AI Policy · Daily

Trump confirmed his team is weighing equity arrangements with frontier AI labs, with Sam Altman pitching an OpenAI version modeled on the Alaska Permanent Fund and meeting Sen. Bernie Sanders the same week. New York lawmakers passed a one-year moratorium on large data centers, with companion bills on surveillance pricing and digital stalking now pending Gov. Kathy Hochul's signature. Senate Democrats began rolling out a coordinated AI legislative package this week, led by a Coons-Reed bill requiring human control over autonomous weapons and barring AI from nuclear launch decisions. Labor unions, consumer groups, and House Democrats opposed the Great American AI Act, the Obernolte-Trahan draft that would preempt new state AI laws for three years for developers above $500 million in revenue.

I.Top Stories

Trump confirms federal equity stake talks with frontier AI labs; Altman pitched OpenAI version, met Sanders the same week

President Trump told reporters Friday aboard Air Force One that his team is examining the idea of AI companies giving Americans an equity stake in their firms, saying he will meet AI executives at the White House "probably next week." CNBC reported the Trump administration is in active discussions with OpenAI about an equity arrangement, with the company's "Public Wealth Fund" concept among the structures under consideration. A person familiar with the talks told Semafor that Altman first floated a hybrid of the Alaska Permanent Fund and Trump Accounts when he pitched the idea to Trump in 2025. Altman also met privately with Sen. Bernie Sanders for nearly an hour in Sanders' Senate office at Altman's request, telling Sanders he supports public equity but cannot back the senator's 50% threshold, people familiar with the conversation told the AP.

Read at CNBC ↗ • Read at AP ↗ • Read at Semafor ↗

New York legislature passes one-year moratorium on large data centers, awaiting Hochul signature

The New York State Senate and Assembly passed a one-year moratorium on new large data centers along with two related bills on surveillance pricing and digital stalking, The Verge reported. The bill applies to facilities above a peak demand threshold and directs the state environmental agency to produce an impact report covering electricity, water, land use and pollution. Gov. Kathy Hochul has not said whether she will sign. The moratorium is the first statewide ban of its kind.

Read at The Verge ↗

Senate Democrats roll out coordinated AI bill package; Coons-Reed defense AI bill introduced Monday

Sen. Adam Schiff and other Senate Democrats are unveiling a coordinated package of AI bills this week framed by the Wall Street Journal as a "challenge to tech giants," including legislation expanding Pentagon AI oversight. Sens. Chris Coons (D-Del.) and Jack Reed (D-R.I.) introduce the Responsible Artificial Intelligence in Defense Act Monday, Axios reported, citing a memo first shared with the outlet. The bill would direct the Pentagon to require humans in the loop and a manual override capability for AI enabled autonomous weapon systems until reliability thresholds are met, prohibit military AI use for mass surveillance of persons inside the U.S., and bar AI from making the decision to launch a nuclear weapon.

Read at WSJ ↗ • Read at Axios ↗

Labor unions, consumer groups and House Democrats publicly oppose Great American AI Act preemption

Resistance to the 269-page bipartisan Great American AI Act is consolidating into a public opposition coalition, with labor unions, consumer protection groups and House Democrats now publicly aligned, TechTimes reported, citing Roll Call coverage. The draft from Reps. Jay Obernolte (R-Calif.) and Lori Trahan (D-Mass.) would preempt new state AI laws for three years and apply to frontier developers with more than $500 million in annual revenue.

Read at TechTimes ↗ • Read at Rep. Trahan ↗

OpenAI publicly commits to comply with Trump pre-release model review executive order

George Osborne, Head of Countries for OpenAI, told CNBC the company will comply with Trump's executive order requiring 30-day government access to frontier models before public release. "It's quite right that democratic governments have a big role to play in how this technology is used and deployed," Osborne said. Trump signed the order June 2. OpenAI is publicly accepting the voluntary regime even as its June 3 policy paper continues to ask Congress to mandate a tougher civilian led review.

Read at CNBC ↗ • Read at Engadget ↗

II.China Watch

China's National Data Administration sets a 2028 target for industry AI datasets

China's National Data Administration released a draft action plan calling for a batch of industry high quality datasets covering key sectors and validated through application by the end of 2028, Jiemian reported. The plan asks regulators to develop data driven AI applications, build specialized data companies and talent, and codify standards for dataset construction. China has already built more than 100,000 high quality datasets nationwide by the end of 2025, and more than 30 new standards covering public data, data infrastructure and AI agents are expected to be issued this year. The administration was set up in 2023 to serve as Beijing's central data regulator. The directive puts the National Data Administration at the center of China's data and AI governance stack, formalizing a state led pipeline for training data supply that has no direct U.S. counterpart.

