---
date: 2026-06-29
subject: "Lutnick eases Mythos 5 ban | Voters back AI safety reviews | Trump digital-tax tariff | Apple courts blacklisted chipmaker"
---

**Commerce Secretary Lutnick partially lifted the Mythos 5 export freeze**, clearing the Anthropic model for more than 100 vetted U.S. companies and federal agencies while keeping Fable 5 offline for public use. **A bipartisan majority of voters wants mandatory safety reviews for powerful AI** before public release, an AI Policy Institute poll found, going further than the administration's current opt-in approach. **Trump threatened a 100% tariff** on any country that imposes a digital services tax on American tech firms, singling out European capitals weeks before a July 4 trade deadline. **Apple is lobbying the administration to buy memory chips from CXMT**, the blacklisted Chinese supplier on the Pentagon's military-companies list.

# 1. AI Policy Today

- **Lutnick clears Mythos 5 for 100+ U.S. companies and federal agencies, Fable 5 still blocked** — Commerce Secretary Howard Lutnick sent Anthropic chief compute officer Tom Brown a June 26 letter authorizing access to the Claude Mythos 5 model for trusted partners, partially reversing the June 12 export directive, per TechCrunch. More than 100 U.S. companies and federal agencies, plus their non-American employees and Anthropic's own non-American employees, can now use Mythos 5, with many drawn from Anthropic's roughly 100-company Project Glasswing roster. Fable 5, the public facing Mythos-class model, remains offline with no announced timeline for restoration. The Foundation for Individual Rights and Expression's John Coleman said "no one knows how these companies are picked and why everyone else is excluded," and OpenAI CEO Sam Altman wrote on X that he did not like "the idea of the government picking the customers." [Politico](https://www.politico.com/news/2026/06/26/white-house-makes-peace-with-anthropic-for-now-00965675) [TechCrunch](https://techcrunch.com/2026/06/26/trump-admin-releases-anthropic-mythos-to-be-used-by-more-than-100-us-companies-agencies/) 

- **Voters across parties back mandatory safety reviews for powerful AI, AIPI poll finds** — An overwhelming majority of likely voters want advanced AI systems to undergo mandatory safety reviews before public release — going further than the Trump administration's current opt-in policy — according to a new survey from the AI Policy Institute (AIPI) reported by NBC News. Republicans were more enthusiastic about government-led safety testing than Democrats, though more than half of voters backed it regardless of party. Among the 1,007 likely voters surveyed June 10 and 11, two-thirds preferred AI with guardrails over an outright ban, more than 60% of both Republicans and Democrats said the federal government — not AI companies — should set safety standards, and more than 80% said companies shouldn't build smarter-than-human systems until they can prove they control them. "I think people in the White House who have been trying to push a no-rules-whatsoever perspective are out of step with the American people," said Peter Wildeford, director of policy at the AI Policy Network. [NBC News](https://www.nbcnews.com/tech/tech-news/voters-both-parties-want-tighter-ai-regulation-poll-finds-rcna351815) [AI Policy Institute](https://theaipi.org/poll-ai-safety-majority/)

- **Trump threatens 100% tariff on any country that imposes a digital services tax on U.S. firms** — President Donald Trump on Friday threatened a 100% tariff on imports from any country that imposes a tax on digital services from U.S. companies, in a Truth Social post that singled out European nations he said were close to implementing such levies. Trump said the tariff would supersede previously negotiated trade deals and would be imposed immediately if a country proceeded with a digital services tax. European Commission spokesperson Olof Gill said the EU would respond decisively to defend its regulatory autonomy if Washington pursued the threat. Roughly half of Europe's OECD members have proposed, announced or already implemented a digital services tax, according to the Tax Foundation, with such taxes designed to capture revenue from large U.S. technology companies that operate in their economies without a local taxable presence, per CBS News. The post landed eight days before Trump's July 4 deadline for finalizing U.S.-EU implementation of the May trade deal capping most EU tariffs at 15%, which left digital taxes unresolved. [The Washington Post](https://www.washingtonpost.com/business/2026/06/26/trump-tariff-europe/bfde159c-717d-11f1-8730-e7fd0e2a6404_story.html) [CBS News](https://www.cbsnews.com/news/trump-digital-services-tax-tariff-europe/) [FT](https://www.ft.com/content/5d886d47-c509-44a4-9077-bcd25158b61e)

- **Apple lobbies Trump administration to buy memory chips from blacklisted Chinese supplier CXMT** — Apple is lobbying the Trump administration for permission to source memory chips from ChangXin Memory Technologies, the Chinese supplier on the Pentagon's Section 1260H Chinese Military Companies list, per the Financial Times. The iPhone maker has approached the White House, the Commerce Department and other administration officials seeking assurances that purchases would not trigger future restrictions, citing pressure from rising DRAM and NAND prices driven by AI related demand. CXMT is China's largest memory chipmaker and has been approved for addition to the Commerce Department's Entity List. U.S. exports of goods, software and technology to listed firms generally require a license that is typically denied. Apple recently raised iPad and MacBook prices by 20%, with CEO Tim Cook calling the memory cost jump unlike anything he had seen in over 40 years.  [FT](https://www.ft.com/content/d72a25e2-7bde-4aa9-bd8d-0c4f3d6cb2cb) [CNBC](https://www.cnbc.com/2026/06/25/apple-macbook-ipad-price-hike-memory.html)

