AI Daily Digest
Your daily briefing on AI developments that matter to your business.
🎧 Listen to this digestThursday, June 25, 2026
Government & Policy
2Pro-AI Super PAC's $100M Campaign Trail Spending Yields Mixed Primary Results
Leading the Future, a pro-AI industry super PAC backed by tech executives and investors, has spent heavily across nearly 30 races to support lighter AI regulation, but Tuesday's primary results showed no clear pattern of influence. Axios reports that the PAC notably did not claim credit for any wins, and at least one targeted candidate who leaned into anti-AI-money messaging plans to continue pushing AI policy in the state legislature. The results suggest AI industry money is not yet reliably decisive at the ballot box.
See the business automation angle on this story — free with signup.
Trump's June 2 AI Executive Order Creates New Voluntary-but-Consequential Framework for Industry Cybersecurity Collaboration
President Trump signed Executive Order 14409 on June 2, establishing a new framework for government-industry collaboration on AI cybersecurity and the secure deployment of advanced AI models. Legal analysts at Law360 note that while the framework is framed as voluntary, its implications for companies working with the federal government carry mandatory weight in practice. The order builds on existing AI security initiatives and is designed to govern how advanced AI models are shared with and used by government agencies.
See the business automation angle on this story — free with signup.
Major AI Players
6Anthropic Accuses Alibaba of Running the Largest-Ever AI Extraction Campaign Against Claude
Anthropic has written to U.S. senators accusing Alibaba-linked operators of conducting nearly 29 million exchanges with Claude through thousands of fraudulent accounts in what it calls the largest 'distillation attack' campaign ever recorded. The goal, Anthropic alleges, was to harvest Claude's software engineering and agentic reasoning capabilities to train Alibaba's own Qwen AI models. Anthropic says such attacks allow Chinese companies to replicate expensive U.S. AI capabilities at a fraction of the cost.
See the business automation angle on this story — free with signup.
OpenAI Unveils Its First Custom AI Chip, 'Jalapeño,' Built With Broadcom
OpenAI has revealed its first in-house AI processor, called Jalapeño, developed in partnership with Broadcom to handle inference workloads for ChatGPT and Codex. The chip is an ASIC (not a GPU) and Broadcom's CEO claims it matches the performance of Nvidia's Blackwell chips and Google's TPUs. The move is a significant step toward OpenAI becoming a full-stack AI platform less dependent on Nvidia.
See the business automation angle on this story — free with signup.
Claude Reportedly Expresses Unease About Its Role in U.S. Military Targeting System
The Atlantic reports that a version of Anthropic's Claude is integrated into the U.S. military's Maven Smart System, which fuses satellite imagery and sensor feeds to generate battlefield target lists in minutes. When asked about this role during testing, Claude reportedly said it found the situation 'genuinely troubling.' The report links the system to a February strike in Iran that killed approximately 168 people at an elementary school.
See the business automation angle on this story — free with signup.
Google Delays Gemini 3.5 Pro Launch to July for Additional Tuning
Google has pushed back the release of its next flagship model, Gemini 3.5 Pro, from June to July, according to Business Insider. The company is spending extra time gathering feedback from early testers and incorporating lessons learned from its recently released Flash 3.5 model. The delay signals Google is taking extra care with its frontier model as competitive pressure from Anthropic and OpenAI intensifies.
See the business automation angle on this story — free with signup.
Google Loses More Top AI Talent to Anthropic and OpenAI as Pre-IPO Equity Proves Irresistible
Two more senior Google researchers who worked on Gemini — Jonas Adler and Alexander Pritzel — are departing for Anthropic, following Nobel laureate John Jumper and legendary AI researcher Noam Shazeer (who is joining OpenAI) in recent weeks. Business Insider reports that the primary lure is pre-IPO equity at Anthropic and OpenAI as both companies prepare to go public. Analysts say the talent drain could slow Google's AI development pace while strengthening its rivals.
See the business automation angle on this story — free with signup.
Washington Post Study Finds AI Chatbots Skew Left on Political Questions, With ChatGPT Leading
A Washington Post analysis testing five major AI models on 29 political topics found that GPT-5.5 gave left-leaning answers roughly 80% of the time and right-leaning answers just once. Anthropic's Claude Opus 4.8 gave left-leaning answers 43% of the time and never gave exclusively right-leaning answers, while Google's Gemini offered both-sides responses over 90% of the time. Grok 4.3 was the most ideologically balanced, splitting answers roughly between left, right, and both-sides.