Read at Jiemian ↗

Alibaba folds AI teams into new Token Foundry unit reporting to CEO Wu

Alibaba upgraded its large language model organizational structure on Monday, merging the Tongyi large model team with the Future Life Lab to form a new Token Foundry business unit reporting directly to CEO Wu Yongming, Jiemian reported. The reorganization names Zhou Jingren as chief scientist and stands up an AI Future Research Institute. The move builds on Alibaba's March 2026 creation of the Alibaba Token Hub group, also placed under Wu's direct control, treating token output and distribution as the company's core commercial AI layer. Alibaba's Qwen models sit inside this restructured stack.

Read at Jiemian ↗

Huawei chips clear a model training milestone widely thought to require Nvidia hardware

A consortium led by Shenzhen Hetao College, with Harbin Institute of Technology Shenzhen, the Shenzhen Big Data Research Institute and Huawei, completed full-parameter post-training of the 1.6 trillion-parameter DeepSeek-V4-Pro model on a Huawei Ascend 910C cluster, Pandaily reported. Until this run, domestic accelerators were viewed inside the industry as suitable for inference but not for end-to-end training at frontier scale. The Mixture-of-Experts training workload requires every expert module to update in parallel, producing all-to-all communication volumes dozens of times higher than standard dense models. The achievement narrows the strategic premise that Bureau of Industry and Security controls on Nvidia's H100 and H200 chips cleanly bottleneck Chinese frontier model development, a premise underpinning current export license design.

Read at Pandaily ↗

Chinese AI models outpace U.S. counterparts in global API traffic for a sixth straight week

Chinese AI models have led American models in total weekly API call volume on OpenRouter for six consecutive weeks, with Chinese models drawing 14.19 trillion tokens against 3.2 trillion for U.S. models in the most recent week tracked, Pandaily reported. DeepSeek-V4-Flash held the top spot and Tencent's Hy3 preview took second, while Shanghai-based MiniMax M3 broke into the global top three for the first time and Xiaomi's MiMo-V2.5 entered the rankings at fourth. The report credited aggressive pricing and open-weight distribution for the developer adoption gap. OpenRouter is one of the largest AI model routing platforms, providing a single, comparable usage signal across vendors. The sustained lead sharpens the U.S. policy debate over whether export controls and any forthcoming federal rules on open-weight models should target inference reach and downstream deployment, not chips alone.

Read at Pandaily ↗

Moonshot AI's valuation jumps sevenfold to $30 billion in six months

Moonshot AI, the developer of the Kimi assistant, opened a new fundraising round at a pre-money valuation of $30 billion, up from about $4.3 billion at the end of December 2025, TechNode reported, citing National Business Daily. The round follows a previous financing that closed at roughly $20 billion post-money in early May. Meituan Longzhu partner Wang Xinyu said Kimi's annual recurring revenue surpassed $200 million in April after the release of the K2.5 model. The deal lands Moonshot in a four-company front rank of Chinese model developers alongside DeepSeek, Zhipu and MiniMax. The compressed funding cadence shows private capital pricing in continued growth at Chinese model labs even as the Commerce Department tightens enforcement of Nvidia and AMD advanced chip license requirements on offshore affiliates of Chinese firms.

Read at TechNode ↗

JD.com and Tencent tie up to push AI shopping agents into Chinese smartphones

JD.com and Tencent have agreed to collaborate on AI agents, pairing JD's e-commerce and fulfillment systems with Tencent's WeChat user network, TechNode reported, citing IThome. JD's agent has been wired into native AI assistants on Huawei, Oppo and Honor devices through an Agent-to-Agent protocol, letting users issue shopping requests that JD's logistics and customer service network then fulfills. Tencent is separately working with Xiaomi and Vivo on similar A2A integrations for WeChat. The setup gives Tencent's WeChat a gatekeeper role over how third-party AI agents call into handset level assistants. Concentrating agent traffic through WeChat and JD's commerce stack hands the Cyberspace Administration of China, already the licensing authority for generative AI services, a narrower regulatory chokepoint as agent-to-agent activity scales inside Chinese consumer apps.

Read at TechNode ↗

III.Federal Policy Tracker

Sriram Krishnan to leave White House AI adviser role at end of June

White House senior AI policy adviser Sriram Krishnan confirmed Saturday on X that he will step down at the end of June 2026, TechCrunch reported. Krishnan has been the principal architect of Trump's AI policy framework since the second term began, including the strategic AI plans and the model review executive order. He is starting a new institution to continue shaping Trump AI policy from outside.

