---
date: 2026-06-10
subject: "WH-Blackburn AI preemption trade | CAISI halted | Claude Fable released | Taiwan AI chip curbs"
---

**The White House is bargaining with Sen. Marsha Blackburn** over federal preemption of state AI laws, offering movement on the Kids Online Safety Act, NO FAKES Act and age verification rules in return. **The administration paused public operations of Commerce's AI testing unit CAISI**, citing national security concerns and clouding the office's future. **Anthropic released Claude Fable 5 to the public**, the first widely available version of its Mythos model, routing cybersecurity and biology queries to Claude Opus 4.8, while deploying Mythos 5 to approved organizations. **Taipei is weighing criminal penalties on AI chip shipments** to any buyer in China, a tightening that would mirror U.S. controls, with Gigabyte and Asustek shares dropping in early Taipei trading.

# 1. AI Policy Today

- **White House negotiating preemption of state AI laws in exchange for kids safety, deepfake bills, with Sen. Blackburn leading** — The White House is negotiating with Capitol Hill to secure federal preemption of some state AI laws in exchange for tech policies on kids safety and deepfake protections, The Hill reported. Sen. Marsha Blackburn (R-Tenn.) is leading the talks, with the package including the Senate version of the Kids Online Safety Act, the NO FAKES Act and age verification requirements, per Axios. Her spokesperson said the proposal is not "blanket preemption of all laws regulating AI or kids safety." The maneuvering indicates the Obernolte-Trahan Great American AI Act is no longer the likely vehicle for AI policy this Congress. The administration is also meeting AI companies this week on benchmarking for the June 2 voluntary pre-deployment executive order. [The Hill](https://thehill.com/policy/technology/5916062-artificial-intelligence-federal-preemption-negotiations/) [Axios](https://www.axios.com/2026/06/08/white-house-hill-relaunch-effort-block-state-ai-laws)

- **White House halts public work of federal AI testing unit at Commerce, throwing its future in doubt** — The White House has ordered the Center for AI Standards and Innovation (CAISI), the Commerce-based unit that tests frontier AI models, to stop releasing public assessments, the Wall Street Journal reported. The pause, communicated by officials including National Cyber Director Sean Cairncross, will hold while last week's AI executive order takes effect. Cairncross and Treasury Secretary Scott Bessent had lobbied to put national security at the center of model evaluation, and the halt reflects their growing influence. Companies including OpenAI have urged the administration to preserve CAISI's role. [WSJ](https://www.wsj.com/politics/policy/white-house-reins-in-ai-testing-unit-as-national-security-concerns-grow-8bd33fbb?mod=rss_Technology)

- **Anthropic releases Claude Fable 5 to general public with cybersecurity, biology guardrails routing dangerous queries to Claude Opus 4.8** — Anthropic on Tuesday released Claude Fable 5, the first publicly available version of its Mythos model, with hard safety limits that route queries about cybersecurity, biology, chemistry and distillation to Claude Opus 4.8, the Wall Street Journal and TechCrunch reported. Anthropic said an external bug bounty produced no universal jailbreaks across more than 1,000 hours of testing before release. Pro, Max, Team and seat-based Enterprise plans get Fable 5 at no extra cost through June 22, after which it will require usage credits. Anthropic plans to restore it as a standard subscription feature. With the launch Anthropic will require 30-day retention on all traffic, including for enterprises previously holding zero-retention agreements, to defend against complex jailbreaks. [TechCrunch](https://techcrunch.com/2026/06/09/anthropic-released-claude-fable-5-its-most-powerful-model-publicly-days-after-warning-ai-is-getting-too-dangerous/) [WSJ](https://www.wsj.com/tech/ai/anthropic-claude-fable-ai-model-f41fb5d7?mod=rss_Technology) [Anthropic](https://red.anthropic.com/2026/mythos-preview/?hl=en-US)

