AI Daily Digest
Your daily briefing on AI developments that matter to your business.
🎧 Listen to this digestWednesday, June 24, 2026
Government & Policy
2Kids' Online Safety and Federal AI Preemption Deal Is Stalling in Congress
A White House effort to pair federal preemption of some state AI laws with the Kids Online Safety Act is running into resistance from lawmakers, advocates, and administration officials with conflicting priorities. A new Senate Judiciary Committee push to attach online safety provisions to the annual defense bill is further complicating the legislative picture. With Congress running short on calendar time, major federal AI legislation may slip past 2026.
See the business automation angle on this story — free with signup.
AI Regulation Becomes a Congressional Primary Battleground as NY-12 Race Turns Into Proxy Fight
New York State Assemblymember Micah Lasher won the Democratic primary for Rep. Jerry Nadler's Manhattan House seat after the race evolved into a proxy battle over AI safety legislation. Lasher's main rival, Alex Bores, had championed AI regulation bills signed into state law. Given the district's deep blue lean, Lasher is heavily favored in November, potentially shifting the balance of pro-regulation voices in Congress.
See the business automation angle on this story — free with signup.
Major AI Players
5Anthropic Launches Claude Tag in Slack as an Always-On AI Teammate
Anthropic has released Claude Tag in research preview, allowing businesses to summon an AI agent inside Slack by typing '@Claude.' The agent reads channel history, builds persistent memory about company workflows, and can pick up tasks mid-conversation — any team member can see what it has been working on. It is currently available in beta for Claude Enterprise and Team customers, with expansion to other platforms planned.
See the business automation angle on this story — free with signup.
Anthropic's Mythos Model Found Vulnerabilities in Classified U.S. Government Systems Under Project Glasswing
Anthropic's Mythos AI model identified security vulnerabilities in highly sensitive U.S. government computer systems during a restricted testing program called Project Glasswing, conducted in partnership with intelligence agencies. The revelation adds context to the export ban on Mythos — the model's unprecedented offensive cyber capabilities appear to be a central concern. Separately, Axios reported that OpenAI released a comparable cybersecurity model (GPT-5.5-Cyber) with far less political fallout, raising questions about what specifically triggered action against Anthropic.
See the business automation angle on this story — free with signup.
Legal Tech Startup Sues U.S. Government Over Anthropic Export Ban, Calling It an Existential Threat
Legion LegalTech Corp has filed a federal lawsuit challenging the Commerce Department directive that forced Anthropic to shut off its Fable 5 and Mythos 5 models for all customers, including its Canada-based engineering team. Legion argues the order unlawfully cut off access to tools critical to its business survival. The case is the first formal legal challenge to the AI model export controls that have roiled the industry.
See the business automation angle on this story — free with signup.
Meta Is the Only Frontier Lab That Hasn't Agreed to U.S. Government AI Model Vetting
The Trump administration has been pressing Meta to voluntarily submit its AI models to review by the Commerce Department's Center for AI Standards and Innovation (CAISI), but Meta has not yet signed on. Anthropic, OpenAI, Google, xAI, and Microsoft have all agreed to the arrangement. Meta says it shares the administration's goals and hopes to sign soon, but the holdout is notable given the political and security context.
See the business automation angle on this story — free with signup.
Meta Halts Employee Tracking Program for AI Training After Nearly 2,000 Workers Sign Protest Petition
Meta has paused a program that monitored employee activity to generate data for AI model training, after an internal petition gathered nearly 2,000 signatures demanding the initiative be cancelled. Workers objected to surveillance being imposed without consent, even as many said they support improving Meta's AI competitiveness. Meta had initially offered a partial opt-out of 30 minutes at a time before suspending the program more broadly.
See the business automation angle on this story — free with signup.
Global Developments
3China's LineShine Supercomputer Tops Global Rankings for First Time Since 2017, Displacing U.S. El Capitan
A supercomputer called LineShine, built by the Shenzhen Cloud Computing Center, has debuted at number one on the Top500 list of the world's most powerful computers, displacing the U.S. system El Capitan. Notably, LineShine runs entirely on conventional CPUs rather than the GPUs typically used for AI workloads. It is the first time a Chinese system has topped the list since 2017, reigniting debate about the U.S.-China tech race.
See the business automation angle on this story — free with signup.
Blackstone Plans $30 Billion AI Data Center Investment in Japan Over Five Years
Blackstone is in discussions to develop AI-focused data center infrastructure in Japan with capacity exceeding 1 gigawatt, representing a $30 billion commitment over three to five years according to Nikkei. The investment reflects surging institutional appetite for digital infrastructure tied to AI demand across Asia. Blackstone also plans to accelerate private equity activity in Japan alongside the data center buildout.
See the business automation angle on this story — free with signup.