Read at TechCrunch ↗

Space Force awards SpaceX $4.16 billion for satellites tied to AI airborne tracking architecture

A Wall Street Journal investigation mapped SpaceX's growing Pentagon embeddedness, with the company holding billions of dollars in active Department of Defense contracts and continuing to land new awards. The Space Force announced May 29 that SpaceX won a $4.16 billion contract under the Space-Based Airborne Moving Target Indicator program for a satellite constellation that can track and target airborne threats globally, Space.com reported. SpaceX is the first of nine companies disclosed for the SB-AMTI vendor pool, with the constellation slated to be operational by 2028 and feed the Golden Dome missile defense effort. The award lands as the fiscal 2027 NDAA moves to House floor consideration with vendor concentration risk an active question.

Read at WSJ ↗ • Read at Space.com ↗

Whistleblower says DOGE plan would have falsely marked 2.7 million people as dead, including U.S. citizens

Former Social Security Administration IT modernization head Jeremiah Schofield filed a 49-page whistleblower disclosure to Sens. Elizabeth Warren (D-Mass.) and Richard Blumenthal (D-Conn.) alleging that DOGE and DHS officials proposed adding 2.7 million living people, including U.S. citizens and lawful permanent residents, to the Social Security Death Master File as an immigration enforcement tool, The Washington Post reported. Schofield said a smaller version was carried out last year, adding 6,100 immigrants to the file. SSA said in a statement it "did not add a list of 2.7 million names" and that internal controls were maintained. Warren and Blumenthal sent letters Thursday seeking answers from SSA and from named former DOGE officials Antonio Gracias, Jon Koval and Payton Rehling.

Read at The Hill ↗ • Read at Washington Post ↗

Meta confirms 20,225 Instagram accounts hijacked through flaw in its AI customer support tool

Meta confirmed that attackers exploited its AI powered High Touch Support system to hijack 20,225 Instagram accounts, BleepingComputer reported. The High Touch Support tool failed to verify whether email addresses were associated with the target accounts, allowing attackers to reset victim passwords by submitting any email. The breach was logged in a Maine Attorney General data breach notice.

Read at BleepingComputer ↗ • Read at Maine AG ↗

War on the Rocks analysis says Pentagon AI edge is eroding through model distillation, not breach

Sebastian Elbaum wrote in the War on the Rocks "Cogs of War" series that adversaries do not need to penetrate Pentagon systems to acquire frontier U.S. AI capability because they can distill capable smaller models from public outputs of U.S. frontier systems. The piece frames the question under U.S.-China strategic competition and recommends rethinking how Pentagon procurement and Bureau of Industry and Security export controls account for distillation pathways. Elbaum cites Stanford AI Index data showing the gap between top U.S. and Chinese models on the Arena leaderboard has narrowed to 2.7%, down from 17% in 2023.

Read at War on the Rocks ↗

IV.Industry & Market Watch

Microsoft unveils seven in-house MAI models at Build 2026 to cut OpenAI dependence

Microsoft launched a family of seven in-house AI models at its Build developer conference, including MAI-Thinking-1, its first reasoning model trained from scratch on commercially licensed data without third-party distillation, and MAI-Code-1-Flash, a coding model rolling into GitHub Copilot and Visual Studio Code, Forbes reported. Microsoft AI chief Mustafa Suleyman said MAI-Thinking-1 was preferred over Anthropic's Claude Sonnet 4.6 in blind evaluations by independent rater Surge and matches Claude Opus 4.6 on coding benchmarks, per Euronews. Microsoft also announced its Majorana 2 quantum chip and a new Cobalt 200 silicon line.

Read at Forbes ↗ • Read at Microsoft Blog ↗

Chip selloff erases more than $1 trillion in market value as Alphabet upsizes AI capex raise to $84.75 billion

The PHLX semiconductor index fell roughly 8.5% Friday, with Nvidia shedding more than $300 billion in market capitalization, Micron Technology down 11% and Marvell Technology down 12%, Reuters reported. The selloff followed Broadcom's weak quarterly report on custom AI chips. Markets opened lower Monday, with the Guardian linking the broader equity slide to AI boom valuation concerns and the Iran-Israel oil shock. Alphabet upsized an equity offering to $84.75 billion to fund AI infrastructure, including a $10 billion private placement to Berkshire Hathaway, with capital expenditure now budgeted at $180 billion to $190 billion this year.