# 2. China Watch

- **Beijing puts 20 Japanese defense linked firms under export controls, citing "remilitarization"** — China's Ministry of Commerce added 20 Japanese entities to its export control list today, blocking Chinese exporters from sending dual use items to four Japanese defense research institutes and nine Mitsubishi defense subsidiaries, Caixin reported. A second batch of 20 Japanese firms was placed on a separate watch list. A ministry spokesperson said the move was a response to Japan's "remilitarization" and nuclear ambitions, adding that it targets only a small number of entities and would not disrupt broader China-Japan trade. China's dual use export regime, first sharpened on rare earths and chip making inputs, is now being applied to a U.S. treaty ally, expanding the ministry's retaliation toolkit against allied suppliers feeding U.S. AI and defense supply chains. [Caixin](https://www.caixin.com/2026-06-29/102458626.html)

- **DeepSeek founder Liang says Anthropic's flagship forced him to abandon the lab's no-funding stance** — DeepSeek founder Liang Wenfeng decided to break the lab's longtime "no funding, no commercialization, no roadshow" stance after watching Anthropic's Claude Mythos demonstrate capabilities trained on massive compute and data, QbitAI reported. Anthropic released its first Mythos preview in April; DeepSeek's funding rumors emerged the same month, and the round has since closed at roughly $7.4 billion. Every DeepSeek department will at least double headcount, with the newly stood up Harness AI agent team already interviewing daily for engineers to convert DeepSeek's models into autonomous agents. The pivot puts DeepSeek into open dependence on outside capital for the first time, pulling the lab squarely inside the U.S. Commerce Department export control regime that already denies it access to Nvidia's frontier GPUs. [QbitAI](https://www.qbitai.com/2026/06/439863.html)

- **China's models extend their global API lead to a ninth straight week as the U.S. share keeps falling** — Chinese AI models held the global lead in API call volume on OpenRouter, the AI model marketplace, for a ninth consecutive week through June 22-28, with DeepSeek-V4-Flash, MiMo-V2.5 and MiniMax M3 topping the charts, Pandaily reported. Chinese models drew 20.39 trillion tokens that week against 4.25 trillion for U.S. models, with platform-wide volume at 46.7 trillion. The U.S. share of OpenRouter token requests has slid from 72% to 33% over the period covered, citing ExponentialView research. The shift extends the picture AIPD's [June 8th edition](https://aipolicydaily.org/archive/daily/2026-06-08/) flagged at six weeks. Chinese open-weight models such as DeepSeek, GLM and MiniMax are pulling developer demand from U.S. counterparts at a pace current Commerce Department export control design was not built to address. [Pandaily](https://pandaily.com/china-ai-model-api-calls-nine-weeks-jun2026)

- **Z.ai's open-weight GLM-5.2 reportedly matches Mythos in cybersecurity scenarios** — Chinese AI startup Z.ai, formerly Zhipu AI, released its open-weight GLM-5.2 model, with some researchers claiming parity with Anthropic's Mythos in bug finding and cybersecurity scenarios. GLM trails U.S. frontier labs on broader general purpose tasks, per The Verge. CNBC said GLM-5.2 makes open source AI a real contender on intelligence-per-dollar against Anthropic and OpenAI. The WSJ reported the assessment is fueling concern in Washington that the clampdown on top U.S. AI is handing Beijing a cyberwarfare advantage. [The Verge](https://www.theverge.com/ai-artificial-intelligence/958804/chinas-z-ai-glm-52-mythos-cybersecurity) [CNBC](https://www.cnbc.com/2026/06/26/china-zhipu-z-ai-open-source-anthropic-openai.html) [WSJ](https://www.wsj.com/tech/ai/chinese-ai-anthropic-mythos-cybersecurity-574b02c2?mod=rss_Technology)

# 5. Industry & Market Watch

- **Advanced chip packaging emerges as AI's compute choke point, deepening U.S. reliance on TSMC** — Advanced chip packaging, which bundles silicon components into the palm-size modules that have become essential for AI accelerators, has shifted U.S. compute supply dependence further onto Taiwan than ever, the New York Times reported. UCLA professor and former IBM technologist Subramanian Iyer said U.S. leadership in the packaging field has slipped while Taiwan Semiconductor Manufacturing Co. has overtaken the segment alongside its lead in leading node chip manufacturing. Nvidia and other chip giants have turned to packaging as the principal way to deliver more capable AI semiconductors as transistor-shrink gains slow, per The Indian Express. TSMC's struggle to keep up with demand has made the packaging bottleneck a focus of Silicon Valley supply planning. [NYT](https://www.nytimes.com/2026/06/26/technology/ai-advanced-chip-packaging-tsmc.html) [The Indian Express](https://indianexpress.com/article/technology/artificial-intelligence/how-a-niche-technology-became-a-choke-point-for-ai-10761049/lite/)