Nissan Partners With Applied Intuition to Cut Software Development Time From Months to Hours Using AI
Nissan is partnering with AI startup Applied Intuition to dramatically accelerate software-defined vehicle development, aiming to compress timelines from months to hours in order to compete with Chinese automakers. The move reflects the broader automotive industry's recognition that software development speed is now a key competitive battleground. China's software-defined vehicle makers have set a pace that legacy automakers are struggling to match.
See the business automation angle on this story — free with signup.
Robotics
4ExRobotics Launches UL-Certified Autonomous Inspection Robot for North American Oil and Gas Market
ExRobotics has launched its UL-certified ExR-2.5 autonomous inspection robot in North America at the Energy Drone & Robotics Summit in Houston. The robot is designed for hazardous oil and gas environments where skilled labor is scarce and infrastructure is aging, and has logged thousands of missions for operators including Shell, Repsol, and BP. The UL certification is significant as it signals the inspection robot market has matured beyond pilots into regulated industrial deployment.
See the business automation angle on this story — free with signup.
Commerce Secretary Lutnick Signals Potential Action Against Chinese Robotics Imports After Closed-Door Industry Meeting
Commerce Secretary Howard Lutnick told executives at a closed-door meeting that his department is reviewing state-subsidized Chinese robotics imports and could take strong action once the review is complete. Administration officials argue the U.S. has lost critical manufacturing base needed to build next-generation robots. The government is also in the process of underwriting loans for at least two U.S. robotics companies — Foundation Robotics and Standard Bots — to help rebuild domestic capacity.
See the business automation angle on this story — free with signup.
Agility Robotics, Maker of the Digit Humanoid Used by Amazon, to Go Public in $2.5 Billion SPAC Deal
Agility Robotics, whose Digit humanoid robot is deployed in Amazon warehouses and Schaeffler manufacturing plants, has agreed to go public through a SPAC deal valuing the company at approximately $2.5 billion. The listing would make Agility one of the first humanoid robotics companies to trade publicly, providing a benchmark for a sector that investors are increasingly betting on. The deal reflects growing confidence that humanoid robots are transitioning from research projects to commercial products.
See the business automation angle on this story — free with signup.
Morgan Stanley Doubles China Humanoid Robot Forecast — Again — Projecting 50,000 Units This Year
Morgan Stanley has raised its 2026 China humanoid robot shipment forecast for the second time this year, now projecting 50,000 units — nearly double its prior estimate of 28,000. The bank estimates China's humanoid robot market will reach $2 billion in 2026 and grow to $15 billion by 2030. Chinese manufacturers have accelerated deployment in factories, convenience stores, and restaurants faster than analysts expected.
See the business automation angle on this story — free with signup.
Transportation
3AI Memory Demand Is Creating a Chip Shortage That Could Slow the Autonomous Vehicle Industry
Deutsche Bank has warned that soaring demand for high-bandwidth memory and DRAM from AI data center buildouts could leave automakers short of the chips needed for autonomous vehicles. Future Level 4 autonomous cars may require more than 300GB of memory, but memory manufacturers are prioritizing higher-margin AI infrastructure customers. The shortage represents an unexpected second-order consequence of the AI infrastructure boom.
See the business automation angle on this story — free with signup.
Finland May Approve Tesla FSD Before EU-Wide Vote, Becoming Second European Country to Clear the System
Finland's transport authority Traficom has indicated it may grant approval for Tesla's Full Self-Driving assisted driving system ahead of the EU-wide committee vote expected in October 2026. The Netherlands became the first European country to grant provisional FSD approval in April. Traficom is evaluating driver takeover times, Finnish road conditions, and speed offset behavior before making a final decision.
See the business automation angle on this story — free with signup.
Tesla Confirms FSD Was Active in Fatal Texas Crash but Blames Driver for Flooring Accelerator to 100%
Tesla has now confirmed its Full Self-Driving system was engaged when a Model 3 struck and killed a 76-year-old woman inside her Texas home, but says the driver overrode it by pressing the accelerator to 100% — reaching 73 mph in a residential area. This is a new development from the story first covered June 21: Tesla had initially disputed FSD involvement, but its own data now places the system active at the moment of impact. NHTSA has opened an investigation.
See the business automation angle on this story — free with signup.
Healthcare & Biotech
3Cadence Raises $100M and Announces Duke Health and Texas Health Resources as New Partners
Remote patient monitoring startup Cadence has raised $100 million and announced new partnerships with Duke Health and Texas Health Resources to expand its AI-assisted chronic care management platform. The funding reflects continued investor appetite for AI tools that operate at the intersection of clinical workflows and continuous monitoring. Cadence's platform uses AI to help care teams manage patients with conditions like hypertension and diabetes between visits.
See the business automation angle on this story — free with signup.
