Meta, Google and Microsoft lean on natural gas to power AI boom

The Blasting News piece captures a growing tension: the same hyperscalers racing to scale AGI‑capable models are quietly locking in long‑term natural‑gas supply to keep their data centers running. As AI workloads drive up electricity demand, climate‑conscious utilities and regulators are struggling to reconcile commitments to decarbonize with the need for firm, dispatchable power that current renewables can’t fully provide. The result is what the article calls a shadow grid—private gas‑backed capacity orbiting around AI hubs.([it.blastingnews.com](https://it.blastingnews.com/tecnologia/2026/04/meta-google-e-microsoft-lai-spinge-sul-gas-naturale-allarme-clima-004012059.html))

For the AGI race, energy is becoming as strategic as GPUs. If power constraints and carbon budgets tighten, they could slow the pace at which ever‑larger models are trained, or shift activity to regions willing to tolerate higher emissions. Conversely, aggressive gas build‑outs could keep the scaling curve alive at the cost of blowing through climate targets. Either way, energy policy is now de facto AI policy.

This story also has reputational and regulatory implications. Tech firms have spent a decade branding themselves as climate leaders; AI‑driven gas demand risks eroding that narrative and inviting stricter oversight of data‑center siting, energy sourcing and emissions disclosures. In the medium term, it may accelerate investment in fusion, advanced nuclear or long‑duration storage as the only politically acceptable way to support AGI‑scale compute.