In partnership with

Discover laptops with built-in intelligence.

Shift your workforce into hyper-productivity mode with Dell Pro laptops powered by Intel® Core™ Ultra with Intel vPro®. Shop on Dell Premier.

⚖️ POLICY & REGULATION

Anthropic Pitches Lawmakers A Confidential Frontier AI Bill: A $1 Billion Or 50-Casualty Trigger, A D.C. Judge As The Only Off Switch, And Preemption Of State Disclosure Rules: Anthropic representatives have privately pitched Capitol Hill on a draft frontier AI safety bill, Politico reported Friday evening, Oct. 16, citing three people with knowledge of the effort; the company has not backed a specific congressional bill before. Under the draft, which Politico obtained, the government could block a model's release or use only for a foreseeable, imminent risk of more than $1 billion in damages or more than 50 deaths or serious injuries, and only from a biological weapon, a model evading its developer's controls or a cyberattack, language that mirrors California's SB 53. The Commerce Department would have to ask a federal judge in the District of Columbia for a suspension, and no other "department, agency, or officer" could halt development, deployment or internal use. Independent evaluators would register with Commerce, could be paid by the labs as long as payment does not depend on results, and would report an "imminent enumerated risk" within 72 hours of finishing an assessment; labs would publish safety frameworks and transparency reports for new frontier models, protect whistleblowers and report critical safety incidents to Commerce. The bill would preempt state laws on public risk disclosure, third-party evaluation and critical-incident reporting, the preemption Democrats have refused in order to protect California, Illinois and New York. "These are sensible policies we believe Congress should act on," a spokesperson said, pointing to the company's Advanced AI Framework. It lands while Thune, Cruz and Klobuchar negotiate their own frontier bill over Cantwell's objections, and the Institute for Law & AI's Charlie Bullock called it sensible but weaker than the House FRONTIER Act. Link

Anthropic Commits 5 Percent Of Its Shares To A Public Benefit Fund Barred From Lobbying, With An Escalator Tied To Valuation, Two Days After Reporting Its Junior Technical Hires Fell From 27 To 9 Percent: Announced Sunday, Oct. 18, and subject to stockholder approval, Reuters reported: the fund will be a wholly owned LLC rather than a 501(c)(3), will receive another 2 percent of shares at the board's discretion for every 50 percent rise in the company's value above an undisclosed baseline, and none of its equity may be used for lobbying or political giving in the U.S. or abroad. Named uses are large nonprofit donations, direct distributions to specific populations and work on AI's labor-market effects, building on the $200 million Economic Futures fund; with the founders' pledge to give 80 percent of their equity and an employee matching program, about 20 percent of Anthropic's shares would be dedicated to public purposes before any escalator. Governance, structure and the valuation thresholds come "in the coming months," with a public-benefit report to follow; Elon Musk said he had recommended "giving the shares directly to the public." The fund arrives two days after the company's Oct. 16 report on its own workforce: the share of junior technical hires fell from 27 percent to 9 percent since early 2025, junior postings dropped sharply after December 2025, interactive agents worked six times the hours of their human handlers by August, and the authors write that private steps alone will not address the junior slowdown and call for public policy on retraining and income support. Anthropic plans to go public as a public benefit corporation. Link

Beijing Proposes Export Controls On 23 AI Technologies, Including Models Of 2 Trillion Parameters Or More, And Bans On Models Trained On Sensitive National Data: China's Ministry of Commerce and Ministry of Science and Technology published a draft update Friday, Oct. 16, to the catalogue of technologies subject to export bans or restrictions, adding 40 items: 23 in AI, from large models to agents, 12 in solar power including polysilicon production, five in biomedicine including cell and gene therapy, and chip-design technologies for AI processors and memory, the AP reported. Core technologies for large models trained on sensitive national data, and AI agents built for defense and critical information infrastructure, would be prohibited outright; exports of models with 2 trillion or more parameters "or with certain world-leading capabilities" could be restricted, a line Moonshot's Kimi K3 and Alibaba's Qwen 3.8-Max cross, though the draft names no models. The commerce ministry called the move "necessary" for national security and said it aims at "reinforcing the security bottom line" and keeping AI "under human control." The AP account carries no consultation deadline. It is the mirror image of the June episode in which the U.S. Commerce Department asked Anthropic to bar foreign nationals from Fable 5 and Mythos 5 before lifting the restriction, and it lands four days after Meta released Muse Spark Open as the most powerful U.S.-developed open-weight model, the other side of a two-way market in model weights that both governments now treat as controlled goods. Link

STATES

Shapiro Petitions To Intervene In The PPL Data Center Complaint, The First Governor In A PUC Case In At Least 15 Years, Seeking A Separate Data Center Rate Class: Gov. Josh Shapiro petitioned Thursday, Oct. 15, to intervene in the Office of Consumer Advocate's complaint against PPL Electric Utilities at the Public Utility Commission (Docket C-2026-3065756), his office said Friday, calling it the first time in at least 15 years that a governor has entered a PUC proceeding. The consumer advocate argues PPL's rates and service classifications do not shield existing customers from the costs and outage risk of large data centers; as a party, the governor will press for a separate customer class for large data centers, allocation of the costs their demand causes to that class, stronger requirements for transmission upgrades tied to data centers, coverage of regional reliability and resource-adequacy program costs, and safeguards for existing customers if supply cannot meet demand. "Pennsylvanians should not be left paying higher electric bills because of data center developers," Shapiro said. It follows his Executive Order 2026-05 directing the Special Counsel for Energy Affordability to seek curtail-first and cost-causation rules, and the PUC's Oct. 1 tentative order doing both; PPL's base-rate settlement this year bars another base-rate case until 2028, which makes this complaint docket the place where PPL's data center class gets decided. Link

