The Singapore-based ST Telemedia Global Data Centres operator, on Monday, said that it has reduced carbon intensity by 70.5 per cent from 2021 levels, three years ahead of its own schedule — with renewable energy now covering 83.2 per cent of its electricity use, according to the ESG Report 2025. 

Bruno Lopez, President & Group CEO, ST Telemedia Global Data Centres, said, "In 2025, we made meaningful progress towards our carbon neutrality goal by 2030. Renewable energy accounted for 83.2% of our electricity consumption, while carbon intensity was reduced by 70.5% against our 2021 baseline. We also achieved a year-on-year reduction in absolute emissions despite continued growth in our portfolio and electricity consumption. Operational efficiency continued to improve, with an average PUE of 1.44 (a 13% improvement from our 2020 baseline) and a 41.2% improvement in WUE over the same period. These outcomes reflect our continued focus on efficient operations, responsible resource management and continuous improvement."

The timing isn't incidental. Data centres everywhere are under pressure to keep up with AI workloads that consume vastly more power than traditional computing, and STT GDC's leadership is framing that tension as the defining problem the industry now has to solve.

"The next phase of digital growth will be defined by how the industry resolves the tension between rising demand — particularly from AI — and the finite nature of energy, water and land," said Lopez. He was blunt about what that means in practice: this isn't a side commitment anymore. "Responsible scaling is therefore not a sustainability commitment alone; it is a commercial and operational imperative that shapes where we build, how we design, and how we run our data centre platform."

Scope 1 and 2 emissions dropped 15.2% year-on-year — a steep one-year cut for a company that's simultaneously trying to expand capacity. Power Usage Effectiveness, the industry's standard yardstick for how much energy actually goes toward computing versus overhead like cooling, came in at 1.44, down 13% from 2020. Lower is better in that metric, and 1.44 puts STT GDC comfortably ahead of the industry average, which typically sits closer to 1.5-1.6.

Water usage effectiveness improved by 41.2% since 2020, which the company attributes to rethinking how it balances energy and water in its cooling systems — a trade-off that's become a live issue for the industry as regulators in water-stressed regions start scrutinising data centre water permits more closely.
Nearly half the company's facilities — 48% — now carry green building certification.

Buried under the emissions headlines is a safety record worth flagging on its own: zero work-related fatalities and a Total Recordable Incident Rate of just 0.1 across more than 41 million hours worked — figures more typical of a mature manufacturing operation than a company still in aggressive construction mode.

Women make up 21.7% of the workforce, a figure that's low in absolute terms but consistent with where the broader data centre and infrastructure sector currently sits. The company also pointed to its DC Power Up programme and partnerships with 10 higher-education institutions as evidence it's trying to build a talent pipeline rather than just poach from competitors — a growing necessity in an industry where skilled technical staff are in short supply globally.

Every employee completed anti-corruption training in 2025, according to the report, and the company said it's folded climate, cybersecurity, and operational risk into a single enterprise-wide risk framework rather than treating them separately — including executive-level cyber exercises designed to stress-test how leadership responds under pressure, not just how the IT team does.

None of this happens in a vacuum. Data centre operators globally are facing louder demands — from regulators, investors, and increasingly from customers themselves — to show their numbers, not just state their intentions. STT GDC's report leans into that expectation directly, positioning early target achievement as proof of "disciplined execution" rather than a one-off win.

Whether that discipline holds as the company scales further, particularly with AI workloads continuing to climb, is the question the industry as a whole is still working out.