As the adoption
of artificial intelligence (AI) and generative AI (GenAI) accelerates, data
centres are facing unprecedented challenges in power consumption and
sustainability. According to Gartner, the electricity demand for AI data
centres is set to surge by 160% over the next two years, potentially
constraining 40% of existing AI data centres by 2027 due to power availability
issues.
Power Consumption and Operational Constraints
The rapid
expansion of hyperscale data centres to support GenAI applications is creating
an insatiable demand for electricity. Gartner’s VP Analyst, Bob Johnson,
highlights that utility providers may struggle to expand their capacity quickly
enough to meet this demand. This could lead to significant operational
constraints, limiting the growth of new data centres and impacting the
deployment of AI technologies.
Rising Electricity Costs
With power
shortages on the horizon, electricity prices are expected to rise, increasing
the operational costs for AI data centres. Organisations must prepare for these
higher costs by securing long-term power contracts and incorporating these
expenses into their future planning. This economic pressure will inevitably be
passed on to AI and GenAI product and service providers.
Sustainability Setbacks
The push to meet
the growing power demands may also hinder sustainability efforts. In some
cases, fossil fuel plants scheduled for retirement may need to remain
operational, leading to increased CO2 emissions. This scenario poses a
significant challenge for data centre operators striving to achieve zero-carbon
goals.
Strategic Recommendations
To navigate these challenges, Gartner recommends that organisations:
- Assess the risks associated with potential power shortages on their
products and services.
- Plan for higher power costs and negotiate long-term contracts for data
centre services.
- Re-evaluate sustainability goals in light of future power requirements.
- Explore alternative approaches such as edge computing and smaller
language models to reduce power consumption.
Conclusion
The data centre industry must proactively address these power and
sustainability challenges to support the continued growth of AI and GenAI
technologies. By planning strategically and considering innovative solutions,
organisations can mitigate the risks and ensure the reliable operation of their
data centres in the years to come.