Four Data Center Decisions This Week: Littleton's First Reading And Memphis's Twice-Delayed Vote Tuesday, Champaign County's NDA And No-Abatement Measures Thursday, And Aurora's Special Meeting Tonight: Littleton, Colo., takes a first reading Tuesday, Oct. 20, of a six-month moratorium on new data center development while staff study land-use, energy, water, noise, air-quality, infrastructure and economic effects, with the public hearing and second reading Nov. 3; data centers are a conditional use in its mixed-use districts and permitted in business and industrial zones, and the ordinance's findings say their characteristics "may be incompatible with nearby residential neighborhoods," the Denver Gazette reported Friday, which would make Littleton the sixth Colorado city with a pause adopted or pending after Denver, Brighton, Fort Collins, Grand Junction and Aurora, where the question is on tonight's agenda. Memphis's council votes the same day on JB Smiley Jr.'s 12-month moratorium after two postponements, one on a 7-5 vote. Champaign County, Ill., corrects our Oct. 15 item: the County Board's Committee of the Whole advanced the no-abatement resolution for hyperscale data centers and an ordinance barring county officials and employees from nondisclosure agreements on voice votes Oct. 13, and final adoption comes at the full board Thursday, Oct. 22; the resolution binds only the county's share of taxes, so municipalities and schools would need their own. Mahomet's Village Board, which ended its Clean Cloud Energy talks Oct. 5, wraps up Tuesday after Trustee Bruce Colravy asked staff for an NDA ban and a moratorium matching the county's. Aurora, Colo., meets at 6 p.m. tonight on Councilmember Amy Wiles's proposal to stop accepting data center applications until the city adopts rules. Link

💰 POWER & CAPITAL

Microsoft Walks Away From Nscale's Loughton Data Center After Grid Officials Said Power Would Not Arrive Until The 2030s, Weeks Before An IPO That Counts Microsoft And Anthropic As 85 Percent Of Contracts: Microsoft has terminated its agreement to anchor Nscale's first British data center in Loughton, Essex, the Telegraph reported Sunday, Oct. 18, and Reuters confirmed; the Telegraph gave no reason and Microsoft did not comment. National grid officials had told Nscale not to expect power at the site until the 2030s, after UK Power Networks could not deliver the connection in time for the planned 2027 opening. Nscale said its relationship with Microsoft "remains strong and long-term," that it stays committed to Loughton and to the UK, and that it is exploring other power sources and other UK sites; "this unfortunate London-wide grid issue has no impact on our other contracts," a spokesperson said. The Nvidia-backed company filed last month for a U.S. IPO seeking up to $3 billion, with Microsoft and Anthropic expected to be its major customers, about 85 percent of contracts per the filing, and Reuters put its total contracted value above $103 billion after two and a half years; Microsoft has separately signed a letter of intent for an Nscale campus in West Virginia. A flagship site with planning permission and a signed tenant that cannot get a connection this decade is the queue problem every grid-reform promise in Britain was written to fix. Link

NiSource Signs Microsoft To A 20-Year, 1,100-MW Agreement For Its Northwest Indiana Campus Through The Generation Subsidiary Built To Keep The Cost Off NIPSCO Customers: Announced Friday, Oct. 16: Microsoft takes service in 2028 under a 20-year term, about 1,100 MW of new generation is added to serve its previously announced northwest Indiana campus, and Microsoft funds all of the generation and transmission required for its load. The agreement runs through NiSource Generation Company LLC, the GenCo structure designed so the new infrastructure is not recovered from existing Northern Indiana Public Service Company customers; the company raised its projected customer savings from its data center contracts to about $1.75 billion over their life from $1.4 billion, and it did not disclose the contract's value. Further operational and financial detail comes on the Oct. 28 third-quarter call. With Duke's large-load tariff in the Carolinas and PPL's complaint docket in Pennsylvania, the Indiana GenCo is the third template on the table for who pays for data center generation, and the only one where the customer buys the plant. Link

📜 DISPATCH PRO

The working layer under this briefing is The Docket, in Dispatch Pro

What moved in all 50 states since the last issue, which AI bills actually matter this week, and exactly what they would make you do, by when: obligations, deadlines and penalties pulled from the bill text and verified, similar bills in other states, every hearing and effective date we have reported, plus a Docket Room Pro seat ($49/mo on its own).

Today for Pro readers: 8 AI bills moved since the last issue, 1 rated Act; this week's three; The Bill on NJ A 5089 with 6 obligations, deadlines and penalties pulled from the text; 1 more AI hearing this week and 49 dated actions in the next 30 days.

$29/mo or $290/yr after the trial, less than the Docket Room seat alone. Cancel anytime. Team licenses on request.

Digitize Dispatch is read by lobbyists, government affairs, policy, and regulatory professionals tracking AI legislation, regulation, and policy developments. Put your brand in front of the people shaping how organizations respond to AI policy.
Interested in advertising with us or have any feedback? Send us an email.