See the business automation angle on this story — free with signup.
Global Developments
2Israel Announces National AI Plan Including Procurement of 100,000 GPUs for Sovereign Infrastructure
Israel's Prime Minister's Office has published a national AI initiative aiming to establish the country as a global AI superpower, including an ambitious plan to procure 100,000 GPUs for sovereign government and national use. Analysts estimate the GPU acquisition alone could cost tens of billions of dollars, with some experts calling full technological independence 'overly ambitious' for a country of Israel's size. The plan also targets sovereign AI infrastructure and workforce development.
See the business automation angle on this story — free with signup.
Anthropic Is Hiring Across Australia and Japan as It Races to Build Asia-Pacific Data Center Capacity
Anthropic is actively recruiting for 13 compute roles focused on data center development, eight of which are based in Australia or Japan, according to CNBC. The company is sourcing data center deals and hiring electrical engineers in Japan specifically, signaling that the Asia-Pacific region is its next major infrastructure frontier after a wave of U.S.-based deals earlier this year. Power access and copyright laws remain challenges in both markets.
See the business automation angle on this story — free with signup.
Robotics
2AMC Robotics Secures Vietnam Manufacturing Facility to Scale NovaArm and Quadruped Robot Production
Nasdaq-listed AMC Robotics has signed a lease for a 6,150-square-meter manufacturing facility in Bắc Ninh, Vietnam to serve as its Southeast Asian production hub. The facility will initially support production of the NovaArm industrial robot arm and later expand to include the Kyro quadruped robotic dog. The move reflects a broader industry trend of robotics firms diversifying manufacturing away from China amid trade tensions.
See the business automation angle on this story — free with signup.
BYD Plans to Deploy Humanoid Robots as Car Salespeople in Every Showroom Within Two Years
BYD executive vice president Stella Li told Business Insider at Cannes Lions that the Chinese automaker wants to place humanoid robots in every one of its car showrooms to demonstrate and sell vehicles. Li predicted the technology would be ready within one to two years and said BYD plans to build its own humanoid in-house, though it would consider purchasing from rivals if needed. BYD is developing its own humanoid as a direct competitor to Tesla's Optimus robot.
See the business automation angle on this story — free with signup.
Transportation
4UN Adopts First-Ever Global Regulatory Framework for Fully Autonomous Vehicles
The UN's World Forum for Harmonisation of Vehicle Regulations has approved a global regulatory framework for vehicles equipped with fully autonomous driving systems, marking what officials call a 'decisive milestone' in self-driving car governance. The framework covers 56 member states across Europe, North America, and Asia and requires manufacturers to demonstrate their systems 'pose no unreasonable risk' and provide continuous performance monitoring. The move aims to prevent a patchwork of conflicting national rules from fragmenting the global autonomous vehicle market.
See the business automation angle on this story — free with signup.
Drone Delivery Pioneer Zipline Expands Into Austin, Texas
Zipline, the world's largest autonomous drone delivery company, plans to launch operations in Austin later in 2026, extending its Texas presence after successful deployments in Dallas and Houston. Customers order through the Zipline app and receive items lowered from the sky via a tether, entirely autonomously. The company raised more than $600 million earlier this year to fund its expansion.
See the business automation angle on this story — free with signup.
Amazon's Zoox Unveils Redesigned Purpose-Built Robotaxi Heading Into Large-Scale Production
Zoox, Amazon's autonomous vehicle subsidiary, revealed an updated version of its ground-up designed robotaxi and announced it is entering large-scale production at its California facility. Unlike Waymo, which uses modified passenger vehicles, Zoox builds its robotaxi from scratch for a fully driverless experience. The company expects to launch ride-hailing services in partnership with Lucid and Nuro in 2027.
See the business automation angle on this story — free with signup.
Waymo Releases 220-Million-Mile Safety Data Showing 94% Fewer Fatal Crashes Than Human Drivers
Waymo published new safety data covering more than 220 million fully autonomous miles, reporting 94% fewer crashes involving serious or fatal injuries compared to human drivers in the same areas. The company also recorded 93% fewer pedestrian-related injury crashes and 84% fewer crashes involving cyclists. The data release comes as Waymo continues rapid expansion and recently announced 500,000 autonomous rides per week.
See the business automation angle on this story — free with signup.