- **Taiwan considering criminal penalties on AI chip exports to all China customers, aligning with U.S. measures** — Taiwan authorities are considering stricter export controls on AI chip sales to China to align with U.S. measures and address semiconductor smuggling, the Straits Times reported, citing people familiar with the matter. The proposed controls would restrict sales to all customers in China, not just specific companies on Taiwan's existing export blacklist, and would let Taipei prosecute AI chip smuggling as a criminal violation for the first time. The move is part of ongoing U.S.-Taiwan trade talks under President Lai Ching-te's administration, per Bloomberg. Taiwanese authorities made their first known detentions of alleged AI chip smugglers in May on charges of falsifying documents. Gigabyte shares fell as much as 3.8% and Asustek 6.2% in early Taipei trading after the news. [Straits Times](https://www.straitstimes.com/asia/east-asia/taiwan-eyes-curbs-on-ai-chip-sales-to-china-to-align-with-us) [Bloomberg](https://www.bloomberg.com/news/articles/2026-06-09/taiwan-mulls-curbs-on-ai-chip-exports-to-china-to-align-with-us)

- **Brussels imposes interim measures ordering Meta to open WhatsApp to rival AI chatbots while antitrust probe continues** — European Commission regulators on Tuesday ordered Meta to restore access to WhatsApp for rival AI chatbot makers under interim measures while the bloc's antitrust investigation continues, per AP. The Commission said the action was needed to prevent harm to competition in the AI assistant market before its investigation concluded. The order remains in effect until June 2029 or until the probe closes, and noncompliance could trigger fines of up to 10% of Meta's annual revenue. Meta said it will appeal, calling the action "regulatory overreach subsidized by the many European companies that pay." [AP](https://apnews.com/article/8b4b48acb23acf5686f141a276bb868d) [FT](https://www.ft.com/content/dda5bc48-d2ef-40ce-98f7-920bb8cf65ca)


# 2. China Watch

- **Apple holds back AI Siri rollout in mainland China over Beijing approval rules** — Apple's redesigned Siri will not launch in mainland China for now because of regulatory requirements, TechNode reported, after Apple unveiled the feature at WWDC on Monday. Chinese rules require generative AI models to pass a local registration process before public release, and Apple has not announced a domestic partner cleared to deliver the assistant. The report, citing Jiwei, said Apple is also holding the feature back in the European Union, where it has objected to the Digital Markets Act. The EU delay was noted in AIPD's [June 9th edition](https://aipolicydaily.org/archive/daily/2026-06-09/); the mainland China block is a separate hurdle tied to the Cyberspace Administration of China's generative AI registration regime. [TechNode](https://technode.com/2026/06/10/apple-delays-ai-powered-siri-in-eu-and-china-over-regulatory-hurdles/)

- **CATL's $1 billion AI compute push and DeepSeek stake sharpen its AI energy strategy** — CATL joined DeepSeek's first funding round alongside Tencent, JD.com and IDG Capital, the latest piece in an AI energy strategy founder Zeng Yuqun has been assembling, Pandaily reported. CATL has deployed roughly $1 billion into AI data center infrastructure over the past year, including a stake in Zhongheng Electric's high voltage direct current business and a 38.1% stake in Chinese data center operator 21Vianet. The DeepSeek raise was previewed in AIPD's [June 4th edition](https://aipolicydaily.org/archive/daily/2026-06-04/); today's piece places that investment within CATL's broader pivot to powering and supplying AI data centers. Chinese policymakers have flagged electricity supply as the binding constraint on the country's AI buildout, giving CATL's energy and compute holdings strategic weight. [Pandaily](https://pandaily.com/catl-invests-deepseek-zeng-yuqun-ai-energy-strategy-jun2026)

- **UniSound launches efficiency focused foundation model in bid for top tier of Chinese AI** — Hong Kong-listed UniSound released U2, a general purpose large language model that reports roughly 25% lower thinking token consumption than comparable models while matching benchmark scores, Pandaily reported. The article said the company is positioning U2 as a general purpose but efficiency-first model with native support for major scaffolding frameworks and lower per-task inference costs. UniSound has built its name on speech AI since 2012; the U2 launch is its bid to enter China's top tier of general purpose providers as inference costs rise with agent workloads. [Pandaily](https://pandaily.com/unisound-u2-token-efficient-large-language-model-jun2026)

