Where AI meets energy

Europe’s AI ambitions increasingly hinge on energy supply. The International Energy Agency puts global electricity consumption by data centres at around 415 terawatt hours in 2024. By 2030, that figure could rise to roughly 945 terawatt hours. For Europe, the IEA expects an increase of more than 45 terawatt hours over the same period, or around 70 per cent. At the same time, the European Union intends to triple its data centre capacity by 2035.[1][2]

This raises a practical question: who can supply new computing capacity with energy reliably, economically and with the lowest possible emissions? Many of the necessary capabilities already exist in the market. They are, however, spread across different players. Energy producers, grid operators, project developers, energy service providers, data centre operators and investors each master part of the task.

Energy becomes a strategic factor in site selection

Data centres draw a great deal of power at a single site and place high demands on power quality and availability. According to the IEA, a typical AI data centre consumes as much electricity as 100,000 households. The largest facilities now under construction may reach twenty times that figure. A data centre with a 100-megawatt power draw, running continuously at full load, would require up to 876 gigawatt hours of electricity a year. Its local impact on the grid and on generation can therefore be considerable.[1]

The first bottleneck is usually firm grid connection capacity. It must be available in the right place and in time for the planned commissioning date. The timelines often fail to align. According to the IEA, a data centre can be built in two to three years, whereas new transmission lines in advanced economies frequently take four to eight. The IEA accordingly considers around 20 per cent of planned data centre projects to be at risk from delays in grid integration.[1]

Technical design adds a further layer. Power Usage Effectiveness, or PUE, expresses the total energy demand of a data centre in relation to the energy demand of its IT equipment. The closer the value is to 1.0, the less energy is consumed by cooling, power conversion and other ancillary systems. A good PUE lowers consumption but is no substitute for secure supply. Uninterruptible power supply, emergency power, grid redundancy and a robust operating model remain indispensable.

Storage, too, solves only part of the problem. Battery storage can absorb short-term load peaks, stabilise power quality and form part of the uninterruptible supply. At today’s costs and scales, however, it is generally not designed to bridge several days of low wind and little sun. Grid supply, renewable generation, storage, flexible loads and back-up systems must therefore be planned as a single system.

The market offers partial capabilities rather than complete solutions

B2B energy solutions providers understand industrial customers, long-term contracts and the operation of energy infrastructure. For large data centres, this alone is often insufficient. What is needed is systems expertise extending from grid connection through new generation and storage capacity to high-availability power supply. International projects add different electricity markets, grid codes, permitting procedures and financing models.

Renewable energy developers are skilled at securing land, obtaining permits, financing and building wind, solar or storage projects. Many of these companies think primarily in terms of assets and projects. Serving a large industrial customer over several years, however, demands a different way of working. The provider must advise the customer early, make technical options intelligible and continue to manage the relationship once the contract has been signed.

Energy producers can supply data centre operators through long-term power purchase agreements, or PPAs. Such contracts give generators and offtakers greater predictability of prices and revenues. They do not, however, create a grid connection, nor do they automatically deliver the required capacity in every hour. Wind and solar farms generate according to the weather. A data centre, by contrast, needs electricity around the clock. The residual load has to be covered by the grid, storage, additional generation or structured supply products. Balancing, profile and price risks must also be managed.

Data centre developers understand land, cooling, connectivity, construction and high-availability operations. Yet they often lack comparable depth in the energy system. Infrastructure investors can provide capital and build platforms. To do so, they need an operating organisation that translates customer requirements, energy architecture and site conditions into concrete projects.

None of this amounts to a blanket judgement on individual companies. Every group includes players that are already broadening their offering. The underlying pattern nonetheless persists: the market is organised by function, while the customer needs a solution that works across those functions.

Customer centricity begins before the first contract

In AI infrastructure, work with the customer often starts long before a contract is signed. At that stage, the site, connection capacity, investment structure and timetable have yet to be settled. The provider must first understand the customer’s load profile, availability requirements, planned expansion phases and climate targets. Only then can a suitable energy architecture be developed.

Customer centricity in this context means continuous accountability. The customer needs a counterpart empowered to act, from the feasibility study through to operation. Grid connection, permits, PPA, financing, construction and operating model are interdependent. Sales, engineering, regulatory affairs and project development must therefore work for the customer as one. Frequent changes in internal responsibility cost both time and knowledge.

A scalable model must combine individual sites with standardised building blocks. Grid conditions, land and permitting differ from country to country. Technical modules, review processes, contract structures and the management of the customer relationship, by contrast, can be designed to be repeatable.

Regulation increases the need for integration

Europe wants more digital infrastructure while limiting its consumption of energy and resources. The Energy Efficiency Directive and Delegated Regulation (EU) 2024/1364 establish a common reporting regime for larger data centres. Building on this, the European Commission is developing a rating scheme and examining minimum standards. The reported metrics include energy consumption, PUE, water consumption, renewable energy and the use of waste heat.[2][3]

In Germany, the Energy Efficiency Act (Energieeffizienzgesetz) sets out specific requirements. Under current law, new data centres must, as a rule, achieve a PUE of 1.2 or lower from July 2026. In addition, there are requirements on energy reuse, on the accounting-based procurement of renewable electricity and on the publication of operating data. Where waste heat is concerned, a technically usable heat flow is not sufficient in itself. It requires an offtaker, suitable temperature levels, pipelines and often heat pumps as well. The data centre, the heating network and municipal planning must therefore be aligned at an early stage.[4][5]

Regulation can create clear standards. It becomes a problem when deadlines, responsibilities and infrastructure planning fail to align. Operators and investors therefore need regulatory expertise as early as site selection and technical design. A review shortly before construction begins comes too late.

