AI Daily Digest
Your daily briefing on AI developments that matter to your business.
Thursday, June 18, 2026
Government & Policy
4Trump's 'Shadow AI Policy' Takes Shape Through Export Controls and Voluntary Testing Frameworks
Without formal legislation from Congress, the Trump administration has assembled a de facto AI policy through export controls, voluntary model-vetting frameworks, and federal procurement guidelines. Axios reports this approach shapes the industry without ever publishing clear rules, leaving companies to interpret shifting case-by-case decisions. The Anthropic export control situation is being watched by other AI labs as a signal of how far the administration will stretch its legal authority.
See the business automation angle on this story — free with signup.
Bernie Sanders Introduces Bill to Create Public Sovereign Wealth Fund Through 50% Tax on AI Giants
Senator Bernie Sanders has unveiled the AI Sovereign Wealth Fund Act, which would impose a one-time 50% tax on the stock of the largest AI companies to seed a government-managed fund giving the public a direct ownership stake in the industry. A seven-person independent commission would manage the fund and use voting shares to block decisions deemed harmful to the public. Notably, President Trump has also mused about government ownership of AI companies, creating a rare ideological overlap on the idea of public stakes in AI.
See the business automation angle on this story — free with signup.
States Push Forward With Targeted AI Laws as Federal Rules Stall
With Congress unable to pass comprehensive federal AI legislation, individual U.S. states are advancing more targeted laws covering areas like children's interactions with AI, employer use of AI systems, and catastrophic risk prevention. The Trump administration has urged Congress to preempt state laws that conflict with its own regulatory framework, calling state-level fragmentation a risk to U.S. competitiveness with China. Businesses operating across multiple states now face a patchwork of potentially conflicting obligations.
See the business automation angle on this story — free with signup.
Rep. Gottheimer Prepares Bill Mandating Government Vetting of Powerful New AI Models
New Jersey Democrat Josh Gottheimer is introducing legislation that would require AI companies developing the most powerful models to submit them to the government for review of national security, cybersecurity, and bioterrorism risks before release. The bill is explicitly tied to concerns raised by Anthropic's Mythos model incident and runs parallel to a broader bipartisan effort to craft federal AI rules that would override conflicting state laws. Unlike the White House's current voluntary framework, Gottheimer's bill would make the review process mandatory.
See the business automation angle on this story — free with signup.
Major AI Players
6Pentagon Discloses Elon Musk's Grok AI Helped Fire 2,000 Munitions at Iran in 96 Hours
In a sworn court filing, the Pentagon's chief digital and AI officer disclosed that xAI's Grok chatbot was used to fire over 2,000 munitions at 2,000 distinct targets within a 96-hour window during strikes on Iran — the first explicit acknowledgment by a senior U.S. official of a commercial AI chatbot being used in active kinetic military operations. The disclosure came as part of a government brief defending xAI against an NAACP environmental lawsuit over its Colossus 2 data center. The revelation is raising urgent questions in Congress about auditability, liability, and oversight of commercial AI in warfare.
See the business automation angle on this story — free with signup.
G7 AI Summit: Western Leaders and Tech CEOs Push for U.S.-Led Global AI Coalition to Counter China
At the G7 summit in Évian-les-Bains, Anthropic CEO Dario Amodei and Google DeepMind CEO Demis Hassabis jointly called for a U.S.-led international coalition to set standards for frontier AI development, with both warning against fragmentation that could benefit China. G7 leaders simultaneously discussed creating a 'trusted partners' scheme to allow allied nations to access advanced U.S. AI models despite current export restrictions. Trump said Anthropic negotiations were 'going fine' while French President Macron warned that the ability to 'turn off the switch' on AI access was eroding trust in American technology.
See the business automation angle on this story — free with signup.
Google's Gemini Co-Lead Noam Shazeer Leaves to Join OpenAI Ahead of Its IPO
Noam Shazeer, the VP of engineering at Google and co-lead of its Gemini AI models, announced he is leaving Google to join OpenAI — a stunning departure given that Google reportedly paid $2.7 billion to bring him back from Character.AI less than two years ago. Shazeer has been credited as a key figure in closing Gemini's performance gap with ChatGPT. The move is a significant talent win for OpenAI as it prepares for its IPO.
See the business automation angle on this story — free with signup.
Anthropic Ships Major Claude Design Update Merging Visual Design and Code in a Single Workflow
Anthropic has released a significant update to Claude Design, integrating it directly with Claude Code so that design systems and code stay synchronized in real time — designers can pull assets into a coding terminal with a single command and vice versa. The update also introduces an administrator role for managing company-wide brand kits, addressing enterprise needs for brand consistency at scale. Claude Design has attracted over one million users since its April beta launch.
