IT Brief US - Technology news for CIOs & IT decision-makers
Modern data center in winter finnish landscape heat recovery sustainability

Nebius reports lower costs & emissions with sustainable AI strategy

Fri, 11th Jul 2025

Nebius Group has released its 2024 Sustainability Report, detailing the company's progress and strategy in integrating sustainability as a fundamental element of its AI infrastructure operations.

The report was prepared with reference to European Sustainability Reporting Standards, outlining how Nebius has positioned sustainability at the centre of its business model. This approach is reported to deliver direct benefits, including lower operational costs, an enhanced ability to meet emerging regulatory requirements, and the capability to appeal to customers with increasing demands for sustainable technology partners.

The report sets out the environmental impact of Nebius's infrastructure, focusing on metrics such as total cost of ownership (TCO), energy efficiency, emissions intensity, responsible hardware lifecycle management, and the broader societal benefits generated by the company's operations.

Operational efficiencies

Nebius reports achieving approximately 20% lower TCO compared to conventional infrastructure through a combination of hardware resilience and energy efficiency measures. The custom-designed servers implemented by the company delivered energy savings of 10 GWh in 2024, compared to using standard hardware provided by third parties.

John Boynton, Chairman of Nebius, commented:

"Sustainability is not something we do on the side - it is integral to how we are building our business to be competitive. This is a strategic approach in which every efficiency gain we engineer into our infrastructure translates directly into better performance. When we save 10 GWh of energy through hardware optimisation or achieve a PUE of 1.1 at our data centres, we are delivering superior economics that enable long-term value creation while reducing environmental impact."

The company's flagship data centre in Mäntsälä, Finland, serves as a case study of this approach. The centre operates with a reported Power Usage Effectiveness (PUE) of 1.1, which is significantly lower than the current industry average. This is made possible through a combination of proprietary hardware, advanced thermal management, water- and refrigerant-free air cooling, and a system to reuse the heat generated by servers.

Community and emissions

A notable feature of the Mäntsälä data centre is its heat recovery system. According to the report, the system provides 65% of the heating needs for the local municipality, turning server-generated heat into a practical resource for the community. Alongside this, 94% of the centre's electricity consumption is sourced from low-carbon energy, resulting in the company reporting market-based emissions intensity of 0.04 tCO2-eq per 1 MWh, a figure the report positions as among the lowest in the technology industry.

The report also highlights Nebius's hardware recycling practices. Nearly all retired hardware was recycled during the reporting period, addressing the problem of electronic waste that is exacerbated by the rapid pace of technological change, particularly within the AI hardware ecosystem.

Sustainable technology in Southeast Asia

Nebius's report underlines the relevance of sustainable AI infrastructure for markets such as Singapore, where both digital transformation and sustainability are pressing national priorities. The company noted that the performance and efficiency gains achieved through sustainable practices can help businesses and governments reduce costs, limit environmental impact, and meet regulatory expectations.

The report's publication is positioned as part of Nebius's ongoing effort to maintain transparency and meet high standards of sustainability reporting, while also illustrating business advantages such as operational cost savings and emissions reduction.

Looking ahead

The 2024 Sustainability Report contains further discussion of Nebius's initiatives in areas such as sustainable computing and educational programmes aimed at expanding digital skills. The company states that its approach to building and managing AI infrastructure is designed to combine cybersecurity, performance, and environmental responsibility.

Follow us on:
Follow us on LinkedIn Follow us on X
Share on:
Share on LinkedIn Share on X