IGC Pharma's AI Platform Cuts Alzheimer's Research Data Harmonization Time by 90%
IGC Pharma has released a beta version of its Agentic Harmonization Assistant (AHA), an AI platform designed to merge fragmented Alzheimer's and aging datasets from different research sources. In representative workflows, AHA reduced data harmonization time by 90%, potentially accelerating the discovery timeline for Alzheimer's treatments. The tool is part of IGC's broader strategy to pair AI-enabled research infrastructure with its own clinical drug pipeline.
See the business automation angle on this story — free with signup.
Major Review: Human-AI Healthcare Teams Only Outperform Humans When Clinicians Remain in Control
A scoping review published in npj Digital Medicine found that AI improves clinical outcomes only when qualified professionals maintain oversight and decision authority — unsupervised AI deployment in critical settings consistently underperforms or introduces new risks. The study analyzed AI effectiveness across clinical tasks and identified human oversight, organizational readiness, and governance frameworks as the key determinants of success. WHO, the EU AI Act, and the FDA all align with these findings.
See the business automation angle on this story — free with signup.
Investment Noteworthy
4AI Memory Startup Engram Raises $98 Million, Claims to Match Frontier Models Using 100x Fewer Tokens
Engram, a 13-person AI startup less than a year old, has raised nearly $100 million to build what it calls the 'learned memory' layer for AI — a system that stores organization-specific workflows and context so models can give smarter answers at a fraction of the usual token cost. The company claims its approach can match or outperform frontier models while using up to 100 times fewer tokens, and already counts Microsoft, Notion, and Harvey among its clients.
See the business automation angle on this story — free with signup.
VCs Are Openly Debating Whether the AI Market Is in a Bubble — Even as Revenues Justify Valuations
At TechCrunch's StrictlyVC event, leading investors described the current AI market as a 'paradoxical era' where valuations appear disconnected from fundamentals, yet the underlying revenue growth is real — with some startups jumping from $10M to $70M ARR in a single year. Investors from M13 and Basis Set Ventures noted that startups now compete not just against each other but against the ten most powerful tech companies in the world. The debate reflects genuine tension between AI's proven near-term value and the long-term sustainability of current valuation multiples.
See the business automation angle on this story — free with signup.
AI Voice Scheduling Startup Assort Health Hits $1.2 Billion Valuation After $120M Raise Led by Menlo Ventures
Assort Health, founded by two 29-year-olds, has raised $120 million in its third funding round in just 14 months, valuing the company at $1.2 billion. The startup's AI voice chatbot handles appointment scheduling for more than 15,000 physicians, primarily at specialty providers like orthopedics and dermatology. The round was led by Menlo Ventures, fresh off its record $3 billion fund close.
See the business automation angle on this story — free with signup.
Menlo Ventures Raises Record $3 Billion Fund, Driven by $14 Billion Stake in Anthropic
Menlo Ventures has closed a $3 billion fundraise — the largest in its 50-year history — anchored by its early bet on Anthropic, which has grown its stake to an estimated $14 billion. The firm originally led Anthropic's Series D with a $750 million commitment, widely described at the time as a bet-the-firm move. Menlo also operates a $100 million startup fund called Anthology in partnership with Anthropic.
See the business automation angle on this story — free with signup.
General AI
3Meta Launches AI-Powered End-to-End Creative Ad Solutions With WPP Integration at Cannes
Meta has unveiled a suite of AI-powered advertising tools including an end-to-end creative solution that learns a brand's identity and goals from existing ads, a unified creator marketplace hub, and expanded AI text-generation options for ad copy. The tools include integrations with agency giant WPP and new AI-powered options for generating multiple message angles and creative variations automatically. The announcements came alongside other major ad industry news at Cannes Lions.
See the business automation angle on this story — free with signup.
Nvidia Unveils BioNeMo Agent Toolkit to Automate Complex Scientific Research Workflows
Nvidia has launched the BioNeMo Agent Toolkit, which converts complex scientific tasks — including protein structure prediction, molecular docking, and genomic analysis — into agent-executable workflows powered by its NIM microservices. The toolkit is built on collaborations with the Arc Institute, the University of Washington's Institute for Protein Design, and the Open Molecular Software Foundation. Nvidia frames it as enabling AI to iterate through biological complexity at speeds no human researcher could match.
See the business automation angle on this story — free with signup.
Software Engineers Are Productive but Anxious as AI Coding Tools Accelerate Faster Than They Can Learn Them
A Business Insider report finds that software engineers are caught between excitement at AI productivity gains and deep anxiety about falling behind as new tools release weekly. Some developers report 'tool paralysis' — the sense that becoming a subject matter expert in any one AI coding tool is futile given the pace of change. The phenomenon is being watched as a preview of the psychological and professional disruption AI will bring to other knowledge work sectors.
See the business automation angle on this story — free with signup.