optimal resource alloca-

tion and adaptive infrastructure de-

velopment, by revealing where and

when resources are being used in real-

time. This can help to forecast de-

mand, optimise energy distribution

and efciently integrate renewable

sources.

Interoperability is not merely aspi-

rational; it is a current technical and

social necessity. Achieving it relies on

standardised data models and frame-

works, which protocols and govern-

ment policy must evolve to support.

This is not only to achieve adequate

data sharing, but also to manage cyber

security risks and establish trust

among market actors. Indeed, like

any technology, incorporating AI into

the operation of critical infrastructure

might introduce new cyber security

vulnerabilities. Robust measures to

protect against this are crucial.

A digital spine

Alongside interoperability of data, AI

model compatibility is essential, and

we are at the centre of developing this

new digital energy system architec-

ture. With Britain’s National Energy

System Operator (NESO), we are

building the Virtual Energy System

Programme, which is the world’s rst

ecosystem of connected “digital

twins” for a national energy system

– enabled by a common data sharing

infrastructure – a digital spine.

These digital twins, digital replicas

of the grid, are designed to improve

search for novel combinations of en-

ergy conservation strategies to

achieve deep reductions in building

energy consumption within cost and

deployment constraints.

The work showed that machine in-

telligence coupled with human re-

view and guidance could lead to up

to 10 per cent greater savings per

building within the same budget.

While this may sound small, an ad-

ditional 10 per cent savings per

building means far less new genera-

tion required on the grid to meet

surging electricity demands.

Similarly, in the UK, we have devel-

oped a data model for energy eibil-

ity markets that makes information

clearer and easier to share, helping

systems work together and support

the use of AI at scale.

AI is also already improving grid

safety and resilience. Computer vi-

sion is now being used to predict the

risk of extreme events, such as

wildres, that might damage grid

infrastructure, helping prevent trag-

edies like the Camp Fire in California

in 2018.

Concept to capability

Undoubtedly, AI has the potential to

rapidly elevate how we model and

optimise grid systems; however, this

is only if it is deployed in a way that

supports human decision-making and

addresses its risks. While data centres

powering AI are consuming large

amounts of energy, the technologies

that increase the need for these data

centres are in themselves essential for

helping us address challenges like grid

decarbonisation.

From predictive maintenance and

digital twins to AI-assisted building

management and wildre detection,

AI is already reshaping how we think

about energy. However, if we want it

to power the next phase of the energy

transition, we must embed it into the

very structure of our systems, not

just as a bolt-on, but as a backbone.

With the right infrastructure, stan-

dards and cross-sector collaboration,

we can unlock AI’s potential not just

to decarbonise, but to decentralise

and futureproof the energy grid. This

will create a smarter, more inclusive

and responsive system that benets

people, planet and economy alike

and it’s a challenge we must rise to,

collectively.

Simon Evans is Global Digital Energy

Leader, Arup.

THE ENERGY INDUSTRY TIMES - JULY/AUGUST 2025

Technology Outlook

There is no silver bullet for the increasingly complex energy landscape. But when paired with human innovation,

ariial inelligene oers a ial ool o help mee he hallenges heaon. rps Simon Evans explains.

The role of AI in empowering the

energy transition