Microsoft Carbon Emissions Hit 34 Million Metric Tons as AI Buildout Accelerates
Microsoft's carbon emissions increased by 25 percent in 2025, according to the company's 2026 Environmental Sustainability Report — a sharp rise that threatens to derail the tech giant's own climate commitments and puts the environmental costs of AI infrastructure firmly in the spotlight. The report, which was first flagged by GeekWire, states that total emissions reached 34 million metric tons "without select interventions," driven primarily by the rapid expansion of its global data center footprint to support artificial intelligence workloads.
For IT decision-makers, cloud architects, and policy professionals who regularly evaluate Microsoft's cloud infrastructure as part of their digital strategy, this disclosure is more than a corporate sustainability footnote. It is a signal that the energy economics of AI — the backbone of services like Azure OpenAI, Microsoft Copilot, and the broader Azure cloud platform — are coming under increasing strain, with real consequences for compliance, procurement, and long-term infrastructure planning.

The scale of this increase is notable even by the standards of a sector known for energy-intensive operations. A 25 percent jump in a single year represents a significant reversal for a company that, as recently as 2020, had pledged to become carbon negative by 2030 — meaning it would need to remove more carbon from the atmosphere than it emits. That goal now looks considerably harder to achieve, and industry observers are watching closely to see how Microsoft responds in the years that follow.
Why Are Microsoft's Emissions Rising So Dramatically?
Microsoft's report attributes the emissions surge to two primary factors. The first and most significant is the expansion of its data center infrastructure. To meet surging demand for AI-powered services — from enterprise customers deploying Copilot integrations to developers building on Azure AI APIs — Microsoft has been on an unprecedented data center construction spree. This infrastructure, which requires enormous amounts of electricity for both computation and cooling, has a substantial carbon footprint, particularly where regional grids still rely heavily on fossil fuels.
The second factor is a policy change: in February of the previous year, Microsoft announced it would stop purchasing what it called "non-additional, unbundled renewable energy certificates" (RECs). While this might sound like a technical accounting decision, it has significant implications. Unbundled RECs allow companies to claim renewable energy usage without directly funding new renewable capacity — critics have long argued they are a form of greenwashing. By moving away from these certificates, Microsoft essentially tightened its own accounting standards, which paradoxically made its reported emissions higher in the short term, even as the company argues the move reflects a more rigorous commitment to genuine sustainability.
"The energy demands of large language models and AI inference workloads are fundamentally different from traditional cloud computing — they are vastly more power-intensive, and the grid simply hasn't caught up with demand."
— Senior cloud infrastructure analyst, commenting on the broader AI energy challengeThis combination — more infrastructure and stricter accounting — created a perfect storm for the company's emissions numbers. However, it also raises a deeper question that is particularly relevant for European IT professionals and policymakers: if the world's largest cloud providers cannot manage their carbon footprint while scaling AI, what does that mean for regulatory frameworks like the EU's Corporate Sustainability Reporting Directive (CSRD) and broader digital sovereignty strategies?
Microsoft Is Not Alone: AI's Energy Problem Is an Industry-Wide Crisis
Microsoft's predicament is emblematic of a broader challenge facing the entire cloud and AI industry. According to the International Energy Agency (IEA), data centers globally consumed around 200–250 terawatt-hours of electricity annually in recent years, and that figure is projected to grow substantially as generative AI workloads multiply. The IEA has flagged data center energy consumption as one of the fastest-growing drivers of electricity demand in advanced economies, a trend that is putting pressure on both national grids and corporate sustainability targets.
Google and Amazon — Microsoft's primary competitors in the cloud infrastructure market — have faced similar scrutiny. Google's own environmental report acknowledged that its emissions had risen significantly in recent years, also attributing much of the increase to AI infrastructure. Amazon Web Services has committed to powering its operations with 100% renewable energy, but reaching that target while simultaneously scaling AI capacity remains a significant logistical and financial challenge, as reported by Reuters and other major outlets.
