The German engine maker Deutz expects its energy-unit revenue to triple within five years as data centres demand resilient electricity, revealing the physical infrastructure costs behind artificial intelligence.
Deutz expects revenue from its energy business to triple over the next five years as artificial-intelligence data centres drive demand for backup generation.
The company’s forecast adds an industrial layer to the AI investment boom. Semiconductor demand attracts most attention, but large computing facilities also require grid connections, cooling, substations and reserve power capable of operating when the network fails.
For Deutz, that creates a growth market in generators and energy systems. For Europe, it raises harder questions about how much electricity data centres will consume and whether backup systems support resilience without locking in additional fossil-fuel use.
Reliability is commercially essential
Data centres sell continuous availability. Even a brief interruption can disrupt cloud services, financial systems, communications and AI workloads.
Operators therefore install multiple layers of protection: batteries and uninterruptible power supplies cover the first moments of an outage, while generators can sustain operations for longer periods.
AI facilities increase the requirement because their power density is high and capital equipment is expensive. A site containing advanced accelerators cannot wait for a weak grid connection or accept frequent interruptions.
This makes backup power a beneficiary of AI investment even if the generators operate only during tests and emergencies.
The grid becomes the constraint
Data-centre developers increasingly compete for electricity capacity. In some regions, new connections take years because transmission and distribution networks were not built for concentrated loads of this scale.
Backup generation can protect a facility but does not solve the underlying shortage. If grids cannot supply normal demand, diesel or gas systems risk becoming more than emergency equipment.
European governments must therefore connect data-centre policy with grid planning, renewable generation and storage. Approving computing facilities without the power infrastructure merely transfers risk to local networks and consumers.
A transition challenge for Deutz
Deutz has a long history in combustion engines. The AI opportunity may support sales of conventional generator sets, but customers and regulators are under pressure to reduce emissions.
The company can respond through higher-efficiency engines, alternative fuels, hybrid systems and integration with batteries. The winning product may not be a standalone diesel generator but a managed resilience system that uses stored power first and combustion only when necessary.
That transition affects the quality of the forecast. Tripling revenue through equipment suited to future emissions rules is more durable than relying on a temporary rush for traditional generators.
AI’s industrial footprint
The Deutz outlook shows why AI is not only a software or chip story. It changes demand for turbines, engines, switchgear, cables, transformers, cooling and construction.
Those supply chains create opportunities for European manufacturers, but they also compete with electrification of transport, heating and industry for scarce grid capacity.
Policymakers should require transparent energy plans from large data centres, including expected consumption, backup arrangements, water use and flexibility during grid stress.
The AI boom may be digital at the point of use. Its infrastructure is unmistakably physical. Deutz’s revenue target is one more sign that the next phase will be measured in megawatts as much as models.
