FERC Issues Fast Lane for AI Grid; Trump Walks Back Anthropic Security Tag

Two federal policy shifts in 24 hours: FERC gives AI data centers a fast lane to the power grid, and Trump signals Anthropic is no longer a national security threat.

🏛️ FERC Orders Grid Operators to Fast-Track AI Data Center Connections

Decoded: The Federal Energy Regulatory Commission on June 18 unanimously approved six orders directing major grid operators to accelerate electricity interconnection for AI data centers and other large-load electricity users. Under the orders, the six grid operators — which collectively manage the majority of U.S. electricity transmission — must submit 30-day reports on available generating capacity and defend or revise electricity rates within 60 days, according to FERC's official announcement. FERC also directed grid operators to accommodate behind-the-meter power sources for data centers and to evaluate alternative transmission technologies including solid-state transformers and superconducting lines. The commission cited wholesale electricity rates up 267% versus five years ago and data center electricity demand projected to nearly triple through 2035 as the structural drivers requiring action. The directive was initiated after Energy Secretary Chris Wright pressed FERC in October 2025 on delays threatening U.S. AI competitiveness. (TechCrunch, FERC official announcement, June 18, 2026)

Why it matters: FERC's fast-lane orders remove a regulatory bottleneck that has quietly constrained the AI infrastructure buildout. Hyperscalers including Amazon (AMZN), Microsoft (MSFT), and Google (GOOGL) have increasingly turned to expensive on-site behind-the-meter power when grid interconnections stalled — a cost that flows directly into data center operating margins. By forcing grid operators to either accelerate interconnections or publicly defend delays on a strict timeline, FERC creates accountability in a process that had drifted for years. The 30-day capacity reporting requirement is the most immediate investor signal: grid operators with constrained spare capacity will need to disclose it, clarifying which geographies can realistically absorb new hyperscaler data center construction through 2027. Investors in AI infrastructure REITs, power utilities, and hyperscaler stocks should watch the 60-day rate revision window — if electricity rates compress in interconnected regions, AI data center operating costs fall, directly improving cloud unit economics at scale.

🤖 Trump Tells Axios He No Longer Views Anthropic as a National Security Threat

Decoded: In an interview with "The Axios Show" published June 19, President Trump was asked whether he viewed Anthropic or its CEO Dario Amodei as a threat to national security. Trump's response: "Well, not now, but a week ago, maybe." Trump added that Amodei had responded cooperatively to the administration's export control directive — issued June 13 — which ordered Anthropic to disable its Fable 5 and Mythos 5 models for all foreign users, citing a claimed jailbreak vulnerability. Trump said he did not rule out using emergency powers against Anthropic but characterized the current relationship as no longer adversarial. Anthropic had publicly disputed the justification for the directive, calling it a "narrow, non-universal jailbreak" — yet cooperative engagement with the administration appears to have shifted the president's position. (Reuters, June 19, 2026)

Why it matters: Trump's Axios interview is the strongest public signal yet that the Fable 5 export control directive is on a path toward modification or reversal. The national security classification provided the legal basis for the June 13 order; if that classification no longer holds — by the president's own statement — the legal foundation for the restriction weakens significantly. For Anthropic, which is preparing for its IPO, the reversal removes the most damaging disclosed risk: zero international revenue from its most capable model tier. For Google (GOOGL), which holds a major Anthropic equity stake, a lifting of the Fable 5 restriction restores the model's commercial revenue potential before the IPO valuation is set. The episode also establishes a de facto precedent: cooperative engagement with the Trump administration — rather than legal challenge — appears to be the more effective lever for AI labs navigating government-imposed model restrictions, with implications for how OpenAI and other frontier labs manage future regulatory friction.

Stay decoded. See you tomorrow.

— The Get AI Decoded Team