For privacy professionals and IT leaders in Europe, this trend has an additional dimension. Many organisations are actively evaluating whether to migrate workloads to European cloud providers precisely because of concerns around data sovereignty, GDPR compliance, and — increasingly — environmental sustainability as part of their ESG reporting obligations. The poor emissions performance of US hyperscalers may accelerate interest in European alternatives that operate on greener, more transparent energy mixes.
| Cloud Provider | Climate Commitment | Reported Challenge |
|---|---|---|
| Microsoft Azure | Carbon negative by 2030 | Emissions up 25% in 2025 due to AI data center expansion |
| Google Cloud | Carbon-free energy 24/7 by 2030 | Emissions rising due to AI workload growth |
| Amazon Web Services | 100% renewable energy target | Scaling AI while hitting renewable targets remains challenging |
| Hetzner (EU) | High renewable energy usage in European data centers | Smaller scale limits AI offering breadth |
What This Means for IT Leaders, Procurement Teams, and Policy Professionals
For organisations operating under European regulatory frameworks, Microsoft's emissions trajectory raises several practical considerations that go beyond environmental optics. Under the EU's Corporate Sustainability Reporting Directive, larger companies — and increasingly mid-sized ones — are required to disclose Scope 3 emissions, which include indirect emissions from their supply chain and purchased services. If an organisation runs significant workloads on Azure, those emissions could eventually appear in its own sustainability disclosures.
This creates a new dimension of vendor risk for procurement and compliance teams. Choosing a cloud provider is no longer purely a question of cost, performance, and data residency — it increasingly touches on carbon accounting, ESG obligations, and reputational risk. IT decision-makers who are helping their organisations navigate CSRD, the EU Taxonomy, or voluntary sustainability commitments should be asking cloud vendors for detailed, granular emissions data — not just high-level pledges.

The EU AI Act, which sets out risk-based requirements for AI systems, does not currently mandate emissions disclosures for AI providers, but discussions within European institutions about expanding the regulatory perimeter to include environmental impact are ongoing. Some policy analysts expect future iterations of AI regulation to include energy consumption thresholds or mandatory transparency requirements for high-compute AI systems — a development that would directly affect how services like Azure OpenAI are marketed and deployed in Europe.
For small business owners and entrepreneurs who use Microsoft 365, Azure, or other Microsoft cloud services, the immediate practical impact is limited. However, for organisations that publish sustainability reports or have made public climate commitments, the emissions profile of their cloud infrastructure is increasingly scrutinised by investors, customers, and regulators alike.
Digital Sovereignty and the Case for European Cloud Alternatives
The emissions controversy arrives at a moment when European businesses and public sector organisations are already reassessing their dependence on US hyperscalers for reasons that span data sovereignty, GDPR compliance, and geopolitical risk. Microsoft's carbon emissions news adds another layer to this calculus.
European cloud providers such as OVHcloud, Hetzner, and Deutsche Telekom's Open Telekom Cloud have positioned themselves as alternatives that combine GDPR-compliant data residency with stronger environmental credentials, given that European electricity grids — particularly in France, Scandinavia, and Germany — have significant shares of nuclear, hydro, and renewable energy. While these providers cannot currently match the breadth of AI services offered by US hyperscalers, the gap is narrowing as investment in European AI infrastructure grows.
The concept of digital sovereignty — the idea that organisations and nations should control their own data, infrastructure, and technology stack — is increasingly being extended to include environmental sovereignty: the idea that the environmental cost of digital services should be transparent, accountable, and aligned with the regulatory environment in which those services operate. Microsoft's 25 percent emissions jump is a case study in why that argument is gaining traction in Brussels and in boardrooms across Europe.
Wired has previously reported on the tensions between Big Tech's climate pledges and the realities of scaling AI, noting that the very systems companies are betting their futures on are the same ones making their sustainability targets harder to reach. That tension is now playing out in Microsoft's own published data.
Microsoft's Response and the Road to 2030
Microsoft has not abandoned its 2030 carbon
Originally reported by The Verge. Summarised and curated by European Purpose.