- **Asian AI labs launch Mythos-like models outside U.S. export control reach** — Asian AI startups are launching new models advertised as Mythos-like in capability and unconstrained by U.S. export controls, TechCrunch reported Saturday. Japan based Sakana AI's Fugu release, an orchestration model built to coordinate other AIs through their APIs, and Chinese cybersecurity firm 360's Tulongfeng, pitched as a head-to-head competitor to Anthropic's Mythos, anchor the piece. A Sakana spokesperson told TechCrunch the launch timing alongside the Mythos export ban was "entirely coincidental," even as the company's website advertises "delivering frontier capability without the risk of export controls." [TechCrunch](https://techcrunch.com/2026/06/27/asian-ai-startups-launch-mythos-like-models-as-anthropics-export-ban-drags-on/) [Sakana AI](https://sakana.ai/fugu-release/)

- **Erin Brockovich organizes nationwide grassroots opposition to AI data center buildout** — Environmental advocate Erin Brockovich, who helped win a $333 million settlement against Pacific Gas and Electric in the 1990s Hinkley contamination case, has turned her attention to local opposition campaigns against AI data center construction, per the Guardian. After posting a callout in April on her website asking residents with data center concerns to write in, Brockovich said 3,862 people replied within a month. She told the Guardian the new wave of AI powered data center construction "feels like Hinkley on steroids." [The Guardian](https://www.theguardian.com/environment/2026/jun/29/were-up-against-forces-that-have-all-the-money-in-the-world-erin-brockovich-on-her-battle-against-ai-datacentres)

# 6. Global & Geopolitics

- **South Korea unveils $880 billion Samsung and SK Hynix package built around southwest chip cluster and 8.4 GW of AI data centers** — South Korean President Lee Jae Myung announced investment plans pairing 800 trillion won ($520 billion) from Samsung Electronics and SK Hynix for four new chip facilities in the country's southwest region with 550 trillion won (about $360 billion) for three new AI data centers, at a Monday Blue House briefing, per Korea JoongAng Daily. The package adds 81 trillion won (about $53 billion) for a high-bandwidth memory packaging hub in Cheonan and Onyang and 30 trillion won (about $20 billion) for next-generation memory and edge AI semiconductors. The government will partner with SK, GS and Naver on data center capacity totaling 8.4 gigawatts, with SK alone slated to scale to 15 GW by 2035 and lifting national capacity to 18.4 GW, per Yonhap News Agency. Samsung Executive Chairman Lee Jae-yong and SK Group Chairman Chey Tae-won joined President Lee at the briefing. [Bloomberg](https://www.bloomberg.com/news/articles/2026-06-28/samsung-sk-reportedly-to-invest-1-3-trillion-over-10-years) [Korea JoongAng Daily](https://www.koreajoongangdaily.com/business/samsung-sk-hynix-to-invest-520b-to-build-chip-plants-in-regional-growth-push/12744777) [Yonhap News Agency](https://en.yna.co.kr/view/AEN20260629006151315) 

- **EIB pledges €3 billion Airbus loan as EU presses tech sovereignty agenda** — The European Investment Bank will provide €3 billion ($3.4 billion) in financing to Airbus SE as the European Union pushes ahead with plans to bolster its technological sovereignty amid intensifying competition from the U.S. and China, per Bloomberg. The first €1 billion (about $1.1 billion) tranche, signed Monday in Brussels, will support Airbus's research, development and innovation programs in France, Germany and Spain across commercial and defense aviation, per The Edge Malaysia. EIB president Nadia Calviño said the bank approved the loan about six months after Airbus submitted its request. The EU has expanded EIB lending in recent years to defense, critical raw materials and clean technology as Brussels treats those sectors as strategic. [Bloomberg](https://www.bloomberg.com/news/articles/2026-06-29/eib-pledges-record-3-billion-airbus-loan-for-tech-development) [The Edge Malaysia](https://theedgemalaysia.com/node/808694)

- **German economy pitched as AI productivity case with €300 billion gain projection from clerical automation** — A made-to-order homebuilder in northwest Germany cut invoice processing time from four working days to two after introducing AI to its back office last year, Bloomberg reported in a case study Monday framing AI rollout as a partial answer to the country's worker shortage. Multiplied across the German economy, the productivity gain from this kind of clerical automation could reach around €300 billion (about $340 billion). The personnel firm Personio has separately put German economic productivity losses from worker disengagement at up to €142.3 billion (about $161 billion), with sector specific AI contribution forecasts running well below that. The German labor market is short hundreds of thousands of skilled workers each year. [Bloomberg](https://www.bloomberg.com/news/articles/2026-06-29/german-ai-rollout-offers-300-billion-fix-for-worker-shortage)