See the business automation angle on this story — free with signup.
Anthropic Restrictions Create Winners: Mistral and DeepSeek Benefit as Open-Weight Models Gain Appeal
Business Insider analysis finds that the U.S. restrictions on Anthropic's Fable 5 and Mythos 5 models are accelerating interest in open-weight alternatives like Mistral and DeepSeek, which customers can download and deploy on their own infrastructure without relying on vendor-controlled access. Mistral CEO Arthur Mensch has doubled down on the company's sovereignty pitch, announcing its upcoming models will be open-weight. Unlike Anthropic's cloud-only approach, self-hosted models cannot be remotely shut down by either the developer or a government.
See the business automation angle on this story — free with signup.
Wired Reveals SK Telecom's China Ties Were the Trigger for Anthropic Export Control Crisis
New reporting from Wired reveals that the Anthropic export control crisis was specifically triggered by the White House ordering Anthropic to revoke South Korean telecom giant SK Telecom's access to Claude Mythos, due to alleged ties to China. The administration concluded it could not trust Anthropic to self-police access to its most advanced models, which led to the blanket order cutting off all foreign nationals. This is a significant new development that explains the specific chain of events behind the shutdown covered in previous digests.
See the business automation angle on this story — free with signup.
Global Developments
2DeepSeek Closes $7.4B Funding Round at $50B+ Valuation With a Strict No-Poaching Clause
Chinese AI lab DeepSeek has closed its first external funding round, raising $7.4 billion and achieving a valuation of over $50 billion — making it China's most valuable AI-only startup. Founder Liang Wenfeng set a non-negotiable condition for all investors: they must not poach DeepSeek's staff or encourage employees to start competing companies. The unusual clause reflects the company's tight control over its most valuable asset as it transitions from research lab to a commercially funded enterprise.
See the business automation angle on this story — free with signup.
China Plans $295 Billion AI Data Center Network With State Telecoms at the Core
China is reportedly preparing to invest approximately CNY2 trillion ($295 billion) over five years to build a nationwide network of AI-focused data centers and computing hubs, with state-owned telecoms like China Mobile and China Telecom taking central operational roles. The plan, being drafted by the National Development and Reform Commission, is part of a broader national infrastructure strategy announced in April. Analysts say this positions Chinese telcos as foundational players in the country's AI ecosystem in a way that has no Western equivalent.
See the business automation angle on this story — free with signup.
Robotics
2Amazon's Next Warehouse Efficiency Drive Uses AI to Optimize How Humans Move, Not Just Packages
Inside Amazon's robot-enabled fulfillment centers, the company is now deploying AI systems to optimize the movement of human workers — not just packages — with internal analysis projecting millions of labor hours eliminated annually if the system scales. The initiative targets what Amazon calls its 'single largest labor automation opportunity': the role responsible for directing workers to the right locations at the right time. The goal, per internal documents, is to 'remove the dependency on manual staffing decisions.'
See the business automation angle on this story — free with signup.
Sanctuary AI Achieves World-Class Performance on Complex Wire-Plugging Task at Automotive Supplier
Sanctuary AI has announced it achieved 'world-class performance' on a wire-plugging task at a global Tier 1 automotive supplier, where its physical AI system handled materials shifting dynamically on a conveyor belt — a task requiring fine dexterity and real-time adaptation. The company's strategy focuses on deploying its AI software platform on existing industrial robots rather than requiring customers to adopt entirely new hardware. This validates Sanctuary's 'physical AI as a software layer' approach to the manufacturing market.
See the business automation angle on this story — free with signup.
Transportation
3Uber Partners With Lucid and Nuro to Launch Premium Robotaxi Service in Houston by Mid-2027
Uber has announced a partnership with EV maker Lucid and autonomous driving startup Nuro to launch a premium driverless robotaxi service in Houston by mid-2027, making it Uber's second major autonomous market after San Francisco. Uber has invested nearly $500 million in Nuro and committed to purchasing at least 35,000 Lucid Gravity SUVs equipped with Nuro's autonomous system. The service will put Uber in direct competition with Waymo in one of the U.S.'s largest metro markets.
See the business automation angle on this story — free with signup.
