Meta Funds 7 Gas Plants for Its Louisiana AI Campus; OpenAI Codex Gets Plugins

Sun, Mar 29 ~3 min read ✓ Reviewed by Get AI Decoded Editorial Team

⚠️ Not financial advice. All content is informational only. We may hold positions in securities mentioned. Always do your own research before making investment decisions. Affiliate Disclosure →

Meta bets on private gas generation for its largest AI campus while OpenAI opens Codex to third-party plugins.

💡 Meta Commits to 7 Gas Plants for Its $27 Billion Louisiana AI Campus

Decoded: Meta and Entergy Louisiana formalized a new energy agreement on March 28 under which Meta will fund the full cost of dedicated energy infrastructure for its hyperscale data center in Richland Parish, northeastern Louisiana — the largest AI data center project under active construction in the United States. Meta pays its complete cost of service, ensuring no costs shift to other utility customers. Entergy Louisiana will construct seven new natural gas-fired combined-cycle power plants totaling more than 5.2 gigawatts of capacity, alongside new transmission lines, battery storage systems, and up to 2,500 megawatts of additional solar capacity. The agreement delivers approximately $2.65 billion in total customer benefits over time. Meta also committed $120 million to Entergy's Power to Care program and $140 million toward energy efficiency initiatives. Construction runs 2026 through 2031. (Entergy Louisiana official press release, March 28, 2026)

Why it matters: Meta (META) has moved beyond leasing grid capacity — it is now funding dedicated power generation at the 5-gigawatt scale, making this campus function more like a sovereign utility than a data center. The 5.2 GW commitment exceeds many U.S. states' entire peak demand growth projections for 2026–2030. For energy investors, the model is instructive: Entergy Louisiana gets guaranteed revenue and rate base growth with no capital risk. For AI capex watchers, this sets the infrastructure floor for serious next-generation AI training facilities: 7 dedicated plants, dedicated transmission, and a multi-year build commitment before a single GPU is installed.

🤖 OpenAI Adds Plugin Support to Codex, Competing Directly with Claude Code and Gemini CLI

Decoded: OpenAI added plugin support to Codex — its agentic coding application — on March 27, allowing third-party developers to bundle workflow prompts called "skills", app integrations, and Model Context Protocol (MCP) servers into distributable packages. Plugins make specific Codex configurations replicable across developer teams inside an organization. The feature matches similar plugin and extension functionality already available in Anthropic's Claude Code and Google's Gemini CLI, both of which have supported MCP-based extensions. (Ars Technica, March 27, 2026)

Why it matters: Enterprise developer tooling is the primary commercial battleground for AI labs in 2026, and Codex's plugin gap versus Claude Code has been a documented procurement friction point for engineering teams choosing vendors. Closing this gap matters for Microsoft (MSFT), which bundles Codex through GitHub Copilot Enterprise: plugin extensibility is a key criterion in enterprise software evaluation. MCP is now the de facto plugin protocol across the three leading agentic coding tools — which reduces switching costs and puts competitive differentiation back on model quality and pricing, not ecosystem lock-in.

Stay decoded. See you tomorrow.

— The Get AI Decoded Team