Healthcare & Biotech
3Iambic AI Launches Drug Discovery Collaboration With Bayer Using AI Small-Molecule Platform
Clinical-stage AI drug discovery company Iambic AI has announced a collaboration with pharmaceutical giant Bayer to develop novel small-molecule medicines using Iambic's AI platform. The partnership adds Bayer to a growing list of pharma majors exploring AI-first drug discovery pipelines. Separately, digital health company Withings acquired respiratory monitoring startup Biosency to expand its connected health portfolio.
See the business automation angle on this story — free with signup.
Mount Sinai Partners With Signal 1 to Centralize Governance and ROI Tracking Across Its Growing AI Portfolio
Mount Sinai Health System is deploying Signal 1's AI Management Platform to unify governance, performance monitoring, and return-on-investment tracking across its diverse AI tools, spanning imaging, generative AI, and emerging agent-based systems. The partnership reflects a growing recognition among large health systems that managing dozens of AI tools simultaneously requires dedicated infrastructure. Mount Sinai's chief digital transformation officer said the priority is ensuring safe, scalable AI deployment without slowing innovation.
See the business automation angle on this story — free with signup.
Medra Launches AI 'Experimentalist' Reasoning Layer for Lab Robotics in Partnership With DARPA
Biotech startup Medra has launched 'AI Experimentalist,' a scientific reasoning layer for its drug discovery robotics platform, in collaboration with DARPA. The system is designed to bridge the gap between AI hypothesis generation and physical experimental validation, addressing what Medra's CEO calls a fundamental 'data problem that is actually a robotics problem.' The platform accounts for subtle physical variables — like pipette angles and reagent timing — that can make or break experimental results.
See the business automation angle on this story — free with signup.
Investment Noteworthy
3Mirendil, Founded by Anthropic Veterans, Raises $200M at $1 Billion Valuation From a16z, Kleiner, and Nvidia
Mirendil, a San Francisco-based AI startup founded by former Anthropic researchers, has raised $200 million at a $1 billion valuation in a round led by Andreessen Horowitz, Kleiner Perkins, and Nvidia. The company is an early-stage bet on what its investors see as the next wave of AI infrastructure or model capabilities. The funding is part of a broader surge in seed and early-stage investments in companies founded by alumni of frontier AI labs.
See the business automation angle on this story — free with signup.
Prosus Invests $460 Million in French AI Health Insurer Alan, Valuing It at $6.3 Billion
Dutch investment giant Prosus has made a $460 million investment in French health tech startup Alan, pushing its valuation to $6.3 billion. Alan offers AI-driven health insurance and telehealth services across France, Belgium, Canada, and Spain, and the new funding will accelerate its international expansion and AI-led product development. Prosus will also provide Alan access to its Large Commerce Model and AI ecosystem.
See the business automation angle on this story — free with signup.
AI Now Accounts for Nearly Two-Thirds of All European VC Deal Value at €21.3 Billion — Ahead of Forecasts
PitchBook's EMEA midyear private capital update finds that AI is dominating European venture investment, representing nearly two-thirds of total VC deal value at €21.3 billion so far in 2026 — well ahead of year-start expectations. The report also notes that the ratio of PE-backed to publicly listed companies in Europe hit a record 2.3x in Q1, driven partly by a contraction in public listings. The data confirms that AI investment concentration is a global phenomenon, not just a U.S. story.
See the business automation angle on this story — free with signup.
General AI
2AI 'World Models' Emerge as the Next Frontier as Top Researchers Shift From Language to Physical Reasoning
A growing cohort of AI researchers, including former LLM specialists, are pivoting to 'world models' — AI systems designed to understand and navigate physical environments rather than just process text. Unlike language models that predict the next word, world models learn to simulate physical cause and effect, making them foundational for robotics and autonomous systems. The shift reflects a broadly held view among top researchers that language model scaling alone is hitting diminishing returns.
See the business automation angle on this story — free with signup.
New Survey Identifies 'Ghost Downsizing' as an Emerging AI-Linked Workforce Trend in U.S. Companies
A 2026 survey by Omni Calculator identifies 'ghost downsizing' — where companies quietly leave roles unfilled after departures and redistribute work to remaining employees using AI — as an emerging workforce pattern linked to AI adoption. The technology and media sector showed the highest level of this type of workforce change. The survey also found a class divide in AI perception: higher-income workers view AI as a productivity tool, while lower-income workers see it primarily as a job threat.
See the business automation angle on this story — free with signup.