# 3. Federal Policy Tracker

- **[Congress] Senate Armed Services continues closed FY27 NDAA subcommittee markups in Emerging Threats and Cybersecurity** — The Senate Armed Services Committee on Tuesday held closed business meetings to mark up the Emerging Threats and Capabilities and the Cybersecurity portions of the National Defense Authorization Act for fiscal year 2027 in Russell 232A. The Emerging Threats and Capabilities subcommittee convened at 9:15 a.m. ET and the Cybersecurity subcommittee at 11 a.m. ET. The full committee began consideration of the FY27 NDAA the same day and continues on Wednesday, per the Federal Newswire. [Senate Armed Services Committee](https://www.armed-services.senate.gov/hearings/to-markup-the-emerging-threats-and-capabilities-portion-of-the-national-defense-authorization-act-for-fiscal-year-2027) [Federal Newswire](https://thefederalnewswire.com/senate-armed-services-committee-announces-fy-2027-ndaa-markup-schedule)

- **[Congress] Senate Judiciary scheduled June 11 markup of S.4591 voice and likeness IP bill, the Senate NO FAKES Act text covering AI deepfakes** — The Senate Judiciary Committee will hold a business meeting Thursday at 10:15 a.m. ET in Hart 216 to consider S.4591, which would protect intellectual property rights in the voice and visual likeness of individuals. The bill is the Senate text of the bipartisan NO FAKES Act targeting unauthorized AI deepfake replicas of voice and likeness, with cosponsors including Sen. Marsha Blackburn (R-Tenn.), Sen. Thom Tillis (R-N.C.), Sen. Amy Klobuchar (D-Minn.) and Sen. Chris Coons (D-Del.), per The Wrap. The same markup takes up S.1133 on federal court media coverage, S.1146 on Supreme Court televising, and judicial nominations including Sixth and Second Circuit seats and the DOJ inspector general post. [Congress.gov](https://www.congress.gov/event/119th-congress/senate-event/338550) [The Wrap](https://www.thewrap.com/media-platforms/politics/no-fakes-act-revised-congress-ai-ban/)

- **[Congress] House Science holds DOE FY27 budget hearing today with Secretary Wright, new Office of AI and Quantum included in $53.91 billion request** — The House Science, Space, and Technology Committee holds an overview hearing on the Department of Energy's fiscal year 2027 budget request today at 10 a.m. ET in Rayburn 2318, with Energy Secretary Chris Wright as the witness, per Hill Heat. The $53.91 billion DOE request includes $7.14 billion for the Office of Science and establishes a new Office of Artificial Intelligence and Quantum (AIQ). The request is a $4.81 billion increase over FY26 levels, with the increase concentrated in defense and nuclear weapons accounts while civilian energy and science accounts absorb cuts, per Legis1. [Hill Heat](https://hillheat.com/events/2026/06/10/overview-of-the-department-of-energy-s-fiscal-year-2027-budget-request) [Legis1](https://legis1.com/news/nuclear-weapons-budget-civilian-science)

# 4. Capability & Research Watch

- **NIST publishes Gödel-style mathematical proof that fixed AI guardrails cannot be universally reliable, urging continuous-monitor-and-update model** — NIST on Tuesday published a mathematical proof from senior scientist Apostol Vassilev arguing that no fixed set of AI guardrails can be universally effective against adaptive adversarial prompts. The proof extends Kurt Gödel's incompleteness theorems to AI systems and appears in IEEE Security and Privacy. NIST framed the finding as supporting a transition to a continuous-monitor-and-update security model for AI systems. [NIST](https://www.nist.gov/news-events/news/2026/06/nist-mathematical-proof-supports-transition-continuous-monitor-and-update)

- **arXiv ABC-Bench finds LLM agents beat median human experts on three dual-use biology tasks including evading DNA synthesis screening** — A new arXiv paper introducing the Agentic Bio-Capabilities Benchmark (ABC-Bench) found LLM agents outperformed the median expert human baseliner on three biology tasks: writing code to operate liquid handling robots, designing DNA fragments for in vitro assembly and evading DNA synthesis screening. In three wet-lab validation experiments, OpenAI's GPT-4o-mini-high produced scripts that successfully assembled DNA with expected sequences on an OpenTrons liquid handling robot. Agents scored highly on tasks drawing on published knowledge and well-documented protocols but more weakly on a task requiring novel bioinformatics reasoning. [arXiv](https://arxiv.org/abs/2606.11150) [OpenReview](https://openreview.net/pdf?id=mo5H9VAr6r)

