The AI Brief

On June 18, FERC ordered the nation's six largest regional grid operators to justify or revise rules governing how data centers and other large power users connect to the transmission system. In a unanimous vote, the five commissioners issued tailored show-cause orders to PJM Interconnection, Midcontinent Independent System Operator, Southwest Power Pool, California Independent System Operator, ISO New England, and New York Independent System Operator. The orders affect regions serving about 200 million Americans across more than 30 states and the District of Columbia. Texas, whose ERCOT grid operates outside federal jurisdiction, is not covered.

The action directly fulfills a 2025 request submitted by Energy Secretary Chris Wright, aimed at drastically accelerating grid interconnection for large-load energy users, specifically AI data centers, amid unprecedented energy demand. Rather than issuing a Notice of Proposed Rulemaking, which typically takes years, FERC used customized show-cause orders via Section 206 of the Federal Power Act to each of the six regional grid operators. Legal experts note this approach is designed to maximize durability against state-level challenges by acting on a region-specific basis rather than a broad national standard. Grid operators and transmission owners have 60 days to respond and explain why their current rules are justified or whether they will make changes across five main categories.

As of the end of 2024, pending interconnection requests had grown to approximately 2,290 gigawatts, nearly twice the roughly 1,280 gigawatts of total installed generating capacity operating across the United States. The queue to get onto the grid had become nearly double the grid itself. Data centers now account for about 5% of US electricity demand nationally; in Virginia, home to the world's largest concentration of data center capacity, that share already exceeds 25% and is on a trajectory toward 40% by 2030. Grid Strategies president Rob Gramlich described the FERC orders as "right up to the line" of agency jurisdiction, stopping short of what Wright had proposed.

John Jumper, the computational chemist who led AlphaFold and shared the 2024 Nobel Prize in Chemistry for it, announced on June 19, 2026, that he is leaving Google DeepMind to join Anthropic. The announcement makes Jumper the most decorated individual scientist ever to change employers mid-career in the AI industry. A day before, Noam Shazeer, the co-lead of Gemini and co-author of the foundational "Attention Is All You Need" transformer paper, announced he was leaving for OpenAI. The timing makes both departures particularly pointed. Google reportedly spent $2.7 billion in 2024 to bring Shazeer back from Character.AI, and losing him to a direct competitor less than two years later carries obvious narrative weight. Jumper's departure to Anthropic, announced within 48 hours, compounds that.

Anthropic has been on a sustained hiring run. A 2025 SignalFire report found that engineers at DeepMind were nearly eleven times more likely to leave for Anthropic than the reverse. Jumper adds a different kind of weight to that list: a Nobel laureate whose most famous work sits at the intersection of AI and scientific discovery, precisely the territory Anthropic has been signaling greater ambitions in. Throughout 2026, Anthropic has been methodically constructing the infrastructure required to do serious AI-for-science work: opening wet labs, publishing research on AI agents designed for biological workflows, and announcing flagship partnerships with the Allen Institute and the Howard Hughes Medical Institute.

Jumper announced he is leaving after nearly nine years and will join Anthropic after taking a break to recharge. He has not yet revealed his exact role at the company. His arrival gives Anthropic something that no amount of language-model benchmark performance could supply: scientific legitimacy at the level of the Nobel Prize.

Disclosure: Anthropic, mentioned in this item, is the company that develops Claude, which generates this brief.

Rather than issuing a Notice of Proposed Rulemaking, which typically takes a long time to finalize, FERC issued customized show-cause orders via Section 206 of the Federal Power Act to each of the six regional grid operators. FERC's stated intent is to move to ensure that Americans have reliable, affordable power as electricity demand and technology accelerates. Legal experts note that FERC's aggressive timelines will likely reshape the AI data center landscape by significantly shortening the multi-year interconnection queue backlog; furthermore, by deploying targeted, region-specific orders restricted to FERC's jurisdiction rather than a broad national rulemaking, FERC strategically maximizes its actions' legal durability against potential challenges from state-level stakeholders.

State-level stakeholders including the National Association of Regulatory Utility Commissioners, representing state utility regulators, and the Virginia SCC had filed official comments with FERC warning the agency not to encroach upon state jurisdiction over the local electricity market. NARUC argued that federal regulation of large-load interconnections would impede states' abilities to adapt quickly to industry changes and protect the affordability of service for retail customers. The show-cause approach lets FERC act while leaving the underlying jurisdictional boundary dispute for another day.

The orders do not apply to the fast-growing data center hub in Texas, which operates a grid outside of federal jurisdiction. Senator Cynthia Lummis introduced legislation the day before FERC's orders, on June 17, to formally codify FERC's authority over large-load interconnection in statute, which would preempt the jurisdictional dispute by resolving it legislatively. Whether Congress acts before the 60-day show-cause deadline determines whether the federal-state boundary holds or expands.

Jumper, the computational chemist who led the development of AlphaFold, shared the 2024 Nobel Prize in Chemistry for it, and spent nearly nine years at Google DeepMind, announced on June 19 that he is leaving the company to join Anthropic. In 2024, Jumper and Hassabis shared half of the Nobel Prize in Chemistry for their work. The Nobel Committee credited them with cracking "the code for proteins' amazing structures" and solving a fifty-year-old problem using AI. Jumper has not yet disclosed his role at Anthropic, and neither party has commented on the specifics.

A day before, Noam Shazeer, the co-lead of Gemini and co-author of the foundational "Attention Is All You Need" paper, announced he was leaving for OpenAI. This makes the most consequential week for Google's talent base since the company first assembled that talent. Jumper had also been a key member of Google's AI coding development team, and his departure further strains the tech giant's efforts in the model race. Google has struggled to sell AI coding tools to businesses, according to former employees.

Jumper's hire signals that the AI-for-science division Anthropic has been quietly constructing is intended to produce something structurally new, not a chatbot interface on top of biology databases, but an organization capable of generating the kind of foundational scientific output that wins Nobel Prizes. Hassabis responded publicly: "What we achieved with AlphaFold changed the world, and showed the field what was possible with AI for science and medicine."

Disclosure: Anthropic, mentioned in this item, is the company that develops Claude, which generates this brief.

In the first three months of 2026, grassroots groups blocked or delayed at least 75 data center projects worth a combined $130 billion, the most ever recorded in a single quarter, according to Data Center Watch. The number of active opposition groups more than doubled between December 2025 and March 2026, reaching 833 across 49 states, organized around concerns over electricity costs, water consumption, and noise.

The scale puts a human-resistance ceiling under the FERC-driven supply-side expansion. Faster federal interconnection queues speed up the grid paperwork; they do not resolve the zoning hearings, water-rights disputes, or county moratorium votes that have become the primary physical chokepoint in markets outside the established Virginia-Dallas-Phoenix corridors. Denmark's experience offers an international warning: its state-owned grid operator Energinet paused all new large-load connection agreements in March after receiving requests totaling 60 gigawatts, in a country whose peak electricity demand is roughly 7 gigawatts. That pause is still in effect.

The $130 billion figure, if it represents even half the true pipeline impact, exceeds the annual capex of all but the two largest hyperscalers. Investors pricing AI infrastructure stocks on the premise that capital spending converts linearly into deployed capacity now face a documented, growing opposition variable that has no federal override mechanism.