OpenAI Breaks With Transparency on GPT-5’s Energy Footprint

OpenAI’s decision to refrain from disclosing the energy use of its highly anticipated GPT-5 language model has thrust the issue of artificial intelligence (AI) sustainability into the global spotlight. Despite mounting calls for greater transparency, especially in the face of surging computational demands, the San Francisco-based company cited “competitive and security reasons” for withholding data on the amount of electricity needed to train and operate its latest generative model.

This move breaks with industry best practices. Competitors, including Google and Meta, have, in recent years, begun releasing details on energy usage and greenhouse gas emissions for their largest AI deployments. In contrast, OpenAI’s reluctance comes at a time when concerns about the environmental impact of AI are becoming more urgent—just as AI systems grow ever more powerful and widely adopted.

Soaring Energy Demands: What Do We Know?

Although OpenAI has kept the specifics under wraps, industry experts and academic researchers have been quick to speculate about the possible scale of the energy appetite behind GPT-5. Earlier OpenAI models, including GPT-3 and GPT-4, were estimated to have consumed thousands of megawatt-hours of energy during both training and ongoing operations. For context, research from the University of Massachusetts Amherst found that training a single large language model (LLM) could emit as much carbon as five cars over their entire lifespans.

With every successive generation—GPT-4 reportedly holding over 1 trillion parameters—energy requirements rise exponentially. While GPT-5’s size is not officially confirmed, experts suggest model scale and training data have likely increased, potentially surpassing the energy budgets of previous iterations. Google’s Gemini Ultra, another state-of-the-art AI, reportedly consumed enough electricity to power hundreds of US homes for months during training, underscoring the industry-wide trend.

Why Energy Transparency Matters

The opacity regarding GPT-5’s energy footprint has met with criticism from sustainability advocates and the broader AI research community. “AI companies have a responsibility not just to shareholders but to the planet and society at large,” noted Dr. Sasha Luccioni, an AI climate researcher. “Transparency is the first step towards accountability.”

Disclosing energy and emissions data is crucial for several reasons. First, it allows industry and government regulators to properly gauge the environmental impact of AI advancements. Second, it empowers consumers and business partners to make informed choices about adopting new technology in line with sustainability goals. Finally, it provides a factual basis for policy debate and corporate sustainability benchmarks in a sector that continues to rapidly evolve.

The AI Industry’s Climate Dilemma

The drive for increasingly capable LLMs has precipitated a surge in demand for high-density datacenters loaded with cutting-edge GPUs, most notably from NVIDIA, whose chips are now the engine rooms of the world’s largest AI models. According to the International Energy Agency (IEA), global datacenter electricity consumption could double by 2026 if current trends in AI adoption persist. In 2023 alone, datacenters were estimated to consume more than 460 terawatt-hours (TWh)—more than some G7 countries combined.

Some cloud infrastructure providers and AI firms claim steps are being taken to mitigate emissions, such as investing in renewable energy, carbon offsets, and greater efficiency. However, critics warn that such moves are insufficient against the sheer scale of energy use, and the lack of independent audits makes industry pledges difficult to verify.

Regulatory and Investor Pressures Rise

Across the globe, policymakers and ESG-minded investors are demanding stronger environmental disclosures from cloud and AI firms. The European Union’s Digital Markets Act and a suite of AI governance bills in the US Congress include provisions requiring the measurement and publication of environmental metrics for large-scale models.

“As AI takes on a larger role across industries—from healthcare and education to finance and creative work—it’s imperative we track and manage its environmental impact with scientific rigor,” said Professor Timnit Gebru, founder of the Distributed AI Research Institute. “Transparency is a bare minimum.”

Investor coalitions with trillions of dollars under management have also begun to scrutinize emissions from the AI supply chain. For public tech companies, including Microsoft (which is closely partnered with OpenAI), quarterly ESG reports now regularly highlight AI energy, water, and climate initiatives under shareholder pressure.

The Path Forward: Balancing Progress and Sustainability

While OpenAI’s GPT-5 undoubtedly represents a major milestone in generative AI, its secretive stance on energy usage has raised difficult questions about the alignment of technological progress and environmental stewardship. The AI industry finds itself at a crossroads: it must balance the unprecedented power and utility of modern AI with the urgent need to limit global emissions and move towards net-zero operations.

Several leading AI companies are voluntarily adopting science-based targets for carbon reduction, experimenting with new model architectures that require less energy, and contributing to open-source benchmarks for green AI development. However, without industry-wide transparency, independent verification, and regulatory alignment, it remains unclear if these steps will be sufficient—or if more stringent oversight will be needed as AI transformation accelerates.