# 5. Industry & Market Watch

- **Broadcom, Apollo and Blackstone launch $35 billion AI infrastructure platform, targeting 20GW total compute through 2028 starting with Anthropic** — Broadcom, Apollo and Blackstone on Tuesday formalized an AI infrastructure platform anchored by Broadcom's custom XPUs and networking, per Data Center Dynamics. The initial $35 billion tranche will finance a 1-gigawatt expansion of Anthropic AI computing capacity at Fluidstack-operated sites starting mid-2026. The platform aims to enable more than 20 gigawatts of compute through 2028 for frontier labs including Anthropic and OpenAI. Google has agreed to backstop lease payments at five Anthropic data center sites, effectively underwriting the equivalent of a $35 billion loan to Anthropic, Bloomberg reported. [Data Center Dynamics](https://www.datacenterdynamics.com/en/news/broadcom-apollo-and-blackstone-launch-20gw-xpu-platform/) [Bloomberg](https://www.bloomberg.com/news/videos/2026-06-10/google-s-backstops-underpin-35-bln-anthropic-chip-deal-video) [Broadcom](https://investors.broadcom.com/news-releases/news-release-details/broadcom-apollo-and-blackstone-establish-landmark-strategic)

- **Seattle City Council voted on one-year data center moratorium, current Amazon employees testified in support** — The Seattle City Council voted Tuesday on a one-year moratorium on new data centers, two months after several companies proposed building five large-scale facilities in the city, The Verge reported. Current Amazon employees joined other Seattle residents to testify in support of the policy last week. Data center protests have spread across the country over concerns about water consumption, local electricity prices and noise. [The Verge](https://www.theverge.com/ai-artificial-intelligence/945809/amazon-employees-seattle-data-center-moratorium) [Seattle City Council](https://council.seattle.gov/2026/06/09/city-council-passes-emergency-data-center-moratorium-and-policy-framework/)

- **Standard Bots raises $200 million Series C at $1 billion valuation for Long Island AI native robot arm production** — Standard Bots raised $200 million in a Series C round at a $1 billion valuation, led by General Catalyst and robotics fund RoboStrategy, up from $63 million raised in 2024, per The Next Web. The capital will expand Standard Bots' Long Island, N.Y. manufacturing facility and engineering hiring, per DC Velocity. The company makes AI native robot arms that learn industrial tasks through demonstration rather than coding. CEO Evan Beard said the company is on pace to do 10% of industrial robot deployments in the United States by next year. [The Next Web](https://thenextweb.com/news/standard-bots-200m-1bn-valuation-us-robots) [DC Velocity](https://www.dcvelocity.com/material-handling/robotics/standard-bots-raises-200-million-in-vc-to-expand-robot-production-in-ny)

# 6. Global & Geopolitics

- **UK Technology Secretary Kendall says under-16 social media ban will proceed despite U.S. embassy intervention** — UK Technology Secretary Liz Kendall said she was "not concerned in the slightest" by the U.S. embassy's intervention against Britain's planned under-16 social media ban, The Guardian reported. The U.S. embassy in London posted a notice warning against the ban, citing concerns about American firms. Kendall said nine in 10 respondents in a government poll supported the under-16 ban. [The Guardian](https://www.theguardian.com/uk-news/2026/jun/09/crackdown-on-tech-platforms-will-go-ahead-despite-us-intervention-says-no-10)

- **UK Medical Protection Society warns doctors and NHS face liability for AI tool errors absent law reform** — The Medical Protection Society called on UK ministers to overhaul medical negligence law to address mistakes by AI tools used in NHS diagnosis and treatment, The Guardian reported. Under the law as it stands, medics and the health service can be held liable for patients being harmed or dying even if AI made the errors. NHS clinicians use AI for scan and X-ray analysis, doctor-patient summaries and letter drafting. [The Guardian](https://www.theguardian.com/society/2026/jun/09/doctors-nhs-could-be-sued-mistakes-ai-tools-medical-protection-society-report) [Medical Protection](https://www.medicalprotection.org/uk/advice-centre-articles/widening-gulf--between-ai-and-the-law-could-leave-nhs-and-doctors-exposed-to-claims)