The rise of Artificial Intelligence (AI) brings enormous benefits, but it also presents challenges—chief among them is the technology’s heavy electricity use.
In partnership with Tema ETFs and the second post in VOLTage week, this visualization highlights how AI is projected to drive growing electricity demand in the years ahead, using data from McKinsey.
AI & Electricity Use
Artificial intelligence consumes a lot of electricity for several reasons. Training a model requires processing massive datasets, which demands huge amounts of computing power. The more complex the model, the more energy it takes to run.
Even after training, energy use continues. Every time an AI answers a query—a process called inference—it requires graphics processing unit (GPU) power, adding significantly to energy demand.
Data Centers
Artificial intelligence needs data centers because the models are too large and complex to run on ordinary computers. Data centers provide the specialized GPUs, storage, and networking required for model training and inference. All of this hardware consumes large amounts of electricity to function.
On top of that, cooling systems are needed to keep servers from overheating, which adds even more power use. As AI adoption grows, so does the electricity demand tied to expanding data center capacity.
AI’s Growing Share of Data Center Capacity
In 2025, artificial intelligence is already projected to consume more data center capacity than all other workloads combined—44 gigawatts versus 38 gigawatts. That gives AI a 54% share of total capacity.
| Year | AI (gigawatts) | Non-AI (gigawatts) | AI Share |
|---|---|---|---|
| 2025P | 44 | 38 | 53.7% |
| 2026P | 62 | 40 | 60.8% |
| 2027P | 83 | 45 | 64.8% |
| 2028P | 102 | 50 | 67.1% |
| 2029P | 124 | 56 | 68.9% |
| 2030P | 156 | 64 | 70.9% |
Looking ahead, the shift becomes even more dramatic. By 2030, demand from artificial intelligence workloads is expected to reach 156 gigawatts, compared to 64 gigawatts for non-AI uses. At that point, AI would account for nearly 71% of all data center demand.
Looking Ahead
Electricity demand is set to climb steadily as AI adoption accelerates and data centers expand. This structural growth could create a powerful tailwind for companies involved in power generation, grid infrastructure, and energy technologies. For investors, the sector could offer a compelling long-term opportunity to capitalize on the digital and energy transitions shaping the future.
The Tema Electrification ETF (VOLT) invests in the companies powering the future—from energy generation to grid modernization and power management technologies. Electrify your portfolio.
https://www.visualcapitalist.com/sp/tema-01-projected-ai-driven-data-center-growth/