Read at WSJ ↗ • Read at CNBC ↗

Amazon unveils next-generation Proteus warehouse robot as part of $11.6 billion Europe push

Amazon unveiled an upgraded AI powered Proteus warehouse robot at its Delivering the Future event in Dartford, England, that responds to conversational prompts and is set to roll out across Europe in the first half of 2027 as part of a €10 billion European fulfillment investment, Reuters reported. Amazon also showed the STARK tote handling system, scheduled to expand to 15 European sites by 2027, and Vulcan, its first robot with a sense of touch. Amazon UK chief John Boumphrey told CNBC robots have "actually driven up employment rather than the reverse" at the company. The announcement lands as Challenger, Gray and Christmas reports AI is the leading cited reason for U.S. layoffs over the past three months.

Read at CNBC ↗

V.Global & Geopolitics

Starmer gives Apple and Google September deadline to block explicit images on children's phones; AI job center tools rolled out

Prime Minister Keir Starmer opened London Tech Week Monday by telling Apple, Google and other device makers they must introduce controls that prevent children from sending and receiving sexually explicit images within three months, or the government will legislate, ITV News reported. The Home Office framed the move as part of a new violence against women and girls strategy. Starmer also announced AI tools for UK jobseekers, including an online "AI assistant" and CV builder, AI training for 400,000 schoolchildren in disadvantaged areas, and an AI bootcamp scheme to roll out across England this summer, The Independent reported. Business Secretary Peter Kyle said the government will "aggressively" take larger stakes in fast growing UK firms to prevent a brain drain. A separate consultation on a full social media ban for under-16s closed at the end of May, with a decision pending.

Read at Guardian ↗ • Read at ITV News ↗ • Read at GOV.UK ↗

South Korea's Lee names former Naver CEO Han Seong-sook as PM with explicit AI mandate

South Korean President Lee Jae Myung on Sunday nominated Han Seong-sook, the incumbent minister of small and medium-sized enterprises and a former CEO of Naver, as the next prime minister, with presidential chief of staff Kang Hoon-sik saying Han is "the right person to flawlessly complete the great AI transformation," the Korea JoongAng Daily reported. If confirmed by the National Assembly, Han would be the Lee government's first female prime minister. Lee, marking his first year in office Monday, said his government will work to secure "absolute competitiveness" in advanced technologies and unveil a large-scale technology investment project, Yonhap reported. Lee also said excess semiconductor tax revenue should be invested in future growth engines.

Read at Bloomberg ↗ • Read at Korea JoongAng Daily ↗ • Read at Yonhap ↗

Netherlands adds AI to foreign investment screening regime effective January 2027

The Dutch government announced it will expand the Vifo foreign investment screening law to cover six additional sensitive technologies, including AI, effective January 1, 2027, Bloomberg reported, with the official Rijksoverheid announcement listing the categories. The expansion brings AI explicitly under a major EU member state's national security investment review and follows a multi-month thread of allied tightening of inbound Chinese capital screens. The move is a template Treasury, CFIUS staff and EU member states are expected to study.

Read at Bloomberg ↗ • Read at Rijksoverheid ↗

Russia pauses Putin surveillance system over AI CCTV targeting risk

Russia paused a surveillance system used in the security envelope around President Vladimir Putin after the use of AI enabled targeting against Iranian leadership exposed the risk that compromised CCTV networks could be used to locate senior officials, the Financial Times reported. The AP, in a separate piece, reported Israeli intelligence hacked into Iran's surveillance network during the Israel-Iran war and used it to gather real-time information for targeting strikes, including assessing damage after Israeli strikes. The Russian operational change confirms that a top adversary is reorganizing its security posture around AI driven CCTV risk.

Read at Financial Times ↗ • Read at AP ↗

Ireland tells data center developers to bring their own power as Dublin grid hits limits

Ireland is requiring new data center developers to bring their own power generation, with the country's grid operator having paused new data center connections near Dublin until 2028, the Wall Street Journal reported. Data centers consumed 21% of Ireland's national electricity last year, the highest reported share to the International Energy Agency for any country, the AP reported. Ireland's environment regulator has also flagged nitrogen oxide pollution from on-site gas or diesel generators. The "bring your own power" framework is the template U.S. states and other small grids are studying as they consider how to host AI infrastructure without grid risk.

Read at WSJ ↗ • Read at AP ↗

UK's Police.AI tells some forces to pause AI use in court statements over accuracy concerns

Alex Murray, head of the UK's Police.AI centre, has intervened at some police forces in England and Wales that were using commercially available AI tools to turn interviews into court statements before the technology had been properly assessed, the Financial Times reported. Murray said any technology used in the criminal justice system must meet a standard of accuracy "beyond reasonable doubt," and flagged disclosure schedules — evidence records owed to the defence — as needing particular caution. He pointed to a "cautionary tale" in which West Midlands Police last year relied on Microsoft Copilot output that fabricated a past match involving Maccabi Tel Aviv in a dossier supporting a ban on the Israeli club's supporters.

Read at Financial Times ↗