New Rankings Show Baidu, Pony.ai, and WeRide Ahead of Tesla and Zoox in Robotaxi Race
A new database from AV research firm Autnmy AI, using a proprietary algorithm to measure real commercialization progress, finds that three Chinese companies — Baidu's Apollo Go, Pony.ai, and WeRide — rank significantly ahead of Tesla and Zoox in the global robotaxi race, behind only Waymo. The rankings assess actual deployment, safety data, regulatory approvals, and revenue generation rather than headlines or hype. The findings suggest China's robotaxi sector is far more advanced commercially than most Western coverage implies.
See the business automation angle on this story — free with signup.
Dutch Road Authority Reports 40,000 Teslas Have Driven 24 Million Km on FSD With No Serious Incidents
The Dutch road authority RDW announced that 40,000 Tesla vehicles in the Netherlands have collectively driven 24 million kilometers using Tesla's Full Self-Driving supervised software since its April approval — the first such European authorization. The agency is monitoring the technology monthly rather than annually and is now seeking EU-wide approval. The data represents an early but significant real-world safety dataset for supervised autonomy in European driving conditions.
See the business automation angle on this story — free with signup.
Healthcare & Biotech
2MSD Signs $510M+ Partnership With Protillion Biosciences for AI 'Lab-in-the-Loop' Drug Discovery
Merck & Co (MSD) has formed a partnership worth over $510 million with Protillion Biosciences, centered on the startup's 'lab-in-the-loop' AI platform that iteratively designs drug candidates, tests them in wet labs, and feeds results back to refine AI models. The approach is gaining traction across pharma as companies seek to accelerate biologics discovery by generating high-quality experimental data at scale. Protillion's CEO argues the rate-limiting step is no longer AI model performance but the availability of high-quality lab-generated data.
See the business automation angle on this story — free with signup.
Harvard and Stanford Study Confirms ChatGPT Outperforms Hundreds of Physicians in Diagnostic Testing
A study by Harvard and Stanford researchers, published in the journal Science, found that ChatGPT outperformed hundreds of physicians in a structured diagnostic obstacle course using real-world patient data. At least one generative AI product has already received FDA approval for clinical reasoning, and hospital systems are beginning to roll out AI-powered diagnostic tools to clinicians. Experts caution that current AI tools are text-only and cannot replicate the full complexity of in-person clinical care, but the performance gap is generating serious discussion about AI's role in medicine.
See the business automation angle on this story — free with signup.
Investment Noteworthy
2World Model AI Startup Odyssey Raises $310M Series B at $1.45B Valuation Backed by Amazon
Odyssey, an LA-based AI startup building world models that simulate physical environments with accurate physics, has raised a $310 million Series B at a $1.45 billion valuation, led by Natural Capital with Amazon, AMD Ventures, GV, and others participating. The company was founded by self-driving vehicle pioneers and offers world models for use cases ranging from video game creation to robotics training. As part of the deal, AWS becomes Odyssey's preferred cloud provider and its models will be optimized for Amazon's Trainium chips.
See the business automation angle on this story — free with signup.
SpaceX Acquires AI Coding Tool Cursor for $60 Billion in Stock
SpaceX has agreed to acquire Cursor, the AI coding assistant developed by Anysphere, in a $60 billion all-stock deal — one of the largest AI acquisitions on record — executed just days after SpaceX's IPO. Before SpaceX's offer, Cursor was on track to close a $50 billion funding round from Andreessen Horowitz, Thrive, and Nvidia. The deal includes a $10 billion breakup fee if the acquisition falls through, signaling how aggressively SpaceX pursued the asset.
See the business automation angle on this story — free with signup.
General AI
2Mathematicians Launch 'Leiden Declaration' to Protect Scientific Integrity From AI Encroachment
An international group of researchers from 15 universities across Europe and the U.S. has published the Leiden Declaration, calling for protections to preserve the core values of mathematics as AI systems increasingly solve problems that were previously the exclusive domain of human researchers. The declaration — endorsed by the International Mathematical Union — raises concerns about mathematical proofs generated by proprietary AI systems that academic researchers cannot inspect, audit, or reproduce. It does not call for a ban on AI but advocates for transparency and open access standards.
See the business automation angle on this story — free with signup.
Enterprises Still Struggling to Measure ROI on AI Spending, Venture Investor Warns
NEA partner Tiffany Luck told TechCrunch's Equity podcast that many large enterprises are hitting unexpected financial hurdles with AI implementation because they lack tools to track costs and measure actual return on investment. Companies are increasingly seeking platforms that help monitor AI spend efficiency rather than just buying AI models. Luck predicts that personal AI agents creating 'magic moments' for consumers will be the next major growth driver — but only after companies solve the ROI measurement problem.
See the business automation angle on this story — free with signup.