The champion becomes the orchestrator

The most promising provider need not own every asset itself. It must, however, understand the system as a whole and be able to steer it with authority. This includes access to suitable sites and grids, new generation, structured power procurement, storage, back-up, waste heat, financing and a resilient customer organisation.

Responsibility for integration therefore becomes the central question. A partner network helps only if a single player brings together the technical architecture, the commercial model and the outcome for the customer. To do this, the orchestrator needs expertise of its own and clear governance. It must be able to decide which services it delivers itself, which it buys in and which risks it assumes.

International customers compare sites across national borders. Key criteria include electricity prices, available connection capacity, carbon intensity, permitting times, connectivity and scope for future expansion. A market leader must understand local rules while delivering a consistent offering across several countries. That calls for common technical standards and strong local teams.

Leadership decides whether integration succeeds

Business logic cannot be integrated through contracts, partnerships or acquisitions alone. Leadership determines whether individual capabilities become a coherent offering. What is needed are executives who can weigh technical, commercial and regulatory questions at the same time. They need not possess the greatest depth in every discipline. They must, however, ask the right questions, recognise conflicting objectives and take decisions across functional boundaries.

Systems competence is particularly important. An executive must understand how grid connection, load profile, PPA, storage, redundancy, waste heat and financing interact. Otherwise each unit optimises its own part while the project as a whole loses time or economic viability. Systems competence is reflected less in detailed technical knowledge than in the ability to identify dependencies early and to assign clear responsibilities.

Genuine customer leadership is equally important. Large data centre projects involve long lead times and many decision-makers. The commercial lead must build access on the customer side, sharpen requirements and secure robust commitments internally. Selling a contract is not enough. He or she must be able to guide the customer through unresolved technical and regulatory questions.

A third requirement is the ability to lead an ecosystem. Grid operators, generators, technology providers, public authorities, municipalities, investors and heat customers pursue different goals on different timelines. The accountable executive must shape them into a workable project structure. That demands clear decision rights, shared milestones, reliable escalation paths and candour about risk.

Then there is international scalability. The central organisation must set standards without crowding out local market knowledge. Executives therefore need experience with decentralised teams, differing regulatory regimes and cross-border customers. They must be able to distinguish between what is prescribed centrally and what has to be decided locally.

A single executive will not be enough

The breadth of capability required is very large for one individual. Companies should therefore resist searching for a supposed all-rounder. A complementary leadership team is the more robust answer. One person can assume overall responsibility, including the P&L. Other key roles lead engineering and operations, customers and business development, and grids, regulation and partnerships.

The decisive factor is how these roles are connected. Shared goals, unified project steering and clear decisions count for more than a perfect organisation chart. The selection criteria must reflect this. An impressive CV from a well-known industry is not enough. What counts is demonstrable evidence. Has the person built new business models? Have they managed complex B2B client relationships over several years? Have they aligned technical and commercial teams behind a single outcome? Have they scaled a business across several countries under differing rules?

The IEA points out that AI-related skills are considerably less widespread in the energy sector than in other industries. Training remains important but will not close the gap on its own. Companies must also look for candidates in adjacent markets. Profiles from digital infrastructure, complex energy solutions, industrial platform businesses or other regulated infrastructure markets may be of interest.[1]

Capability questions must be settled before the search begins

For companies, the work on people therefore begins before any single position is filled. First, it must be clear which capabilities are indispensable to the business model. It then becomes possible to decide which of these must sit within the leadership team, which can be developed and where partners will suffice. This is the foundation on which meaningful roles, decision rights and search profiles can be built.

An executive search partner can support this work but should not stand in for business strategy. Its contribution lies in translating the target picture into a concrete capability architecture, defining realistic search fields and assessing candidates on more than their industry background. In a market without a clear champion, a narrow focus on direct competitors is of little help.

At Odgers, an engagement can therefore begin with a structured stocktake. Which capabilities already exist within the team? Where are the critical gaps? Does the company need a new overall leader, a technical and commercial leadership tandem, or targeted reinforcement at the next level down? These questions yield the relevant target markets and robust selection criteria.

The search should then deliberately reach across industry boundaries. The assessment considers not only what candidates have been responsible for, but how they have managed interfaces, developed customers, weighed risks and built new structures. In this way, executive search becomes part of the organisation’s preparation for the new business, without taking over the company’s responsibility for strategy and execution.

Market leadership rests on clear accountability

Whether the future champion already exists in the market remains an open question. More likely, it will emerge through the deliberate development and combination of capabilities that are still separate today. Capital, generation portfolios and data centre capacity will remain important. What will make the difference, however, is the ability to lead the system as a whole and to deliver reliably for the customer.

Leadership questions are thus part of market strategy from the outset. Companies must clarify who bears responsibility for integration, which capabilities the leadership team needs to cover and how international execution is to succeed. Those who answer these questions early improve their chances of turning individual projects into a scalable business.

Sources

[1] International Energy Agency, Energy and AI, 2025
[2] European Commission, Energy performance of data centres
[3] European Commission, Minimum performance standards for EU data centres
[4] Federal Ministry of Justice and Federal Office of Justice, Energy Efficiency Act (EnEfG), Section 11
[5] Federal Ministry of Justice and Federal Office of Justice, Energy Efficiency Act (EnEfG), Section 13

 ___________________________________________________________

Get in touch to find out more or to receive advice from our experts: Go to the contact form.