2.5 GW of Gas, 0 Change in the Climate Narrative: The Hyperscaler Disclosure Gap

Key Takeaways

• Microsoft's West Texas plant: 2.5 GW, $7 B, startup 2027.
• Ohio approvals add 1,086 MW of behind‑meter gas for data centers.
• Scope classification shifts emissions from Scope 1 to Scope 2.
• Disclosure gap lets firms claim clean energy while using fossil power.

Pulse Analysis

The AI boom is driving hyperscalers to secure dedicated fossil‑fuel power. Microsoft has entered an exclusive agreement with Chevron and Engine No. 1 for a 2.5 GW gas‑turbine plant in West Texas, a $7 billion investment slated for 2027. In Ohio, the Public Utilities Commission approved five behind‑meter gas projects totaling more than 1 GW, directly serving Meta’s New Albany campus and other data‑center sites. These facilities are being built faster than any federal inventory can track, highlighting a rapid, largely invisible shift toward gas‑heavy electricity for high‑performance computing.

Accounting standards create a paradoxical narrative. Under the GHG Protocol, emissions from owned generators fall under Scope 1, while electricity purchased from a third‑party plant is logged as Scope 2. By structuring contracts so that the gas plant is owned by a partner, hyperscalers can report the same emissions as purchased electricity, preserving a cleaner public climate story. The distinction is technical, not substantive: the same amount of carbon is burned, but it appears in a different accounting bucket, allowing firms to claim progress on renewable‑matching and net‑zero goals without disclosing the underlying fossil dependence.

For investors and policymakers, the disclosure gap poses a material risk. Stakeholders rely on Scope 1/2 reporting to assess a company’s true carbon footprint and to gauge the credibility of its sustainability commitments. When large AI workloads are powered by newly approved gas plants, the gap can mask a substantial increase in emissions, skewing valuation models and delaying regulatory action. Greater transparency—requiring detailed reporting of behind‑meter generation and its carbon intensity—will be essential to align corporate narratives with the physical reality of the power system supporting AI.